- The introductory session for the registered students will be on 23.10.2020 at 9:00 over tum-conf Zoom service. Attendance for all the registered students is mandatory.
- This lab course does not involve any regular weekly sessions. However the tasks will be assigned over moodle for each group and the deliverables will be handled over the gitlab repositories. Irregular meetings will be conducted to make sure about the progress of each group.
- Interested applicants must send an email of interest to "shafaei (@) in.tum.de" (set the subject of your email to: DRPPAD20-21) with a copy of their CV, transcripts of results and a brief explanation of their motivation, latest by the last day of the matching period.
- If you're planning to set this lab course as your first priority in the matching system, please write us about it in your email of interest.
The resource planner computes an architectural proposal based on predefined optimization goals for an autonomous driving vehicle by collecting system properties and AI-based application requirements. In case of dynamic reconfiguration, e.g. when the vehicle changes from the highway to city traffic or when drives in a tunnel, the resource planner can kill running individual algorithms based on the context as well as launching other algorithms while ensuring the whole process in terms of functional safety. In this lab course, It is planned to implement a context-based resource planner while considering the computation, communication ans safety aspects. Different groups of students will be assigned to work on different modules of this resource planner, from perception unit to the core of the resouce planner itself.
There will be 5 main modules of "Architecture Design", "Application Development", "Context Management and Monitoring", "Resource Planner Development" and a bonus project on "Fail Operational" for this lab course. Students will be given the chance to work either individually or as a group of two. Each group will be focused mainly on the predefined modules while maintaning the collaboration with other modules/groups to deliver the expected architechture of the resource planning platform at the end. The grading will be solely based on the quality of the deliverables of each group and the final evaluation of the developed architecture.