A control algorithm for adaptive cruise control systems

This thesis aims to develop a configurable control algorithm for the Adaptive Cruise Control (ACC) system, which can generate different behaviors in order to gain wider user's acceptance. The control algorithm includes a basic algorithm structure with several design variables. On the one hand, based on literature found on different control methods for ACC, the linear controller, the fuzzy logic controller and the sliding mode controller are chosen as basic algorithm candidates realized by simulation models. In order to study the flexible potential of each algorithm structure, the control performances of several existing ACCs (with regard to a group of defined driving scenarios) are taken as the performance criteria. On the other hand, the genetic algorithm (GA) optimization routine, which reduces the discrepancy between the desired behaviors and the actual behaviors of the developed control algorithm, is used to obtain optimum design variables. The co-simulation between MATLAB/Simulink and other commercial software is applied to support the optimization process. The potential of each controller is evaluated by comparing the desired behaviors and its simulated results. In addition, the controller containing the best development potential is further improved by modifying the basic algorithm structure and the new control behaviors are very similar to all the desired behaviors. The improved ACC control algorithm containing its optimized design variables is installed in a vehicle and its performance is tested by measurements. The test results indicate that the developed control algorithm can work properly as an ACC system in real traffic situations and the measured output of the controller is similar to the designed output from the simulation.