Model Predictive Control of a Heavy-Duty Truck Based on Gaussian Process

A recent trend on the automotive market is the incorporation of several driver assistance systems into common vehicles. Among the most studied systems, lies the cruise control (speed), either for platooning or to ease off driver's tasks. For the design of this controller, particularly when applied to large and heavy vehicles, a longitudinal model that represents satisfactorily all the features of this complex system is necessary. In this sense, a Gaussian Process model was applied within Model Predictive Control to regulate the speed of a heavy-duty truck. The controller takes the variance produced by the Gaussian Process model into account on the optimization of the control signal. The proposed controller achieved low tracking error even on hard conditions, like steep roads.