Systematic Optimization of Parameters for Adaptive Local Trajectory Planning Using Genetic Algorithm

The following paper deals with an assistance function based on genetic algorithms (GA) to optimize the weighting factors of the driving function “Adaptive Local Trajectory Planning (alTp)”, which serves as a part of the trajectory planning of autonomous driving to compensate for the inability of global trajectory planning in obstacle avoidance in the dynamic traffic environment. Considering the derived requirements, the considered optimization problem is coded and initialized. To ensure a steady optimization of the alTp with the specified weighting factors, the same objective criteria of the alTp are also applied in the fitness function of the assistance function for the quantitative evaluation of the respective solution candidates. Using a test scenario with driverless transport vehicles (AGV) for material transport in intralogistics, the developed assistance function is validated with the help of a virtual test bench and demonstrated.