Abstract
Today modern automobiles are equipped with an increasing number of driver assistance systems. These systems are designed to support the driver in common use or assist him in critical driving situations. For this purpose a huge number of sensors are required, which already provide many information about the driving condition of the vehicle. Based on the existing sensor infrastructure, autonomous driving, that means driving without or only limited intervention of the driver, is a further development of driver assistance systems. This paper presents a concept for an autonomous track guided electro-vehicle. One of its main parts is a path planning strategy. Based upon a nonlinear single track model, which describes the vehicle’s driving dynamics, a method for optimal control of nonlinear ODE is used to generate the reference control values, as there are steering angle and driving force. This allows to lead the vehicle alongside a calculated lane, which is optimal for different objectives, like energy consumption or comfort. Also various constraints like width of the street or maximal lateral and longitudinal accelerations are taking into account. Due to different unknown disturbances it is also necessary to build up an additional closed loop path control system. In contrast to other path control approaches, which work with a virtual look ahead point in front of the vehicle, the system is controlled to the actual position of the vehicle to use the full potential of the path planning strategy. With the actual position of the vehicle and its reference position, also generated from the path planning strategy, it is possible to calculate its present lateral and longitudinal distances from the lane. These distances are used as control variables for two decoupled adaptive controllers for lateral and longitudinal dynamics based on a linearization of the nonlinear single track model. To pre-evaluate the designed system, path planning and path control strategy are implemented in a nonlinear simulation environment using Matlab/Simulink. The operability of the system is shown on the basis of some simple driving experiments.
KEYWORDS – autonomous driving, optimal control, moving horizon, path control, nonlinear system