A Method for Pilot Acceptability Evaluation of Airborne Collision Avoidance System Based on Genetic Algorithm

Compared with the existing Traffic Alert and Collision Avoidance System (TCAS), the development of the new system, Airborne Collision Avoidance System X (ACAS X), uses a model-based optimization method to improve airspace safety and operational efficiency. However, due to the opacity of the model-based optimization method in the process of automatically generating logic, the resolution advisory issued by ACAS X are often misunderstood or difficult for pilots to accept. In this paper, first we introduce the development process of ACAS X, and then analyze the natural shortcomings of ACAS X and the evaluation metrics for ACAS X optimization from the perspective of human factors engineering. Finally, we propose a method for evaluating the pilot acceptability of collision avoidance system based on Genetic Algorithm (GA). This method is used to find high collision risk encounters where ACAS X tends to issue advisories that are not conducive to the pilot's acceptance, so as to evaluate the pilot's acceptability of the collision avoidance system. The final simulation results demonstrate the effectiveness of the method.