String-stable CACC design and experimental validation

The design, analysis and experimental validation of a Cooperative Adaptive Cruise Control (CACC) system are presented. Focussing on the feasibility of the implementation, vehicle-to-vehicle communication with the directly preceding vehicle only is considered. This communication structure is known as semi-autonomous ACC and enables ad-hoc platooning and dealing with non-CACCeqquipped vehicles in a straightforward way compared to other communication structures. Via wireless communication, the acceleration of the directly preceding vehicle is obtained. The signal is used in a feedforward setting in combination with ACC feedback. Hence, in the case of no feedforward, an ACC system instead of a CACC system results. Consequently, if the directly preceding vehicle is not CACC-equipped, or if communication fails, ACC functionality is still available. For ACC, the inter-vehicle distance and relative velocity of the preceding vehicle are obtained by on-board sensors, e.g. radar or lidar. To allow straightforward implementation on different types of vehicles, the (C)ACC system is divided into a lowerand an upper-level controller. The lower-level controller is vehicle-dependent and controls the acceleration of the vehicle by actuation of the throttle and brake system. The corresponding closed-loop system is approximated by a simple process model, called the vehicle model. The vehicle model parameters are easy to identify, e.g. using step response measurements. The desired acceleration is computed by the upper-level controllers, based on the information from on-board sensors and, in the case of CACC, wireless communication. The design variables of the upper-level controller are the ACC feedback controller, the CACC feedforward controller and the spacing policy dynamics. The spacing policy is the desired inter-vehicle distance. Using a frequency domain approach, the presented (C)ACC system is analyzed for stability, performance and string stability, where vehicle dynamics and communication delay are taken into account. String stability considers the propagation of oscillations upstream a string of vehicles, i.e. a platoon. Inputoutputand error string stability functions are derived, which indicate the amplification of these signals upstream a platoon. For a homogeneous platoon, the error-, inputand output string stability transfer functions are identical. For heterogeneous traffic, the output string stability function has to be considered. Considering a velocity-dependent spacing policy, the CACC system enables string stable behavior at smaller inter-vehicle distances and smaller distance errors. Hence, a small inter-vehicle distance can more safely be adopted in the case of CACC compared to ACC. Road experiments with two CACC-equipped vehicles validate this theoretical result.