Nytt

PhD student group meeting with Gurobi workshop 2019.

Practical information:

The second meeting for the PhD student group is scheduled for October 24 to 25th 2019 and will take place in connection to the conference SOAK in Nyköping. For information regarding SOAK, see https://soaf.se/soak/.

The meeting for the PhD student group will start at late afternoon on Thursday October 24th and on the first evening there will be some social activity as well as a dinner. On the second day of the meeting Gurobi https://www.gurobi.com/ will have a workshop until lunch. The registration for the conference and the meeting is at https://www.simplesignup.se/event/154928-soak2019. Thanks to SOAF and Gurobi there is no fee for the meeting. This includes accomodation, dinner, and participation in the Gurobi workshop.

To participate in the meeting for the PhD student group a membership in SOAF is required. If you are not sure if you’re already a member, check if your organisation is a member. Membership is free for PhD students.  Contact the chairman of SOAF at ordforande@soaf.se for more information about becoming a member.

For any questions about the meeting, please contact bjorn.moren@liu.se.

Presentation at the Gurobi workshop:

Among other technical talks, we will be hearing from a Gurobi user and their PhD colleague. Here is an abstract of their talk:

Applications for MIQP programming in control of self-driving tractor-trailer vehicles (a research project collaboration between Linköping University and Scania CV.)

Over the past decades, autonomous transport solutions and advanced driver assistance systems have received rapidly increased interest as the technology in these areas has advanced. In this talk, I will present two applications where MIQP programming has the potential of pushing the performance of these systems. The first application is Platooning where two of more vehicles are driving in a platoon with short inter-vehicular distances to optimize fuel consumption. The Platooning problem is to control the motor torque, the breaking forces and to select engaged gear such that the overall fuel consumption is minimized. The second application is the control of a reversing tractor-trailer vehicle. Since trailers are often changed during daily operation, optical sensors mounted on the tractor have been proposed to solve the online estimation problem of the trailer states (position, orientation and joint angles). Since these sensors typically have a limited field of view (FOV), it is important that the vehicle is controlled such that its joint angles remain in the region where high-accuracy state estimates can be reliably computed online. This region is highly non-convex but can be accurately modelled as a union of convex polygons, and the control task can be formulated as a MIQP problem. In both applications, Gurobi is used as MIQP solver and simulation results show that the proposed solutions have the potential of being deployed in practice.