Detail of the student project

Topic:Metriky hodnotící predikci trajektorie vozidla
Department:Katedra kybernetiky
Supervisor:Ing. Vojtěch Franc, Ph.D.
Announce as:Diplomová práce, Semestrální projekt
Description:In the word of autonomous cars, there is a growing need for reliable trajectory prediction of vehicles based on the driver’s behavior and interaction between drivers. While wide range of methods was proposed in recent years, the metric evaluating their performance is often not thoroughly thought through. The aim of this project is to develop metric, which computes similarity of two trajectories, taking in account heuristic knowledge about different road situations. For example, imagine a scenario where you predict a trajectory of a target car 200 meters ahead. The required precision on longitude and latitude is on a different scale, with lower error tolerance on latitude. However this might change when considering ego car point of view – imagine a car approaching a congestion, the importance of a correct prediction on longitude increases dramatically.
Bibliography:- Stéphanie Lefèvre, Dizan Vasquez, Christian Laugier. A survey on motion prediction and risk
assessment for intelligent vehicles. ROBOMECH Journal, Springer, 2014

- Quehl, Jannik et al., How good is my prediction? Finding a similarity measure for trajectory prediction
evaluation. 2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC)
(2017): 1-6.
Responsible person: Petr Pošík