Home | People | Research | Publications| Links | News | Contact   

Control of High-Speed Autonomous Wheeled Vehicles

The need for reducing the risk for human lives while operating in hazardous or hostile environments has led to the development of unmanned and often autonomous vehicles in both commercial and military applications. A class of vehicles we envision to be completely automated in the future are ground wheeled vehicles (Fig.1a) that operate in hostile off-road environments (e.g., battlefields). A typical mission would be to drive the vehicle from point A to point B, avoid any obstacles, while minimizing the exposure to danger; see Fig.~1b.


                            Fig. 1a                                                                 Fig,1b

In general, minimization of the exposure to danger involves driving through a trajectory in minimum time or maximum average velocity. The fastest off-road vehicles can be found in open country rally-cross races. Racing such a vehicle requires a great deal of practice and skill. In this research our plan is to develop controllers that operate as ``intelligent rally/race-drivers,'' which imitate the driving techniques of race drivers. These driving techniques differ significantly from the techniques used by everyday drivers; they also are significantly different than driving techniques used by expert drivers of closed circuit, track or F1 races. It is evident, even to inexperienced drivers, that race driving involves operating the vehicle at the limits of its handling capacity.

Race drivers operate at the adhesion limits of the tires with very slim safety margins. For instance, oversteer is an unstable condition. The car is trying to spin and the spin must be stopped before one can worry about regaining directional control. Nonetheless, very often race drivers intentionally induce controlled skidding through corners to optimize their trajectory and prepare the vehicle for the next maneuver. This project also investigates the effects of slipping and sliding on wheeled mobile vehicles. In a clear departure from most of the research reported in the literature thus far, our approach will not treat slipping/sliding as a ``nuisance'' or ``uncertainty'' to be reduced or eliminated completely. Instead, we will incorporate the possibility of slipping and sliding as part of the acceptable vehicle operating regime. The control algorithms may even initiate slipping/sliding in a controlled manner if this action will benefit the overall operational objective.


This project is sponsored by the US Army Research Office.

Selected Publications


Home | People | Research | Publications| Links | Contact