PhD student in fast real-time optimization and certification for MPC and autonomous systems with placement at the Division of Automatic control at Linköping University.
Your work assignments
The research conducted within the area of automatic control at Linköping University includes subareas such as learning control, optimization, robotics and autonomous systems, and sensor fusion. We are now announcing a PhD student position targeting research within the next generation of fast real-time certified optimization algorithms for Model Predictive Control (MPC), financed by Sweden’s by far largest individual research program, the Wallenberg AI, Autonomous Systems and Software Program (WASP). The planned research project is a direct continuation of a now finished research project where the, now graduated, PhD student recently received the prestigious IFAC World Congress Young Author Prize 2023 for his work in the project. In the previous project, we, e.g., showed how to a priori bound the computation time of MPC down to CPU-cycle accuracy, which has never been done before. Furthermore, online optimization software was developed that was shown to be among the fastest available in the world for MPC, capable of running on board in real-time on the miniature quadcopter Crazyflie.
There are several goals with the proposed project. First, to improve the performance of our QP algorithm even further by incorporating high-efficiency problem structure-exploiting algorithms and include these improvements into a truly state-of-the-art optimization software for MPC. Second, to extend our real-time certification framework to also include the structure-exploiting algorithms. Third, to optimize the online solver performance using information from the certification framework, a direction which has never been explored before. The work includes development of theory, algorithms, and software. It is highly desirable to demonstrate the result on relevant physical systems, e.g., on drones or autonomous trucks via our industrial partners. In summary, this is a challenging project with very high potential for the ambitious student.
As a WASP-funded PhD student, you get access to a unique graduate school offering courses given by leading researchers within AI, autonomous systems and software, exclusive to WASP PhD students. Furthermore, regular social events give you the possibility to build a large network within Sweden, while regular study trips abroad and the possibilities to apply for a semester abroad gives you the possibility to build an international network.