Optimization-Based Navigation for the DARPA Grand Challenge

From MurrayWiki
Jump to: navigation, search


Dmitriy Kogan and Richard M. Murray
Submitted, 2006 Conference on Decision and Control

This research addresses the path planning problem with a nonlinear optimization method running in real time. An optimization problem is continually solved to find a time-optimal, dynamically feasible trajectory from the vehicle’s position to some receding horizon ahead (20m-70m forward). The locally optimal numerical solver optimizes both the spatial and temporal components of the trajectory simultaneously, and feeds its output to a trajectory-following controller. The method has been implemented and tested on a modified Ford E350 van. Using one stereo pair and four LADAR units as terrain sensors, the vehicle was able to consistently traverse a 2 mile obstacle course at the DGC qualifying event. At the main DGC event, the vehicle drove 8 autonomous miles through the Nevada desert before experiencing non-planning issues. During this time, the planning system generated a plan 4.28 times per second on average. This execution speed, coupled with a feedback-based trajectory-following controller was shown to be adequate at providing smooth and reliable obstacle avoidance even on complicated terrain.