Robust lateral control of heavy vehicles on automated highways

This dissertation deals with the problem of lateral control of automated heavy vehicles (tractor semitrailers). Lateral control forms an integral part of any Automated Highway System (AHS) architecture as it provides lane following and lane changing capabilities to vehicles in the AHS. Over the last couple of decades, AHS technologies have attracted increasing attention from researchers throughout the world. Particularly, AHS technologies for heavy vehicles (HVs) have been pursued vigorously. HV automation has several attractive features which include reduction in fuel cost and driver stress, lower automation vs vehicle cost as compared to passenger vehicles.; The dissertation begins with an analysis of lateral dynamics of tractor semitrailers. Important parameters that affect the lateral behavior, which is captured in the linearized model, of these heavy vehicles include longitudinal speed, road adhesion coefficient, mass and position of cargo load in the trailer amongst others.; Based on these analyses, two types of steering controllers (one degree-of-freedom and two degrees-of-freedom) are designed using an H _∞ loop-shaping procedure for a tractor semi-trailer combination to follow the road centerline on both curved and straight highway sections. These lane following controllers ensure robust performance even in the presence of uncertainties in the control model.; The vehicle longitudinal speed affects the lateral dynamics significantly. Earlier steering controllers have addressed the problem by making the controllers robust to variations in the longitudinal speed of the vehicles. Such methods often lead to conservative designs. An alternative design procedure, highlighted here, is to use a gain-scheduled controller with the speed, a measured variable, as the scheduling parameter. The H_∞ loop-shaping controller has an observer based state-feedback structure, which makes it suitable for gain-scheduling. The effectiveness of the proposed controllers is demonstrated by both simulations and experiments.; A possible secondary control input for lateral control is the differential braking force to the rear wheels of the trailer. It generates a torque, which can be used to control the yaw motion of the trailer for effective handling of yaw instability modes such as fish-tailing. Coordinated steering and braking control for HVs is introduced with explicit consideration of model uncertainties and availability of sensors. The effectiveness of the proposed coordinated controller is demonstrated by simulations.; Another contribution of this dissertation is the formulation of design of the combined lane following and lane changing problem as a hybrid automaton design problem. In the hybrid automaton the reference is switched between zero for the lane following mode and the desired lane changing trajectory for the lane changing mode. Results of the worst case reachability analysis of the hybrid automaton, using the proposed H_∞ controller, demonstrate that a satisfactory safe set is attained. It is also shown that the lane following controllers, presented earlier, can be used for lane changing maneuvers as well.