30 credits – Efficient vehicle motion planning on GPU i Sodertalje

Scania is now undergoing a transformation from being a supplier of trucks, buses and engines to a supplier of complete and sustainable transport solutions.

30 credits – Efficient vehicle motion planning on GPU

Ingress:
The Autonomous Vehicle Motion Planning and Control group at Scania develops the key components for path planning, speed planning and control for autonomous vehicles. A thesis project at the group and Scania is a great opportunity to work at the forefront of autonomous vehicle development, and an excellent way of making contacts for your future working life.

Background:
The vehicle industry is moving towards more and more autonomous vehicles, and the vehicles are becoming equipped with more and more capable sensors and control units. Motion planning is an essential part of the autonomous system, and aims to calculate trajectories for the vehicle to follow. These trajectories must meet the dynamical constraints of the vehicle, as well as the constraints from the surroundings. They must also be collision-free, and checked against both dynamic and static obstacles. This trajectory generation is computationally expensive, and often result in a very large number of possible trajectories, each of which must be collision checked. Therefore the calculations may, depending on the algorithm, be very suitable for parallel processing on GPUs.

Target:
To develop an algorithm for motion planning and collision checking that is suitable for massively parallel implementation.

Assignment:
To propose an algorithm for calculating collision-free trajectories which meet the dynamical constraints of the vehicle, as well as the constraints from the surroundings. The algorithm shall take advantage of the processing power of modern GPUs. It shall be implemented and evaluated in simulation, and preferably also in a real vehicle.

Education:
Master (civilingenjör) in electrical engineering, physics, mechatronics, computer science or similar. Preferably with specialization in control theory, robotics or computer science.
Number of students: 2
Start date: January 2019
Estimated time needed: 20 weeks

Contact persons and supervisors:
Robin Andersson, Autonomous Vehicle Motion Planning and Control, 08 – 553 702 65,

• Title: 30 credits – Efficient vehicle motion planning on GPU
• Business area: Research and Development
• Country: Sweden
• City: Sodertalje
• Last application date: 2018-12-06
• Job Id: 20183681