Content, progress and pedagogy of the module

The module is based on knowledge achieved from the courses in Non-linear Control and Reliability and System Identification and Diagnosis.

Learning objectives

Knowledge

• Have knowledge about the components and control strategies for electronic systems including machine learning models.
• Have knowledge on the principles of system identification to identify models of dynamic systems from experimental data and knowledge on how to use system identification techniques to design control systems.
• Have knowledge about advanced closed-loop control solutions, for instance model predictive control (MPC), shallow and deep neural networks, or other machine learning algorithms.
• Have insight in how to apply reliable electronics and control algorithms for reliable operational performance.

Skills

• Be able to design a complete embedded system operating with industrial electronics or design a learning control system for industrial settings.
• Be able to design hardware and software which successfully can stabilize complicated processes during uncertain and varying conditions applying artificial intelligence.
• Be able to design efficient controllers for disturbance rejections or machine learning frameworks to design end-to-end systems.
• Be able to implement hardware applications related to industrial electronics or integrate deep neural networks in a control system.
• Be able to design control systems for robotic manipulators, mobile robots, autonomous underwater vehicles (AUV), unmanned aerial vehicles (UAV) or other automated systems.

Competences

• Be able to design models for complex control systems for industrial electronics
• Independently analyse complex control system problems of electronic systems and develop control strategies
• Independently understand the concepts of predictive and adaptive control strategies and the application of artificial intelligence in industrial process control
• Be able to control the working and development process within the project theme, and be able to develop feasible solutions within optimisation, control, and/or diagnostic within advanced control of industrial electronics
• Independently be able to continue own development in competence and specialisation
• Be able to follow state-of-the-art literature within the topics of machine learning, optimisation, predictive and data-driven control, and advanced control of industrial systems
• Be able to work collaboratively with others, including engineers, scientists, and technicians, to design and implement data-driven control systems

Type of instruction

The project work must be documented by a scientific paper (max. 8 pages) accompanied by a project summary report. The project summary report should elaborate the project details and conclusions. The maximum length of the summaryreport (report without appendices) is 50 pages. For more information see semester description in Moodle.

The scientific paper will be presented at a conference arranged within the Department of Energy, prior to the project examination.

The project must include an application that includes a power electronic converter, a power source and embedded system or be based on an automated system such as a mobile robot, UAV, AUV or similar. The operating principles for the system must be described and a control problem is formulated including key specifications. A dynamic simulation model is made taking the relevant dynamics into account. Different advanced control methods like predictive, adaptive, and robust control strategies or machine learning algorithms such as shallow, deep neural networks, or other types of artificial intelligence are designed to control the system and evaluated on basis of the simulation model. At least one method is selected for practical implementation in a real system incorporating hardware and software and a real-time digital advanced control system based on a digital signal processor or a microcontroller. Finally, the whole system is tested and the developed control strategies are evaluated with the purpose of verifying the hypothesis, as well as drawing conclusions based on the achieved result.

If there are special technical or scientific documentation requirements, the student documents the project work in a project report, which can be prepared individually or in a group within the project theme. However, the student’s special preferences for the semester must be approved by the Study Board in advance.

Extent and expected workload

Since it is a 20 ECTS project module, the work load is expected to be 600 hours for the student.

Exam

Prerequisite for enrollment for the exam

• It is a pre-condition that the student has submitted a scientific paper and presented the scientific paper at an internal conference prior to the project examination.
• In case of a re-exam, the student will have to present the scientific paper in front of a committee made up of the supervisor and at least one internal adjudicator.

Exams