Bachelor and Master Theses

Title:Developing Information Models to support Systems Engineering 
Subject: Computer Science
Level: Advanced
Description:In the context of industrial systems/software engineering, systems are built from existing software/hardware components, as well as from newly developed components. The advantage of this approach is that components can be developed separately. Often, Functional architectures need to be developed to support product and system families.To life cycle manage these functional architectures as well as support industrial systems development, information models are required to capture all artefacts e.g. requirements, functional archtiectures, system design, phycial architectures etc. The goal of the thesis is detailed as follows: Develop an information model to support modeling of system software aspects. Investigate tool support  
Company: Volvo Construction Equipment, kontaktperson: Jagadish Suryadevara
Proposed:2018-03-01 
IDT supervisor: Jagadish Suryadevara
jagadish.suryadevara@mdh.se, 021- 10 31 16
Examinator:

,
IDT supervisor:
,
Title:Data-driven Modelling on Powered Two Wheelers using Machine Learning 
Subject:
Level: not set
Description:Problem description According to the European Commission (ec.europa.eu), in 2015, 15% motorcycle riders and 3% moped (and similar powered two-wheelers) riders road fatalities. As it can be found in [1-2], there are around 90% of road-traffic crashes caused by driver error (i.e. as inattention, loss of vigilance, mental under/overload) and unsafe behavior (i.e. inadequate training or lack of experience). Improving road safety includes understanding the individual, collective and interaction behaviour of riders. This work proposes a development of a methodology for riding patterns classification by using machine learning techniques. The riding pattern classification problem will be formulated as a classification problem aiming to identify the class of the riding situation by using inertial sensor data. This inertial sensor data was collected from three accelerometer and three-gyroscope sensors mounted on the motorcycle. These measurements constitute experimental database which was valuable to analyze Powered Two Wheelers (PTW) rider behavior. In a previous work [3], the obtained results based on the raw 3D inertial measurements (accelerometers / gyroscopes) data shown the effectiveness of such approach. The project work is subdivided as follows: 1. Literature study and state-of-the-art This task requires a systematic literature review to identify the features, time and frequency domain analysis approaches for feature extractions, approaches for feature selection/ranking, and machine learning approaches for classification for PTW riding manoeuvring. Student requires presenting an analytical summary of the state-of-the-art based on the literature study. 2. Implementation Analysing the PTW's riding data of various events for a given dataset. Student also requires developing an approach using machine-learning algorithms for detection of different riding patterns based on the riding events and sensory dataset. 3. Evaluation Student should evaluate the proposed approach and learning algorithm for detecting the patterns. Make a comparison between different feature extraction and selection methods on the data set that will be given. 4. Report writing It is expected that student provide a report as a completion of the project work. Report consists of background, problem formulation, state-of-the-art, methods, evaluation, and discussion. Reference: 1. Elander, J., West, R, French D, "Behavioural correlates of individual differences in road-traffic crash risk: An examination of methods and findings", Psychological bulletin 113.2 (1993): 279 2. Feyer, A.M., Williamson, A. & Friswell, R., "Balancing work and rest to combat driver fatigue: an investigation of two-up driving in Australia". Accident Analysis and Prevention 1997 Jul, 29, 541-553. 3. F. Attal, A. Boubezoul, L. Oukhellou and S. Espié, "Riding patterns recognition for Powered two-wheelers users' behaviors analysis," 16th International IEEE Conference on Intelligent Transportation Systems (ITSC 2013), The Hague, 2013, pp. 2033-2038.  
Proposed:2018-03-06 
IDT supervisor: Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
IDT supervisor:
,
Title:Machine learning based pedestrian event monitoring using IMU and GPS 
Subject: Computer Science
Level: Basic
Description:Road transportation is a complex system and poses challenges in safe transportation due to dynamic environments changes, where multiple actors (e.g. Vehicle driver, Motorcyclist, Bicyclist, Pedestrian) are involved. It has been reported that 90% of traffic crashes are caused due to drivers'/riders'/Pedestrians' risky behaviour. To improve the traffic safety it is not enough to analyse only the risky behaviour of the driver but also a collective behaviour analysis of several actors such as pedestrians. The aim of this thesis is to investigate and implement a Pedestrian model for different events using IMU sensor, GPS signals. Here machine-learning algorithms should include in the model development to find patterns for different walking/running behaviours. The project work is subdivided as follows: 1. Literature study and state-of-the-art This task requires a systematic literature review to identify the parameters, metrics and indices for Pedestrian that can be used modelling Pedestrians' behaviour in road transport environment. Student requires presenting an analytical summary of the state-of-the-art based on the literature study. 2. Implementation This task involves analysing the Pedestrians' data using IMU sensor and GPS signals of various walking/running events and developing an approach using machine-learning algorithms for detection of these patterns based events. 3.Evaluation Student should evaluate the proposed approach and learning algorithm for detecting walking/running events. 4.Report writing It is expected that student provide a report as a completion of the project work. Report consists of background, problem formulation, state-of-the-art, methods, evaluation, and discussion.  
Proposed:2018-02-09 
IDT supervisor: Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
IDT supervisor:
,
Title:Automatic derivation of data flow graphs from software models for parallel execution 
Subject:
Level: Advanced
Description:In model-driven software engineering, the focus is shifted from programming to modelling, thus from code to models. Code is then automatically generated from models. In order to properly generated code to run on parallel platforms, models need to be analyzed and code generators to be properly instructed. One of the crucial analyses is data flow analysis of the model behaviour. In this thesis, the student(s) will provide a theory and implementation of a mechanism for automatically deriving data flow graphs from software models described in UML in order to identify independent branches to be potentially run in parallel. 
Proposed:2017-10-24 
IDT supervisor: Federico Ciccozzi
federico.ciccozzi@mdh.se, +46 21 151762
Examinator: Jan Carlson
Jan Carlson
jan.carlson@mdh.se, +46-21-151722
Title:Master Thesis: A mapping study on MDE adoption effects in industry (2 students) 
Subject:
Level: Advanced
Description:CONTEXT AND AIM: The complexity of software systems grows continuously due to their pervasive use in almost any aspect of everyday life. To alleviate the intricacy of their development, Model-Driven Engineering (MDE) proposes to shift the focus from coding to design. Models are well-defined abstractions of the reality that allow, indeed, to abstract away those details that do not matter a domain-specific point-of-view. This simplified, yet detailed, representation of a certain sub-problem allows also to perform early analysis of the system and hence to anticipate issues that would be much more expensive to be solved at late stages of the development process. Due to its promises, MDE attracted remarkable attention from industry, and since its introduction around year 2000 has been adopted in every application domain. In this respect, there exist several documented stories of successful adoption of MDE [2, 3]. However, partly due to the immaturity of the methodology and supporting tools, partly due to the inadequacy of companies’ personnel skills, MDE had controversial effects on development processes and in some cases disappointed most of the expectations. In the latest years, more and more empirical studies have been documenting the experiences of companies that adopted MDE, and trying to highlight what are the main gains and open issues for the adopters [4]. Nonetheless, it is still very difficult to reply to the following question: “Given my company characteristics, what would be the most prominent effects I would experience if adopting MDE?”. In other words, it is still very difficult to answer with a simple “Yes” or “No” to the question about the opportunity of adopting MDE in a certain company. GOALS: This thesis work has the goal to investigate the current literature about the adoption of MDE in industry. The aim is to identify, if possible, a set of characteristics relating expectations and outcomes preceding and following the adoption of MDE, respectively. The study will rigorously follow empirical studies guidelines [1]. In particular the work will: - collect relevant publications about the adoption of MDE in industry; - identify trends over time; - relate expectations and outcomes with respect to the MDE adoption effects. THESIS EXECUTION: The work described in this thesis proposal is expected to be developed by two students. The work will be done in collaboration with Bombardier Transportation (BT), which will share the supervision. BIBLIOGRAPHY: 1. K. Petersen et al.: Guidelines for conducting systematic mapping studies in software engineering. Inf. Softw. Technol. 64, C (August 2015), 1-18. DOI=http://dx.doi.org/10.1016/j.infsof.2015.03.007 2. http://www.nyteknik.se/fordon/nya-gripen-flyger-i-simulator-6576208 3. https://www.nasa.gov/sites/default/files/01-03_orion_cre_exploration_vehicle_model_0.pdf 4. John Hutchinson, Jon Whittle, Mark Rouncefield, and Steinar Kristoffersen. 2011. Empirical assessment of MDE in industry. In Proceedings of the 33rd International Conference on Software Engineering (ICSE '11). ACM, New York, NY, USA, 471-480. DOI=http://dx.doi.org/10.1145/1985793.1985858 
Proposed:2017-10-10 
IDT supervisor: Antonio Cicchetti
antonio.cicchetti@mdh.se, +46-21-151762
Examinator: Mikael Sjödin
Mikael Sjödin
mikael.sjodin@mdh.se, +46 70 288 2829
Title:Generation of Mutants for Testing Execution Time 
Subject: Computer Science
Level: Advanced
Description:Targeted mutation is a paradigm for mutation testing of non-functional properties, where the mutations are focused to the parts of the code that are likely to have a significant impact on the property in question [1]. This methodology can be applied to Time analysis, arguing that the parts of the code that can affect the program flow provide the best mutation targets. These parts are easily identified by program slicing. Expected Outcome: The students will investigate the combination of (i) Static Analysis – identifying the parts of the code with a strong influence on the execution time, and (ii) Mutation Testing – injecting changes in the identified parts of the code. The solution will be implemented as a part of a framework that basically explores the relationship between mutants and control flow. [1] B. Lisper, B. Lindström, P. Potena, M. Saadatmand, M. Bohlin: “Targeted Mutation: Efficient Mutation Analysis for Testing Non-Functional Properties”, ICST Workshops 2017: 65-68. http://www.es.mdh.se/pdf_publications/4660.pdf This thesis is defined as part of the TOCSYC project that is a five-year collaboration project funded by the Knowledge Foundation. The project partners are Mälardalen University, Blekinge Institute of Technology, University of Skövde, Karlstad University, and SICS Swedish ICT AB.  
Company: RISE SICS Västerås, kontaktperson: Pasqualina Potena
Proposed:2017-10-02 
IDT supervisor: Mehrdad Saadatmand
mehrdad.saadatmand@mdh.se, +46-(0)21-107336
Examinator: Björn Lisper
Björn Lisper
bjorn.lisper@mdh.se, +46-21-151709
Title:A Decision Support System for medical diagnosis using Data Mining and Machine Learning 
Subject: Computer Science
Level: Advanced
Description:1.Background of the work Today, due to deployment of wireless sensor devices and its communication through the Internet, a huge amount of sensor data could be achieved continuously for example, for ECG. However, the optimal use of this sensor reading is not properly done since a lot of hidden information is not mine properly. Therefore, discovering of new knowledge related with diseases (e.g. anomalies in ECG signal for Heart disease) is an open and challenging issue. This also effects very much on medical decision support in proper decision-making task. Data mining and machine learning techniques through Artificial Intelligence (AI) approaches could take the advantage to develop a Decision Support System (DSS) for medical diagnosis by identifying new/unknown knowledge and relate them with a specific disease. 2.The problem The work will develop a DSS for medical diagnosis using AI techniques such as machine learning algorithms and data mining. The main tasks are: a) Literature Study and State-of-the-art: finding related algorithms for knowledge discovery and DSS, current advancement of the area, identified drawback and limitation of the chosen methods, etc. b) Implement chosen methods (at least 3 to 4) for knowledge discovery and test them with sensor signals such as ECG, identify best methods with new and unknown knowledge. c) Identifying best machine learning methods and test them. d) Developing DSS for diagnosis using best methods and evaluated them with other commonly used methods. 3.Expected outcomes The main outcome of the thesis is to implement a DSS for medical diagnosis based on a literature study, which counts the state-of-the-art of this area as a scientific contribution.  
Proposed:2017-11-14 
IDT supervisor: Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
Title:Towards an Eclipse Modelling Environment for LLVM 
Subject:
Level: Advanced
Description:LLVM is a well-established compiling infrastructure, which allows to both exploit existing compilers and design brand new ones. In model-driven software engineering, the focus is shifted from programming to modelling, thus from code to models. Code is then automatically generated from models. A de-facto standard platform for software modelling is currently represented by Eclipse. Towards the establishment of an LLVM model compiler, a first step is represented by bridging the Eclipse modelling environment and the LLVM compiling infrastructure. In this thesis, the student(s) will investigate existing solutions, identify needs, and build a prototype implementation of a textual model editor for the LLVM intermediate representation language in Eclipse. 
Proposed:2017-10-24 
IDT supervisor: Federico Ciccozzi
federico.ciccozzi@mdh.se, +46 21 151762
Examinator: Björn Lisper
Björn Lisper
bjorn.lisper@mdh.se, +46-21-151709
Title:Data and control flow analysis for industrial systems developed in IEC 61499 
Subject:
Level: Advanced
Description:IEC 61499 is an industrial standard defining a high-level, partly graphical, language for development of industrial control systems. Applications can be defined as collections of interconnected function blocks that, in turn, can be defined by smaller function blocks or by a special type of state machines.
In our previous research, we have developed a new method to analyse control dependencies in systems implemented in IEC 61499, i.e., the different ways in which function blocks can trigger each other to execute. Moreover, an initial version of a corresponding analysis of data dependencies has been developed.
The goal of this thesis project will be to i) further improve the data dependency analysis to correctly handle for example variables in transition guards; and ii) develop a new method for combined control and data flow analysis. The developed analysis algorithms should be formally defined and then implemented as a plugin for the open source development environment 4DIAC. Finally, the implementation should be used to validate the developed analysis in a small case study on a suitable system. 
Proposed:2017-09-19 
IDT supervisor: Jan Carlson
jan.carlson@mdh.se, +46-21-151722
Examinator:

,
Title:A Survey on Service Level Agreement Definition for Cloud Services in IoT 
Subject: Computer Science
Level: Basic
Description:Cloud computing and Internet of Things (IoT) are computing technologies that provide services to consumers and businesses, allowing organizations to become more flexible. A formal document between a service provider and a service consumer that defines the level of service expected from the service provider is called a service level agreement (SLA). The quality of service, integrated in the SLA, is an important issue for both service provider and service consumer. Before the services can be provided to the consumer, both the provider and the consumer must agree on the metrics, level, quality, price and penalties related the services. Various metrics can be part of an SLA and the level of service definitions need to be measurable and specific in each metrics. A few surveys and systematic reviews relevant to the SLAs in cloud computing and IoT have been conducted. The goal of this thesis is to conduct a detailed investigation of one of the existing systematic mapping study [1] and present a detailed state-of-the-art survey on SLA Definition. [1] S. Mubeen, S. A. Asadollah, A. V. Papadopoulos, M. Ashjaei, H. P. Breivold, and M. Behnam, ''Management of service level agreements for cloud services in IoT: A systematic mapping study,'' IEEE Access, to be published.  
Proposed:2018-03-16 
IDT supervisor: Sara Abbaspour Asadollah
sara.abbaspour@mdh.se,
Examinator:

,
IDT supervisor:
,
Title:Deep Learning based Eye Tracking and Head Movement Detection  
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description: When human communicates with each other through speech, several gestures such as facial expressions, eye movement, head movement etc. provide complementary information as communication channels [1]. Eye tracking and head movement detection are one of the most interesting research area in the field of Image Processing and Computer Vision and the tasks are the fundamental contributor for human computer interaction (HCI). Eye tracking and head movement detection are co-related with each other and have been an active research field in the past years as it adds convenience to a variety of applications. Both technologies are considered the easiest alternative interface methods. Eye tracking means the process of measuring either the point of gaze (where one is looking) or the motion of an eye relative to the head. For both eye tracking and head movement detection, several different approaches have been proposed and used to implement different algorithms for these technologies. Examples of different fields of applications for both technologies, such as human-computer interaction, driving assistance systems [2], and assistive technologies need to be investigated. In this project, you need to use image processing and computer vision techniques to build your intelligent system with the help of different artificial intelligence algorithms. The project works can be subdivided as follows: 1. Literature study and survey of existing methods and systems 2. Data Collection using Camera both lab environment and driving situation 3. Develop expected methods 4. Evaluation 5. Outcome Successfully eye tracking and head movement detection will be the expected outcome in different angles, positions and situations. Extra Qualification: Image processing and Computer vision Concept (or highly ambitious to learn)  
Proposed:2017-11-14 
IDT supervisor: Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
Title:An intelligent system for driver cognitive load detection using eye tracking data  
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Vehicle driving requires attention, effort, alertness and quick reactions. Reduced vehicle control due to different mental states can lead to a severe accidents. One factor that affects the driver’s mental state is his/her level of cognitive load. Cognitive load can be defined as effortful and conscious resource allocation within the brain, needed to deal with non-routine or inherently difficult tasks, resulting in controlled performance. Humans have limited capacity for cognitive load and engaging in one cognitively loading task interferes with one’s ability to at the same time engage in other cognitively loading task. Performing cognitively loading secondary tasks, such as talking on the phone, while driving can hence affect the performance in the primary task, i.e. the driving. In the last few years, eye tracking has become a non-invasive method for driver cognitive load detection and analysis. Measures such as pupil diameter and gaze concentration have been shown to correlate with level of cognitive load. Factors such as ambient light and traffic environment also have a great effect on those measures though, making interpretation of them difficult in applied settings. This thesis project aims to develop an intelligent system for driver cognitive load detection using eye-tracking data. It also requires investigation of eye-tracking data that have been collected in a simulator study for driver cognitive load detection and analysis. The project work can be subdivided as follows: 1. Literature study and survey of existing methods and systems 2. Data analysis 3. Development of an intelligent system using machine learning algorithm 4. System evaluation  
Company: Volvo Car, kontaktperson: Emma Nilsson
Proposed:2017-11-14 
IDT supervisor: Shaibal Barua
shaibal.barua@mdh.se,
Examinator: Mobyen Uddin Ahmed
Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Title:Data-driven cognitive load classification system using machine-learning algorithm 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Sitting behind a wheel, i.e., driving requires a high degree of concentration and dynamic and complex activities e.g., visual, cognitive and manual tasks are involved during driving. The driver has to make strategic decision, monitor the roadway and surrounding environment as well as inside the vehicle system, process information and execute control level activities. All these activities impose workload and cognitive load on the driver. Physiological measures are sensitive indices for detecting changes in mental workload. It is suggested in the literature that individuals devote their mental resources to keep up a given level of performance until the point at which their resources are exhausted. For the development of advanced safety system, it may be significant to find out the temporal relationship indicating shifts in physiological arousal due to mental workload before driving performance is impaired. Studies have shown that cardiovascular measures such as heart rate and blood pressure increase with the increasing of cognitive demand. Skin conductance is another measure that also increases with the increasing cognitive demand. Various studies have suggested that these physiological measures may be useful indices of mental workload, however, the overall pattern of cognitive load and driving performance is unclear. The goal of the thesis work is to analyse physiological signals i.e., heart rate variability, skin conductance, and respiration rate for cognitive load detection. It also aims to investigate machine-learning classification for automatically identification of cognitive workload for different task scenarios. The work requires literature study, signal analysis and feature extraction, and implementation of an intelligent system using machine-learning algorithms for classification of cognitive load and evaluation of a prototype system.  
Company: Volvo Car
Proposed:2017-11-07 
IDT supervisor: Shaibal Barua
shaibal.barua@mdh.se,
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
Title:Human Emotion Detection using Deep Learning 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Human face reveals the healthy status of an individual through a combination of physical signs and facial expressions. Facial expressions reflects the overall health status and it is directly correlate to our emotions. Emotions indicate the physiological status. Therefore it is very important to identify the facial expression to know our emotions. The goal of this project is to detect human emotions using facial expressions. In this projects you have to develop a non-contact system to detect correct human emotions using facial expressions classification. The system should detect whether a person is anger, happy, sad, smiley, relaxed, surprised and so on. Your system should have one or several artificial intelligence algorithms to build the intelligent system but you are free to use any classification algorithms. The project works can be subdivided as follows: 1. Data Collection non-contact (Camera ) systems 2. Feature extraction for different facial expressions 3. Feature classifications and detect emotions 4. Evaluation 5. Outcome Expected Output: Detection of Emotional state  
Proposed:2017-11-14 
IDT supervisor: Hamidur Rahman
hamidur.rahman@mdh.se,
Examinator: Mobyen Uddin Ahmed
Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Title:Test-case Generation For Timing Properties From EAST-ADL Models 
Subject:
Level: Advanced
Description:Testing is an important activity in assessing the quality of a software product, and generation of test cases is one way to introduce automation in testing and to make it more efficient. Models can serve as the main source of information to generate test cases from. In this thesis work, the goal is to generate test cases for the timing properties that are specified using EAST-ADL modeling language. To achieve this, for instance, a test-case generation algorithm can be designed and implemented as part of a model transformation which navigates the source model, identifies and extracts properties of interest, and creates as output test specifications. In short, the following sub-tasks and challenges will be addressed in this thesis: 1) categorization of timing properties captured in an EAST-ADL model 2) defining appropriate test cases for each type of timing property 3) designing a test-case generation algorithm for the timing properties 4) implementation of the algorithm. This thesis work is defined in the scope of the TOCSYC project. TOCSYC is a research project establishing an environment combining five Swedish groups in software testing research to advance the knowledge of testing critical characteristics of complex embedded systems. The overall goal of TOCSYC is to enable and support cost-effective testing for critical characteristics in embedded systems by providing Swedish industry with new and improved tools and techniques for efficient and effective testing as well as the decision-support procedures necessary to select the right testing tools or techniques for their context.  
Proposed:2017-09-02 
IDT supervisor: Mehrdad Saadatmand
mehrdad.saadatmand@mdh.se, +46-(0)21-107336
Examinator: Antonio Cicchetti
Antonio Cicchetti
antonio.cicchetti@mdh.se, +46-21-151762
Title:Classifying cognitive load using vehicle’s state and driving behaviour signals 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Vehicle driving requires a high degree of concentration and dynamic, and complex activities, e.g., visual, cognitive and manual tasks are involved during driving. Through cognition process, a person understands the information about an external object or phenomenon in response to the influence of acquired knowledge, memory, and experience. Therefore, a vehicle driver has to make strategic decision, monitor the roadway and surrounding environment, as well as, inside the vehicle system, process information and execute control level activities. These activities impose cognitive load and the level of cognitive load affects the mental state of the driver. Vehicle driving involves the interaction of driving actions and the vehicle’s state. With the increase cognitive load, vehicle’s state and driver’s action differ from the situation in which the driver’s cognitive load is normal. Driver’s action depends on the level of cognitive load and vehicle’s state (e.g., the speed of the vehicle). Hence, the level of driver’s cognitive load can be classified from the combined vehicle’s state and driving behaviour signals. Driving behaviour signals are obtained from vehicular data, which represent “longitudinal control” or “lateral-control” action. The forward motion of a vehicle achieves by longitudinal control action, which can be obtained from the accelerator or brake. On the other hand, the direction of the vehicle achieves by lateral-control action, which can be obtained from steering wheel operation. Another example of lateral control action is yaw and yaw rate; yaw is simply an indication of a vehicle's rotation from it's vertical axis or how far is the vehicle angled to the left or right away from its centre or how far has the vehicle deviated from it's straight course. In addition, the level of cognitive load can be measured using the steering-entropy method that quantifies the roughness of steering when a driver operates under various loads. The steering-entropy method was developed to investigate the degree to which the operating vehicle equipment burdens the driver. This thesis aims to develop an intelligent system for driver’s cognitive load classification using vehicle’s state and driving behaviour signals. It requires investigation of steering-entropy method, data analysis of vehicular signals, and implementation of machine learning algorithm to classify level of cognitive load. The project work can be subdivided as follows: 1. Literature study and state-of-the-art on methods and systems 2. Driving behaviour i.e., vehicular data analysis 3. Development of an intelligent system using machine-learning algorithm 4. System evaluation  
Proposed:2017-11-07 
IDT supervisor: Shaibal Barua
shaibal.barua@mdh.se,
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
Title:Evaluating an Industrial Internet of Things Protocol in the Fog and Cloud 
Subject: Computer Science
Level: Basic
Description:Description: This thesis work is defined in the scope of the Future Factories in the Cloud (FiC) project. It envisions future factories being designed by compositions of smart connected components, with a large part of the intelligence residing in the Cloud. This will enable increased flexibility and evolvability of manufacturing, as well as pave the way for new business models where production facilities can be accessible as cloud services. Cloud computing is an Internet-based computing paradigm that provides an on-demand access to shared resources. Although Cloud computing is able to handle huge amounts of data from IoT devices, but sending huge amount of data to Cloud servers is a challenge due to the limitation of bandwidth and unpredictable number of users. Thus, there is a need to another paradigm that shifts the processing closer to IoT devices, which is Fog computing. Fog computing is a novel paradigm in computing that aims to process data near data source, where it enables new applications and services. Fog computing significantly decreases the data volume that was supposed to be sent to Cloud servers. Fog computing enables data analytics and knowledge generation to occur at the data source. It also helps to attain better accuracy that enables new applications. In this Thesis, we are envisioning a two-level closed-loop control system, composing of hosts, Fog server and Cloud server. This system is supposed to be implemented in one of the common IoT platforms, namely Contiki OS. It considers wireless link between hosts and the Fog server, and getting advantage of the current protocol stacks in Contiki, such as RPL routing using IEEE 802.15.4 radio. There are some aggregation functions that are supposed to be applied to the host, Fog and Cloud devices. This Thesis will focus on the functionally of such a network with both local and remote control loops. The main goal is to enable measuring the responsiveness and reliability of both control systems, and to classify services based on their timing requirements on different control units. Problem statement: There are various Cloud Computing platforms in the literature. However, there is a need to devise a system that encompasses both Cloud and Fog servers for IoT applications that provides the opportunity of employing different IoT protocols. Main outcome: The main outcome of this Thesis is to show the performance of IoT applications in the existence of Fog servers. Tasks: - literature review on the Fog and Cloud - run mobility plugin in Contiki - define different tuples including time (year,month,day,hour,min,sec), location (x,y,z), RSSI, … - develop closed-control loop between host and Fog - develop closed-control loop between host and Cloud - apply aggregation functions to the Fog and Cloud - evaluate network performance metric in different situations This thesis is suitable for 2 students. 
Proposed:2018-01-04 
IDT supervisor: Maryam Vahabi
maryam.vahabi@mdh.se,
Examinator: Mats Björkman
Mats Björkman
mats.bjorkman@mdh.se, +46-21-107037
Title:Generic case based recommender module using AI and machine learning 
Subject: Computer Science
Level: not set
Description:Recommender funktionalitet are used in many different contexts, such as recommender systems and decision support systems. By developing a recommender function that can take cases, domain knowledge and a seed and return "similar" cases with explanation why they are similar would be a valuable funktion. By designing a clear interface and implement such a function the function can be explored and evaluated. There are databases with cases available that can be used for evaluation. 
Proposed:20180115 
IDT supervisor: Peter Funk
peter.funk@mdh.se, +46-21-103153
Examinator:

,
Title:A configurable slider for enhanced model understanding 
Subject:
Level: Basic or Advanced (contact supervisor)
Description:Modelling and model-driven engineering are current practices in industry for efficiently developing complex software. Models are used to both simplify human communication and enable automatic generation of the actual software product. Complex software usually leads to complex models. Nevertheless, some stakeholders still need to be able to abstract from certain details carried by these models depending on their specific needs. In order to allow a swift visual simplification of complex models, this thesis aims at providing a configurable slider, which allows the stakeholder at hand to augment/diminish the level of detail carried by the model both syntactically and semantically. 
Proposed:2017-10-24 
IDT supervisor: Federico Ciccozzi
federico.ciccozzi@mdh.se, +46 21 151762
Examinator: Antonio Cicchetti
Antonio Cicchetti
antonio.cicchetti@mdh.se, +46-21-151762
Title:Cloud computing simulator for industrial systems 
Subject:
Level: Advanced
Description:CONTEXT AND AIM: Cloud computing has attracted a lot of attention in the last years both in academia and industry [1]. However, carrying out research in cloud computing requires a suitable research platform [2], which is typically developed ad-hoc for specific research purposes, and sometimes it might also be far from realistic scenarios. This calls for reliable and modular simulation platforms that are validated over real platforms. This becomes even more relevant when a cloud infrastructure is not present yet, and preventive analysis is required for the deployment in an industrial context. The scope of this thesis can be adapted to the students background, and it can cover different aspects, ranging from identifying and validating models of computational resource usage of different cloud applications, up to the analysis and the development of vertical and horizontal scaling techniques. GOALS: This thesis work has the goal to investigate the current literature about cloud simulation in industrial contexts. Moreover, depending on the students background and interests, it will include analysis of state-of-the-art techniques for vertical and horizontal scaling or models for computational resource usage in cloud applications. THESIS EXECUTION: The work described in this thesis proposal is expected to be developed by two students. The work will be done within the project Future factories in the Cloud (FiC) (http://www.es.mdh.se/projects/430-FiC), in which companies like Volvo and e-on are part of the steering committee, providing interesting industrial use cases. BIBLIOGRAPHY: 1. Georgia Sakellari, George Loukas, A survey of mathematical models, simulation approaches and testbeds used for research in cloud computing, Simulation Modelling Practice and Theory, Volume 39, December 2013, Pages 92-103, ISSN 1569-190X, http://dx.doi.org/10.1016/j.simpat.2013.04.002. 2. Calheiros, R. N., Ranjan, R., Beloglazov, A., De Rose, C. A. F. and Buyya, R. (2011), CloudSim: a toolkit for modeling and simulation of cloud computing environments and evaluation of resource provisioning algorithms. Softw: Pract. Exper., 41: 23–50. doi:10.1002/spe.995 
Proposed:2017-10-06 
IDT supervisor: Alessandro Papadopoulos
alessandro.papadopoulos@mdh.se,
Examinator:

,
Title:Activity monitoring in daily life using Shimmer sensing  
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Background Wearable devices with advanced sensors become more and more used not only for fitness but also for identifying potential health issues. In an aging population fall detection and fall risk assessment is crucial. In Sweden the cost for society is 26.4 billion every year. Sensors can play an important role in early warning system to assess the risk of falling. A wearable device collects data from the senior citizen and an expert system in the cloud will analyze the time series to detect deviations from a normal behavior. This is especially useful for elderly persons living alone in their home environment. A first step was made in the “fallen angel” project spring 2016 focusing on fall detection. Now it is time to look at predicting increased risk of falling. Thesis task The task is to identify activities in daily life (ADL), communicate it to a cloud server and provide trends and changes in user behavior over time. The ADL recognition and analysis will be done in an Android app but also communicated to an existing cloud service. The ADL is checked regular every 5-15 minute and the activity is tested against a pre-defined set of ADL common to most elderly persons. The task includes a theoretical analysis of AI algorithm to be used, collection of training data to recognize activities and a model for identify “normal activity” and “deviations from normal activity” for an individual. The project will use any suitable wearable device such as Shimmer sensing. Expected results - ADL recognition using machine learning algorithms - A working android application monitoring ADL and changes in user behavior - Database with training sets for at least 10 ADL Lessons learned from the implementation  
Company: lifescience technology, kontaktperson: Peter Eriksson
Proposed:2017-11-14 
IDT supervisor: Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
Title:Study Impact of Interference and Mobility on Network Performance in WSNs 
Subject: Computer Science
Level: Basic
Description:Wireless sensor networks (WSNs) are formed by small and low cost devices that are able to sense and transmit data over wireless medium. These networks are characterized by their low-power radios, limited battery power, and weak processing capabilities. Network performance in WSNs is highly affected by two main issues: (1) interference, and (2) node's mobility. In fact, WSNs operate in the unlicensed ISM bands, where they share the radio spectrum with several other devices. For example, in the 2.4 GHz frequency band, WSNs compete with WiFi and Bluetooth devices, cordless phones, and microwave ovens. These devices generate external interference that highly affects network performance. Traditionally, sensor networks were considered as static networks, where nodes. However, in future IoT applications, it is mandatory to employ mobile sensors, where a subset of nodes are allowed to move. For instance, in health monitoring applications, patients with sensors attached on the body are free to move, while data is collected through wireless medium. Mobility has a large impact on the wireless link dynamics, and consequently it influences network performance. The speed of mobile node and mobility pattern are also two main parameters that must be studied. This thesis proposal aims at studying the impact of interference and mobility on WSNs, by reviewing the relevant works, and providing qualitative comparison between the parameters involved. We are also aiming to conduct some real-world experiments with sensor nodes in the existence of interference and node mobility. It involves varying the relevant parameters, such as producing different types of external interference (periodic and sporadic) with weak/strong transmission power, and provide various mobility patterns with different speeds. This thesis is suitable for 1-2 students. 
Proposed:2017-10-19 
IDT supervisor: Hossein Fotouhi
hossein.fotouhi@mdh.se,
Examinator: Mats Björkman
Mats Björkman
mats.bjorkman@mdh.se, +46-21-107037
Title:Study the Impact of Link Quality Indicator on the Reliability of IoT Networks 
Subject: Computer Science
Level: Basic
Description:Due to the low-power data transmission in wireless sensor networks (WSNs), wireless links are highly unreliable. This means that the quality of links is unpredictable, and may vary over time and space. Link quality estimation techniques have been devised for WSNs, in order to calculate, and predict links quality. Although the accuracy of a link quality estimator has a strong impact on the network performance, since packets are supposed to be transmitted over reliable links. Providing an accurate link quality estimator for WSNs is a challenging task as links are more dynamic and unreliable. Several link quality estimators have been devised for WSNs, which are using both hardware and network parameters. The common hardware parameters are received signal strength indicator (RSSI), signal to noise ratio (SNR) and link quality indicator (LQI). These estimators are directly read from the radio transceiver. Thus, they have no additional computational overhead to the protocol design. Packet reception ratio (PRR), traffic level, and rank level are known to be network-based parameters, which require some packet exchanges in order to compute and update these information. In this project, we concentrate on LQI parameter, which is a hardware-based parameter. The LQI parameter has been proposed in the IEEE 802.15.4 standard, and the CC2420 radio reads this radio parameter. There is a correlation between the LQI and the PRR, which depends on network condition. Unfortunately, no mapping table exists in the literature to correlate LQI and PRR parameters in WSNs. Finding a rule that would define their correlation is a novel area of research, which would also benefit this project. In this project, first we study the main correlations between these parameters. We are also aiming at conducting some experiments collecting LQI and PRR values by varying relevant parameters such as transmission power and interference. This thesis is suitable for 1-2 students. 
Proposed:2017-10-19 
IDT supervisor: Hossein Fotouhi
hossein.fotouhi@mdh.se,
Examinator: Mats Björkman
Mats Björkman
mats.bjorkman@mdh.se, +46-21-107037
Title:Data Center Operation Simulation 
Subject:
Level: Advanced
Description:Tasks Data centers are one of large energy consumers with over 35% of energy used for cooling. Energy efficiency is an important issue for data centers. Data center infrastructure modeling and operation simulation are one approach to investigate energy saving potential. The tasks of this thesis work is to further develop data center infrastructure simulation models with MATLAB/SIMULINK to cover subsystems, for example, data center IT system, cooling system and power supply system. Modeling and simulation work will focuses on power consumptions and interactions among subsystems. The simulation models will be used for optimal operation of data centers as industrial processes with various operation scenarios. Proposed activities • Literature survey and study earning knowledge on state-of-the-art. • Develop data center infrastructure simulation models in MATLAB/SIMULINK for specified systems • MATLAB/SIMULINK simulations for operation scenarios • Verify results from the developed models with available data • Master thesis project report and presentation  
Company: ABB CRC, kontaktperson: Xiaojing Zhang (xiaojing.zhang@se.abb.com)
Proposed:2017-06-13 
IDT supervisor: Ning Xiong
ning.xiong@mdh.se, +46-21-151716
Examinator:

,
Title:Combining Runtime Verification and Automated Test Generation for PLC Embedded Software 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Software testing is typically a process where human testers manually write test inputs and expected test results and manually or automatically executing the software on these tests. As a solution to this challenge, automated test generation has been suggested to allow tests to be created with less effort and at lower cost. Runtime verification combines formal verification and testing and is suggested as a practical that can help in finding many errors in software by monitoring software execution. To automate the testing process several researchers have proposed the combination of automated test case generation with runtime verification. In this thesis, we undertake the combination of automated test generation and runtime verification for testing PLC embedded software. The technique will be evaluated on industrial systems.  
Proposed:2017-10-25 
IDT supervisor: Eduard Paul Enoiu
eduard.paul.enoiu@mdh.se, +46-21-101624
Examinator:

,
Title:Using Combinatorial Testing for Testing Industrial Control Software: A Study on Timed Systems 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Combinatorial testing was proposed as a suitable technique for testing software based on a specific choice of input parameters. Test cases are created based on this strategy by varying the values of the inputs. However, this strategy might not be as effective when used on industrial control software for testing timed behaviour. The thesis will investigate the use of extending these techniques in order to incorporate the state and time related behaviour and it will evaluate these techniques on industrial programs written in the IEC 61131-3 programming language.  
Proposed:2017-08-25 
IDT supervisor:
,
Title:Tailoring Combinatorial Testing to Real-Time Embedded Systems 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Combinatorial testing was proposed as a suitable technique for testing complex systems based on a specific choice of input parameters. Test cases are created based on this strategy by varying the values of the inputs. However, this strategy might not be as effective when used on real-time embedded systems for testing complex timed behaviour. The thesis will investigate the extension of these combinatorial techniques in order to incorporate the state-full and timed related behaviour and it will evaluate these techniques on industrial embedded systems. 
Proposed:2017-10-24 
IDT supervisor: Eduard Paul Enoiu
eduard.paul.enoiu@mdh.se, +46-21-101624
Examinator:

,
Title:Model-level timing analysis for UML-RT capsules 
Subject:
Level: Advanced
Description:UML-RT is a subset of UML specifically targeting modelling and development of embedded real-time systems. However, the available timing analysis mechanisms are still fairly limited considering the intended domain. The goal of this thesis is to investigate how an application modelled in UML-RT can be analysed to establish the worst case execution time for the different activities in the system. This requires analysis of the state machines of individual capsules (the main building blocks of UML-RT applications) as well as at system level where capsules are composed into systems. The developed analysis algorithms should be formally defined and then implemented as a plugin for an Eclipse-based UML-RT development environment.  
Proposed:2017-09-19 
IDT supervisor: Jan Carlson
jan.carlson@mdh.se, +46-21-151722
Examinator:

,
Title:Systematic literature review: Impediments for combining model-driven development and continuous integration/delivery 
Subject:
Level: Advanced
Description:Model-driven development is a well-established technique to manage the increased complexity of software systems by allowing developers to define structure and behaviour at a higher level of abstraction compared to traditional programming languages. At the same time, there is a strong current trend towards more agile development with shorter development cycles, striving towards continuous integration, delivery and deployment, also in industrial domains that have previously worked in longer development cycles. The combination of these two is far from straightforward, and is affected by impediments such as insufficient support across the tool chain, weak configuration management at model level, and lack of traceability between different artefacts. The goal of this thesis is to make a in-depth systematic literature review to identify existing research that have identified specific problems related to the combination of model-driven development and continuous integration/delivery, as well as proposals for how to overcome them. 
Proposed:2017-09-19 
IDT supervisor:Jan Carlson
jan.carlson@mdh.se, +46-21-151722
Title:SDN Controllers in Smart Factory 
Subject:
Level: Advanced
Description:Description: This thesis work is defined in the scope of the READY project (Research Environment for Advancing Low Latency Internet, https://ready-sidus.se/). This project aims at reducing latency in various networks, services and applications. Interactive services are disrupted for shorter or longer periods of time, and it becomes hard to deliver media services with low latency. Web services suffer from long delays per transaction which is frustrating for end users because of long waiting times and bad interactivity. Wireless sensor networks (WSNs) are the main building blocks of IoT applications by providing sensing and communication capabilities. IoT applications are enabling connectivity through sensor networks. Smart factory is the current trend of automation and data exchange in manufacturing, where it involves IoT technologies. There are some standards and protocols implemented for industrial automation. REALFLOW is a routing protocol designed for industrial applications in order to support reliable communication by providing multipath routing strategy, however, it lacks immediate routing updates during network changes. This issue can be targeted by applying SDN controllers. Problem statement: Current standard SDN networking have been designed and implemented for wired networks. There are some efforts on implementing SDN for wireless communication. This Thesis aims to develop and test an SDN enabled solution especially for WSNs. Main outcome: This Thesis must provide some solution for connecting an existing SDN controller to the REALFLOW routing protocol for industrial applications. Tasks: • Review the existing SDN solution in wired and wireless domain • Connecting SDN controller to the REALFLOW (developed in OMNeT++) • Evaluation and testing the performance of the network. This thesis is suitable for 1-2 students. The work is done in collaboration with the industry. Qualifications: To be successful in this thesis work the candidate(s) would need the following: • MSc studies in Computer Science or similar area. • Excellent programming skills in C/C++ • Good knowledge of wireless communication systems • Be fluent in English.  
Company: ABB, kontaktperson: Johan Åkerberg
Proposed:2017-10-30 
IDT supervisor: Maryam Vahabi
maryam.vahabi@mdh.se,
Examinator: Mats Björkman
Mats Björkman
mats.bjorkman@mdh.se, +46-21-107037
Title:Heterogeneous IoT Networks for Remote Health Monitoring 
Subject:
Level: Advanced
Description:Description: This thesis work is defined in the scope of the ESS-H research profile (Embedded Sensor Systems for Health Research Profile, http://www.es.mdh.se/projects/324-ESS_H) and ecare@home project (https://ecareathome.se/). These projects aim at providing remote health monitoring of patient through wireless medium. There is a need to utilize both environmental and physiological sensors in health monitoring applications in order to get more accurate information. Thus, it is mandatory to employ various types of sensing devices that measure different parameters while being able to communicate. It is challenging to maintain network reliability while employing various radios with different protocol stacks and standards. Problem statement: In this Thesis, we are focusing on the problem of implementing a heterogeneous network with Bluetooth and IEEE 802.15.4 enabled devices for remote health monitoring. Main outcome: Implementing a test bed that models a heterogeneous network with different radio technologies and evaluate the network performance in terms of network reliability. Tasks: • Run 6TiSCH protocol in an IoT operating system (e.g., Contiki or OpenWSN) on a platform (e.g., OpenMote/Telosb motes) using IEEE 802.5.4 radio • Connecting a network of Bluetooth-enabled Shimmer sensors to a network of OpenMote sensors through a Gateway, where the Gateway is a laptop collecting measurements • Evaluating network reliability under various network size and link conditions through extensive experiments This thesis is suitable for 1-2 students. 
Proposed:2017-10-19 
IDT supervisor: Hossein Fotouhi
hossein.fotouhi@mdh.se,
Examinator: Mats Björkman
Mats Björkman
mats.bjorkman@mdh.se, +46-21-107037
Title:Towards attack models of autonomous Systems of Systems 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:We are witnessing fast technological and industrial advances within area of autonomous systems of systems (SoS). Systems like SoS are built as a collection of several systems that share their resources and capabilities in order to achieve new functionalities, provide better performance or higher level of efficiency, when compared to traditional systems. Systems like this comes with higher level of complexity and providing analysis of its properties is one of the major challenges, since their behaviour might evolve due to the dynamic nature of such systems. It is expected that fully autonomous and cooperating systems increase the production efficiency, and decrease (if not completely replace) the human effort in harmful environments. To enable this, one needs be able to guarantee critical properties of SoS, such as safety and security. It is not sufficient anymore to analyse and guarantee these properties independently, but one has to be able to address safety and security in a joint effort. This thesis will consist of the following: - Given a set of safety requirements for an example of an autonomous SoS (an autonomous construction site), one should explore the interdependencies between the safety and security concerns and identify possible assets of the system that should be protected. - Based on the collected information the thesis is expected to provide details on possible attack models and corresponding safety requirements that can be generated using the combined safety and security reasoning. - Finally, using Goal Structuring Notation (GSN) work should provide an argument that a system is acceptably safe to operate given the set of attack models identified in the previous step. The work will be performed as a part of KKS Prospekt project SAFSEC-CPS - Securing the safety of autonomous cyber-physical systems and in an active collaboration with involved companies in the project (Knightect AB; ABB Robotics; Volvo Construction Equipment). 
Company: Knightec Ab, kontaktperson: David Wenslandt
Proposed:2017-10-15 
IDT supervisor: Aida Causevic
aida.delic@mdh.se, +46-21-107011
Examinator: Kristina Lundqvist
Kristina Lundqvist
kristina.lundqvist@mdh.se, +46 (0)21 101428
Title:Performance Study of Time Sensitive Networks (TSN) 
Subject: Computer Science
Level: Advanced
Description:Background: Distributed embedded systems are nowadays found in many applications, for example, in automotive and automation industries. In these type of applications the amount of data is growing, while exchanging of data within the distributed embedded systems have constraints with respect to timing. This type of systems are commonly known as real-time distributed embedded systems. In order to deal with the mentioned requirements, Ethernet solutions are being considered due to the high bandwidth support. In this regard, the Time-Sensitive Networking task group, which is part of the IEEE 802.1 standardization process, is working on many projects to deliver solutions according to the demands. As part of the IEEE standard, IEEE 802.1Q provides forwarding and queuing techniques for the messages transmitted through a switched Ethernet network. The standard divides the traffic into different classes according to their priorities and adds traffic shapers to prevent burst of traffic on the switch ports. By applying the traffic shapers and providing different classes of traffic, the timing constraints of traffic can be respected. Nevertheless, within the standardization process, two important amendments are being discussed, which are the IEEE 802.1Qbv and IEEE802.1Qbu amendments. The former provides a technique to transmit the traffic which are scheduled a priori with a short latency, whereas the latter provides a preemption mechanism to suspend non-critical messages when a critical message should be sent. Combination of these techniques may provide different performance for critical messages. In this thesis, we are aiming at studying the performance of critical messages when using the amendments by means of simulation or real experiments. In this context, the performance of critical messages refer to their delays and jitters in various network settings. Objective: In this thesis, we are aiming at studying the performance of critical messages when using the amendments by means of simulation or real experiments. In this context, the performance of critical messages refers to their delays and jitters in various network settings. Tasks: 1- study the related work 2- study the state of the arts on distributed embedded systems and time-sensitive networking 3- design a plan for performance study 4- performance study 5- provide an extensive report on the results Time-line: To be carried out for a duration of 20 weeks at the Westermo R&D facilities in Västerås and is suitable for up to two persons. 
Company: Westermo, kontaktperson: Peter Johansson, peter.johansson@westermo.se
Proposed:2017-11-07 
IDT supervisor: Mohammad Ashjaei
mohammad.ashjaei@mdh.se, +46-21-151772
Examinator: Moris Behnam
Moris Behnam
moris.behnam@mdh.se, 021-107094
Title:Speech recognition for autonomous vehicles 
Subject:
Level: Advanced
Description:Background of thesis work Department Emerging Technologies at Volvo Construction Equipment is responsible for introducing future technologies internally, via research projects and demonstrators. Speech recognition by SW is a standard interface technology today that will significantly increase its presence in the future. And it will be a very important interface towards autonomous machines in construction, but also to our regular machines and to our infrastructure. We therefore want to run a thesis work to investigate state-of-art for the technology, and to gain knowledge. Questions will be: • How good is the easy reachable technology – e.g. Cortana from Microsoft? • What is a good speech interface for a human user with a machine in a construction scenario? • What do we need to provide a feeling of trust and friendship when we use it? • How do you command an autonomous machine, and how does it feel like? Suitable background Computational engineer at master level, or equivalent education profile, with specific interest in audio, speech recognition, robotics, and human soft values. Description of thesis work The thesis work is for 1-2 students, takes place at Volvo CE’s facilities in Eskilstuna, and is paid according to Volvo’s standard routines. The thesis work will on a high level consist of: • State-of-art in speech recognition investigation • Design of a use-case scenario with an autonomous machine • Defining important soft features for a human-machine speech interface, especially about trust and ease of operation, by interviews, literature studies, and analytical work • Implementation of the use-case • Verification, evaluation, and demonstration of implemented work • Written report  
Company: Volvo CE, kontaktperson: Torbjörn Martinsson
Proposed:2017-10-23 
IDT supervisor:
,
Examinator: Mikael Ekström
Mikael Ekström
mikael.ekstrom@mdh.se, +46-21-101671
Title: Quality property ontology population through text/data mining 
Subject:
Level: Basic or Advanced (contact supervisor)
Description:Today, software is everywhere. Such a prevalence calls for the ability to develop high quality software at a faster pace. An important step in this direction requires that software engineers better comprehend which quality properties should be evaluated and how. However, the current body of knowledge on quality properties is huge, which makes it challenging for software engineers to be aware of the most suitable methods to evaluate a given property in their current context of work. To help organizing the knowledge on quality properties, an ontology of property models was proposed by Sentilles et al[1]. However, the ontology still needs to be populated with appropriate and accurate information. The goal of this thesis is to investigate how text mining, data mining and deep learning can be used to retrieve relevant data for the ontology. The thesis is suitable for 1 or 2 students. The scope of the work will be limited to a predefined subset of properties in line with the student's profile. The thesis work will on a high level consist of: - State-of-the-art and on text mining, data mining, deep learning, and quality properties to identify the most suitable technique to apply in the context of this work - Definition of a protocol on how the selected technique will be used (e.g., manual training, tagging, syntactic parsing, disambiguation, sources of information, features…) - Evaluation of the quality of the retrieved data - Classification of the retrieved data according to the property model ontology  
Proposed:2017-10-24 
IDT supervisor: Séverine Sentilles
severine.sentilles@mdh.se, +46-21-10 70 38
Examinator:

,
Title:Experimentations with graph generation algorithms 
Subject:
Level: Basic or Advanced (contact supervisor)
Description:Modern software applications provide more and more smart features, where for smart is intended any "intelligent" automated customisation to a given user profile. In order to do so, the applications have to be designed over a data repository, on which they create relationships between users' characteristics, try to derive corresponding interconnections among them, and hence adapt appropriately. In some cases, the future data to be accumulated in the repository is not available publicly, which means that artificial repositories have to be created in order to test and validate application features. In this respect, graph generation algorithms are a possible solution to create artificial data repositories based on certain data patterns. This thesis project is devoted to the experimentation of graph generation algorithms. Students are expected to: - propose a well defined experimentation protocol (how the experiment should be set-up to make the exercise repeatable); - propose a number of metrics to measure and compare the performances of the generation algorithms; - demonstrate/validate their proposal on a concrete case study. The thesis project will be done in the context of a decision support application.  
Proposed:2017-10-26 
IDT supervisor: Antonio Cicchetti
antonio.cicchetti@mdh.se, +46-21-151762
Examinator:

,
Title:Using Coverage Observers for Test Generation and Monitoring with UPPAAL SMC 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:UPPAAL SMC is a statistical model checker that can be used to generate random simulations of the system and can perform statistical analysis in order to check whether the system satisfies a certain property with a given degree of confidence. This enables UPPAAL SMC to analyze large systems, and recent studies [1, 2] have shown that the model checker can be used for random test case generation. In this thesis, the students are expected to develop a framework around the existing test case generation tool that will allow for: (i) automatic monitoring of coverage, and (ii) guiding the model checker towards maximizing coverage. The framework will be evaluated on an industrial use case, where observers will be used for automatic test case generation and measuring the achieved coverage for different types of coverage criteria. Reference: [1] Raluca Marinescu, Eduard Enoiu, Cristina Seceleanu, and Daniel Sundmark. Automatic Test Generation for Energy Consumption of Embedded Systems Modeled in EAST-ADL. International Conference on Software Testing, Verification and Validation Workshops, pages 69–76. IEEE, 2017. [2] Jonatan Larsson. Automatic Test Generation and Mutation Analysis using UPPAAL SMC. Bachelor of Science Thesis Report, MDH Diva, 2017.  
Proposed:2017-10-26 
IDT supervisor: Raluca Marinescu
raluca.marinescu@mdh.se,
Examinator:

,
Title:Robotic waste bin emptying in an office environment 
Subject:
Level: Advanced
Description:

Background

The Volvo Group is one of the world’s leading manufacturers of trucks, buses, construction equipment and marine and industrial engines under the leading brands Volvo, Renault Trucks, Mack, UD Trucks, Eicher, SDLG, Terex Trucks, Prevost, Nova Bus, UD Bus, Sunwin Bus and Volvo Penta. Volvo Group Trucks Operations encompasses the production of state-of-the-art products for the truck brands of the Volvo Group, as well as Volvo Group engines and transmissions, through an international world class industrial environment. With Volvo Group Trucks Operations you will be part of a global and diverse team of highly skilled professionals working with energy, passion and respect for the individual to become the world leader in sustainable transport solutions.

Background of thesis project

Robots acting in an environment where also humans are active becomes more and more close to reality. Not only in applications aimed for testing and demonstrations, but also closer to real, and productive usage. One of the big questions still, is how to interact with humans, giving the humans a natural feeling for a safe and efficient machine. At Volvo GTO, there are projects on-going having autonomous transporters in our manufacturing plants and in applications in public areas (e.g. Unicorn) with this focus. To get a broader and better understanding for this interaction we would also like to introduce moving robots also in the office environment. That’s the driver for this thesis proposal.

Thesis project task

Based on development and finding made at MDH in the Unicorn project, we would like to set up an office system with one or two truly autonomous robots for emptying dedicated trash bins. The task will consist of:
  • Design and realize the mechatronic design of robots and bins
  • Design and realize the mechatronic design of a small container where the robots can empty their payload
  • Design and implement the functionality emptying the bins on a regular basis
  • Implement a SLAM-based SW for the robots for navigate and manoeuvre in the office space
The work can be carried out where appropriate, but installation shall be made at an office space at the Volvo GTO headquarter in Gothenburg.

Suitable background

Robotics, Mechanical engineering, Automation & Mechatronics (M/Z/F).

Methodology

  • Understand the needs and prerequisites at Volvo office
  • Set up a theory for the human machine interaction
  • Study theories various sensor systems, e.g. lidar, stereo cameras
  • Study ROS system implementation
  • Implement and run demonstration

Language

Thesis is to be written in English. This thesis is suitable for two students  
Company: Volvo GTO, kontaktperson: Per-Lage Götvall
Proposed:2017-11-06 
IDT supervisor: Mikael Ekström
mikael.ekstrom@mdh.se, +46-21-101671
Examinator:

,
Title:Software Defined Networking for a Heterogeneous Network 
Subject: Computer Science
Level: Basic
Description:Software Defined Networking (SDN) is a key enabler to decrease the complexity of network management. SDN allows networks to employ a unified control plane in order to manage the network devices in a centralized manner. The control platform, known as the SDN controller, coordinates the traffic forwarding by applying forwarding rules to the network devices, remotely. The complexity of network management can be further decreased by applying network virtualization together with SDN. Network virtualization is a technique in which a physical network is partitioned into several virtual networks, known as slices, and each slice is controlled by an SDN controller. The performance of SDN networks (without network virtualization) have been studied in various applications. In this thesis, we are aiming at studying performance of an SDN network where the physical network is partitioned into a number of slices. We also consider a network as a combination of wired and wireless links. The main intention is to perform several experiments on a real hardware to obtain various performance parameters, such as latency, jitter and throughput. This thesis work is a continuation of previous thesis work (Network Virtualization in Multi-Hop Heterogeneous Architectures) and the student will be provided with the architecture from that.  
Proposed:2017-11-07 
IDT supervisor: Mohammad Ashjaei
mohammad.ashjaei@mdh.se, +46-21-151772
Examinator: Saad Mubeen
Saad Mubeen
saad.mubeen@mdh.se, +4621103191
IDT supervisor:
,
Title:Use of service choreography principles in the context of systems of systems with focus on safety and security - a mapping study 
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Systems of systems (SoS) are built as a collection of systems that share their resources and capabilities in order to achieve new functionalities, provide better performance or higher level of efficiency, when compared to traditional systems. The major challenges in such systems are their complexity and providing analysis of properties of SoS, given that system behaviours may change. Ensuring safety of such systems is becoming a challenging task due to their dynamic nature, as well as the possible impact of security threats on safety. In order to address these challenges adequately, one has to learn about interaction of these properties and ways how they can be handled. Service choreography defines ways of service composition between several services through some interaction protocol. The main idea is that at run-time, each service in choreography executes its part, and contributes to the overall purpose of the large service. The focus of this thesis is to explore possibilities to use service choreography principles in context of SoS using a mapping study approach. The thesis outcomes should describe the elements of service choreographies and map them to the appropriate element of SoS.  
Proposed:2017-11-20 
IDT supervisor: Aida Causevic
aida.delic@mdh.se, +46-21-107011
Examinator: Hans Hansson
Hans Hansson
hans.hansson@mdh.se, +46 21 103163
Title:Design and Evaluate a Multi-Radio Sensor Network 
Subject:
Level: Basic
Description:Description: This thesis work is defined in the scope of the ESS-H research profile (Embedded Sensor Systems for Health Research Profile, http://www.es.mdh.se/projects/324-ESS_H) and ecare@home project (https://ecareathome.se/). These projects aim at providing remote health monitoring of patient through wireless medium. Wireless Sensors Networks (WSNs) are considered as one of the key Internet of Things (IoT) technologies, and are widely used in various application areas such as environmental and structural monitoring systems, industrial automation, healthcare systems, traffic management and logistics, and public safety. An efficient IoT-enabled healthcare system aims to provide a remote health monitoring of a patient health status in real-time, the prevention of critical patient conditions, life quality improvement of the elderly through the smart environment, medical and drugs' database administration, and wellbeing services. The smart IoT sensors for healthcare enable accurately measuring, monitoring and analyzing a variety of vital health status indicators together with environmental parameters. Sensor readings are then collected and transferred to the end-devices or to a Cloud server. The SensorTag device contains 10 low-power MEMS sensors in a small package. This device includes following sensors: light, digital microphone, magnetic sensor, humidity, pressure, accelerometer, gyroscope, magnetometer, object temperature, and ambient temperature. This multi-radio device supports both BLE and 802.15.4 radios. It is based on the CC2650 wireless MCU, offering 75% lower power consumption than typical Bluetooth low energy products. This allows the SensorTag to be battery powered, and offer years of battery lifetime from a single coin cell battery. This is the main benefit of SensorTag that makes it a potential candidate for IoT applications. This device has the possibility of running IoT technologies such as 6LoWPAN and 6TiSCH. Problem statement: It is common to use physiological sensors for health monitoring. Studies have revealed that it is mandatory to consider both environmental sensors and physiological sensors. The problem is that there is no health monitoring system consisting of both environmental sensors and physiological sensors. Main outcome: The main outcome of this Thesis is to connect Sensortags to the database, and visual the real-time data collection in a GUI. This Thesis will show the functionality of SensorTag mult-radio device in terms of packet reception ratio and RSSI fluctuation. Tasks: collect sensor parameters, RSSI, number of transmitted packets, and number of lost packets show the results during run-time in a GUI setup a repeatable testbed for experiments study impact of mobility study impact of interference study impact of number of nodes study impact of sampling frequency 
Proposed:2018-01-04 
IDT supervisor: Hossein Fotouhi
hossein.fotouhi@mdh.se,
Examinator: Mats Björkman
Mats Björkman
mats.bjorkman@mdh.se, +46-21-107037
Title:Efficient time-series storage for enabling IoT device intelligence 
Subject: Computer Science
Level: Basic
Description:Background: During the last few years there has been a staggering pace of development, attention, and increasing maturity, of cloud enabled services. Even in the popular press there is every week articles on how Artificial Intelligence will change many aspects of our everyday life. Intelligent devices are also often mentioned in the same context, the so called Internet-of-things (IoT) devices. However, to develop intelligent systems and devices there is often need for huge amounts of data from many different sources. For example, Facebook has billions of users which each day perform actions which is one important driving force for steering future development of the services Facebook provides. The access to this vast amount of data, and the capabilities of transferring this data to the cloud for processing and storage, is not as easy for industrial systems. Often industrial systems are in remote locations with less than ideal connectivity options. Still, providers of industry systems also want to develop the same kind of intelligence for their systems and devices. One big difference though, is that industry systems typically wants to collect data with millisecond sampling rates, or even nanosecond rates, while Facebook and many others work with a second as the lowest resolution. One important source of data for developing machine intelligence is time-series data, and there are already many existing open-source time-series databases (TSDB) focusing on this particular problem. One reason for having specific databases for time-series data is due to the need to compress the data in order to be able to increase the amount of data stored and/or the frequency with which data is stored; traditional databases are simply too in-efficient in this respect. Typically a time-series consists of a time stamp (8bytes) and a value (typically 8 bytes), i.e., 16 bytes per sample. Current open-source TSDBs can on a typical case reduce the storage need from 16 bytes per sample to 1.3 bytes per sample, i.e., more than a 10 time reduction in storage need. However, there is one caveat with this. Many of the open-source TSDBs are focusing on time-series data generated from machine logs in large datacenters, where the sampling frequency is at most 1 second. Consequently, their TSDB implementations are optimized for that particular use-case and cannot directly be used in an industry context. (Facebook has an open-source TSDB here: https://github.com/facebookincubator/beringei, an extensive list of existing of TSDBs can be found here: https://misfra.me/2016/04/09/tsdb-list/ ) This thesis project will be conducted at ABB Force Measurement in Västerås, Sweden. At Force Measurement we develop, produce, and sell advanced measurement equipment and control systems for use in process industry. Several of our products are used in cold-rolling mills, hot-rolling mills, paper production lines, and even cruise ships. This thesis project could possibly become one important part of the puzzle of enabling further intelligence in our products. Goals, problems: Benchmarking of a select set of existing open-source time-series databases to determine their suitability for using in an industrial context. Based on the benchmarking and study of features of the set of TSDB solutions there will most likely be a gap compared to what is needed by industry. The second phase of the project is concerned with proposing changes to one of the open-source TSDBs, implementing the proposed changes, and finally benchmarking the implementation to see if the goals have been met. The proposed solution also needs to consider how the data is gathered in an industry site and transferred to the cloud. To be more specific, there will most likely be a need for an in-memory TSDB, with possibilities for persisting it to disk. Expected outcomes: Thesis report documenting: - Result of benchmarking and selection criteria used in evaluating the set of open-source TSDBs - Proposed design changes to at least one existing TSDB Prototype implementation: - Extend one open-source TSDB to become useful in the industrial context - If time permits, also look into how this data can be transferred to the cloud for further processing and storage 
Company: ABB Force Management, kontaktperson: Markus Lindgren
Proposed:2018-01-29 
IDT supervisor:Dag Nyström
dag.nystrom@mdh.se, +46-21-107042
IDT supervisor:
,
Title:Data-driven actors modelling for road transportation  
Subject: Computer Science
Level: Basic or Advanced (contact supervisor)
Description:Road transportation is a complex system and poses challenges in safe transportation due to dynamic environments changes, where multiple actors are involved. It has been reported that 90% of traffic crashes are caused due to driver risky behaviour. To improve the traffic safety it is not enough to analyse only the risky behaviour of the driver but also a collective behaviour analysis of several actors such as two wheelers, pedestrians i.e., the road users. It also requires analysing interaction between these users and safety related measures. To understand the dynamic of road users and road safety the work focuses on four actors that are: 1. Vehicle driver 2. Motorcyclist 3. Bicyclist 4. Pedestrian The project aims for a systemic literature review to understand and identify parameters, metrics and indices for all actors. The review will also cover the interaction between different actors and other measures for both safe road usage and road crashes. Vehicle driving requires a high degree of concentration and it includes various dynamic and complex activities. Through cognition process, a person understands the information about an external object or phenomenon in response to the influence of acquired knowledge, memory, and experience. Among many actions driving depends on the vehicle's state e.g., the speed of the vehicle, lateral position. In addition, driving behaviour signals are obtained from vehicular data, which represent "longitudinal control" or "lateral-control" action. Hence the second aim of this project is to investigate and implement models for different driving events using driving behaviour signals. Here machine-learning algorithms should include in the model development to find patterns for different driving behaviours. The project work is subdivided as follows: 1. Literature study and state-of-the-art This task requires a systematic literature review to identify the parameters, metrics and indices for all actors that can be used modelling actors' behaviour in road transport environment. Student requires presenting an analytical summary of the state-of-the-art based on the literature study. 2. Implementation This task only involves analysing the drivers' vehicular data of various driving events for a given dataset. Student also requires developing an approach using machine-learning algorithms for detection of different driving patterns based on the driving events and vehicular dataset. 3. Evaluation Student should evaluate the proposed approach and learning algorithm for detecting driving patterns. 4. Report writing It is expected that student provide a report as a completion of the project work. Report consists of background, problem formulation, state-of-the-art, methods, evaluation, and discussion.  
Proposed:2018-02-19 
IDT supervisor: Mobyen Uddin Ahmed
mobyen.ahmed@mdh.se, +46-21-107369
Examinator: Shahina Begum
Shahina Begum
shahina.begum@mdh.se, +46-21-107370
IDT supervisor:
,
Title:Övergången från traditionella nätverk till software-defined access 
Subject:
Level: Basic
Description: 
Proposed:2018-01-07 
IDT supervisor:
,
IDT supervisor:
,
Title:CONTROLLING CACHE PARTITION SIZES TO OPTIMIZE APPLICATION RELIABILITY 
Subject:
Level: Basic
Description:Single-chip multiprocessors(CMP) are commonly used today due their better overall performance compared to uniprocessor architectures. While CMPs are beneficial in terms of performance, CMPs also comes with its own share of problems. One of these problems stems from how shared cache memory is organized and utilized in a CMP alongside employed eviction policies, this problem is known as cache contention. Cache contention occurs as consequence of unconstrained usage of shared cache memory. As cores in a CMP all have equal access to shared cache memory, they implicitly compete for available capacity. If the issue of cache contention is left unmanaged, process performance can be negatively effected as the Least Recently Used(LRU) eviction policy might evict stored data indiscriminately. This performance impact becomes more troublesome if executing processes are time-critical, since execution time is affected due to eventual cache misses. A solution to solve the cache contention problem is use cache partitioning, which splits cache memory into different independent partitions. This enables processes to execute in isolation from each other, therefore minimizing risks of inter-process interference. In this study we intend to employ cache-partitioning combined with cache-coloring rules to try and solve the cache contention problem. As result of this we intend to investigate whether process reliability can be increased by controlling cache partitions.  
Proposed:2018-03-18 
IDT supervisor: Jakob Danielsson
jakob.danielsson@mdh.se,
Examinator: Masoud Daneshtalab
Masoud Daneshtalab
masoud.daneshtalab@mdh.se,
IDT supervisor:
,
Title:EXPLORING ISOLATION OF HARDWARE RESOURCES USING HYPERVISORS 
Subject:
Level: Advanced
Description:Virtualization has become an important aspect of the IT world by improving scalability and workloads which can lead to benefits such as optimizing resources and simplifying management of processes [1]. In a virtualized environment, a hypervisor, also known as a virtual machine monitor (VMM), acts as a platform for running virtual machines and allows for the consolidation of physical computing resources such as CPUs, memory and storage [2]. A VMM is a piece of software that virtualizes the hardware and enables isolation between the operating system (OS) processes [3]. It handles resource and memory allocation of the host for all the virtual machines, ensuring that they cannot interrupt each other. However, when multiple virtual machines concurrently run on the same hardware, they share the same available resources of physical host and it may lead in the high level of unpredictability of performance. Predictability is especially important in time sensitive systems such as avionics. The avionic industry enforces the ARINC-653 standard, which is the software interface for avionic processes. The standard furthermore specifies how an operating system should consolidate shared resources, which is by enforcing total isolation of a process and its hardware resources during the process timeslice. Methods for achieving different degrees of isolation include partitioning of different hardware resources, scheduling processes to enforce I/O isolation and also using virtualized environments to limit the CPU resource sharing. 
Proposed:2018-03-18 
IDT supervisor: Jakob Danielsson
jakob.danielsson@mdh.se,
Examinator: Moris Behnam
Moris Behnam
moris.behnam@mdh.se, 021-107094
IDT supervisor:
,
Title:Load Balancing Workloads Through DRAM Bank Partitioning 
Subject:
Level: Basic
Description:Reliability has long been seen as one of the major bottlenecks when incorporating multi-core systems in embedded environments. The many different shared hardware resources can cause unwanted memory contention, which is caused by different processes competing for the same memory. However, measures such as partitioning can be taken in order to guarantee exclusive ownership of shared hardware resources. Hardware partitioning is a technique that partitions hardware on the chip without any software support and can be both robust and effective. Hardware partitioning is however not always applicable to general Commercial Off the Shelf (CoTS) hardware platforms such as Intel, since there may be a lack of support for such functionalities. Another technique that can be used, is the software based partitioning method page-coloring. Page-coloring isolates hardware resources using software techniques. Itis a much more general approach since it is applicable to all systems using a feasible Linux kernel. The page-coloring technique is applicable to both DRAM and cache memory. In this work, we will focus on investigating the page-coloring technique for DRAM memories, called DRAM bank partitioning. DRAM bank partitioning is used for over-writing the memory translations from the virtual to the physical memory, in order to reduce memory contention on the DRAM memory. 
Proposed:2018-03-18 
IDT supervisor: Jakob Danielsson
jakob.danielsson@mdh.se,
Examinator: Masoud Daneshtalab
Masoud Daneshtalab
masoud.daneshtalab@mdh.se,
IDT supervisor:
,


  • Mälardalen University |
  • Box 883 |
  • 721 23 Västerås/Eskilstuna |
  • 021-101300, 016-153600 |
  • webmaster |
  • Latest update: 2018.03.15