30 credits –Trajectory Prediction of Surrounding Vehicles and the Effect of Varying Host Vehicle Maneuvers
Scania is now undergoing a transformation from being a supplier of trucks, buses and engines to a supplier of complete and sustainable transport solutions.
A thesis project at Scania is an excellent way of making contacts for your future working life. Many of our current employees started their career with a thesis project.
Highly automated driving systems are required to make robust decisions in many complex driving environments, such as urban intersections and settings with a high level of interaction between vehicles. In order to make as informed and safe decisions as possible, it is necessary for the system to be able to predict the future maneuvers and positions of other traffic agents.
In many driving environments, the actions of one vehicle can affect the future actions of of other vehicles. Therefore, when choosing which action the host vehicle should take, it is of interest for the decision-making and planning modules to know how the different hypothetical maneuvers of the host vehicle will change the predicted trajectories of the surrounding vehicles. This information can then be used to choose optimal behavior that reduces risk to the host vehicle as well as other traffic participants.
To evaluate and develop a prediction framework that is capable of adapting its predictions for the same traffic scenario based on different hypotheses for the host vehicle action. This is to be done using an existing prediction framework as a basis, with the possibility to conduct a comparison with an alternative method later.
The assignment can be broken down into several tasks: 1. Explore prediction methods that are capable of multi-agent prediction in traffic scenarios with a high level of interaction; 2. Evaluate the capability of an existing prediction framework to adapt its predictions to varying host vehicle actions in the same traffic scenario. 3. Build upon or modify the existing method further in case the evaluation shows an inability to adapt. 4. Implement one of the found alternative methods and conduct a comparison between the two, either in simulation or with real-world data on a research platform.
Master (civilingenjör) in computer science, electrical engineering, or applied mathematics, preferably with specialization in artificial intelligence algorithms and a knowledge of Deep Learning.
Number of students: 1
Start date: January 2020
Estimated time needed: 20 weeks
Contact persons and supervisors:
Joonatan Mänttäri, Ph.D Student in Autonomous Systems, KTH , firstname.lastname@example.org
Christoffer Norén, Senior Engineer in Autonomous Motion, email@example.com , 08 – 553 811 48
Enclose CV, cover letter and transcript of records.
Sista sökdatum: 2019-12-20
Jobb Id: 20194285