Accessibility navigation


Multi-vehicle planning using RRT-connect

Kala, R. and Warwick, K. (2011) Multi-vehicle planning using RRT-connect. Paladyn. Journal of Behavioral Robotics, 2 (3). pp. 134-144. ISSN 2081-4836

Full text not archived in this repository.

To link to this item DOI: 10.2478/s13230-012-0004-5

Abstract/Summary

The problem of planning multiple vehicles deals with the design of an effective algorithm that can cause multiple autonomous vehicles on the road to communicate and generate a collaborative optimal travel plan. Our modelling of the problem considers vehicles to vary greatly in terms of both size and speed, which makes it suboptimal to have a faster vehicle follow a slower vehicle or for vehicles to drive with predefined speed lanes. It is essential to have a fast planning algorithm whilst still being probabilistically complete. The Rapidly Exploring Random Trees (RRT) algorithm developed and reported on here uses a problem specific coordination axis, a local optimization algorithm, priority based coordination, and a module for deciding travel speeds. Vehicles are assumed to remain in their current relative position laterally on the road unless otherwise instructed. Experimental results presented here show regular driving behaviours, namely vehicle following, overtaking, and complex obstacle avoidance. The ability to showcase complex behaviours in the absence of speed lanes is characteristic of the solution developed.

Item Type:Article
Refereed:Yes
Divisions:Faculty of Science
ID Code:32262
Uncontrolled Keywords:Autonomous vehicles, rapidly exploring random trees, RRT-Connect, multi-robot path planning, coordination, robocars, planning, intelligent vehicles.
Publisher:Versita; Springer

University Staff: Request a correction | Centaur Editors: Update this record

Page navigation