Improving on our ICRA paper, this new approach enables real-time probabilistic object tracking. We currently use this method to track all dynamic obstacles seen by our autonomous vehicle, in real-time, with significantly improved accuracy compared to our previous Kalman-filter based approach. Tracking-based semi-supervised learning, as originally presented at RSS, was an offline algorithm. This is fine in some contexts, but ideally a user could provide new hand-labeled training examples online, as the system runs, without retraining from scratch.
Autonomous Vehicle Control at the Limits of Handling | Dynamic Design Lab
As autonomous vehicles enter public roads, they should be capable of using all of the vehicle's performance capability, if necessary, to avoid collisions. This dissertation focuses on facilitating collision avoidance for autonomous vehicles by enabling safe vehicle operation up to the handling limits. The new control approaches first rely on a standard paradigm for autonomous vehicles that divides vehicle control into trajectory generation and trajectory tracking. A trajectory generation approach calculates emergency lane change trajectories, defined in terms of path curvature, that allows an autonomous vehicle to perform emergency lane changes up to its handling limits. Analysis also provides insights into when and to what extent a vehicle should brake and turn during an emergency lane change to maximize the number of situations in which a collision can be avoided.
Autonomous Vehicle Control at the Limits of Handling
To browse Academia. Skip to main content. Log In Sign Up. Autonomous Vehicles 4, Followers.
Many road accidents are caused by the inability of drivers to control a vehicle at its friction limits, yet racecar drivers routinely operate a vehicle at the limits of handling without losing control. If autonomous vehicles or driver assistance systems had capabilities similar to those of racecar drivers, many fatal accidents could be avoided. To advance this goal, an autonomous racing controller was designed and tested to understand how to track a path at the friction limits. The controller structure was inspired by how racecar drivers break down their task into i finding a desired path, and ii tracking this desired path at the limits. Separating the problem in this way instead of integrating both path planning and path tracking into one problem results in an intuitive structure that is easy to analyze.