Autonomy Talks - Erdem Bıyık: Learning how to Route Autonomous Vehicles on Shared Roads

Autonomy Talks - 01/11/2021

Speaker: Erdem Bıyık, Stanford University

Title: Learning how to Route Autonomous Vehicles on Shared Roads

Abstract: Road congestion induces significant costs across the world, and road network disturbances, such as traffic accidents, can cause highly congested traffic patterns. If a planner had control over the routing of all vehicles in the network, they could easily reverse this effect. In this talk, I will consider a more realistic scenario where a planner controls autonomous cars, which are a fraction of all present cars. I will present the problem as a dynamic routing game, in which the route choices of autonomous cars can be controlled and the human drivers react selfishly and dynamically. As the problem is prohibitively large, I will go over a solution based on deep reinforcement learning to control the autonomous vehicles. This learned policy indirectly influences human drivers to route themselves in such a way that minimizes congestion on the network. To gauge the effectiveness of the learned policies, I will also establish theoretical results characterizing equilibria, and will compare them to the deep reinforcement learning based solution.

Visit the official webpage for more details: https://idsc.ethz.ch/research-frazzoli/autonomy-talks.html

Видео Autonomy Talks - Erdem Bıyık: Learning how to Route Autonomous Vehicles on Shared Roads канала Autonomy Talks