|Optimization of Emergency Braking Pedestrian Collision Avoidance for Autonomous Vehicle Fusing the Fuzzy Neural Network with the Genetic Algorithm
Wei Yang*, Zhen Zhang, Kongming Jiang, Qian Lei, and Ketong Zong
International Journal of Control, Automation, and Systems, vol. 20, no. 7, pp.2379-2390, 2022
Abstract : In order to improve braking performance of the intelligent vehicle, a pedestrian collision avoidance control model for autonomous emergency braking pedestrian (AEB-P) system based on the fuzzy neural network (FNN) and genetic algorithm (GA) theory is proposed. In this research, we construct a backpropagation (BP) feedforward FNN-GA model for the AEB-P system. Aiming to solve the problem that initial training parameters of the FNN model generate randomly, they are optimized by fusing the fuzzy neural network with GA. Simulation results show that training epochs reduce from 800 to 60 and training error changes from [−0.06, +0.06] to [−0.04, +0.04] after optimization. Moreover, the maximum model error lessens from 0.058 to 0.0351 and sample total error decreases from 0.0179 to 0.0068. At last, the proposed scheme is applied to the China New Car Assessment Program (C-NCAP) simulation test scenarios. Output curves of vehicle distance, velocity, and deceleration become smoother and the maximum error decreases from 0.51 m/s2 to 0.26 m/s2 , which verifies the proposed FNN-GA control strategy for AEB-P system is effective and reliable.
Autonomous emergency braking, backpropagation feed-forward fuzzy neural network, genetic algorithm, pedestrian collision avoidance.
Download PDF : Click this link