Research on Autonomous Driving Decision Based on Improved Deep Deterministic Policy Algorithm
2022-01-0178
03/29/2022
- Event
- Content
- Autonomous driving technology, as the product of the fifth stage of the information technology revolution, is of great significance for improving urban traffic and environmentally friendly sustainable development. Autonomous driving can be divided into three main modules. The input of the decision module is the perception information from the perception module, and the output of the control strategy is to the control module. The deep reinforcement learning method proposes an end-to-end decision-making system design scheme. This paper adopts the Deep Deterministic Policy Gradient Algorithm (DDPG) that incorporates the Priority Experience Playback (PER) method. The framework of the algorithm is based on the actorcritic network structure model, and the model is based on continuously acquired perceptual information as input. The output of the vehicle is driving. Continuous control amount of action. Combined with the CARLA simulation environment, the state space of the CNN network based on the input of the car's front view image is designed, and the action space design takes into account the actual situation of the accelerator brake being used at different times. In the design of the reward function, the reward function based on the car state information and the reward function based on the artificial potential field method (APF) are designed respectively. After that, based on the CARLA urban driving environment, the DDPG algorithm and PER-DDPG algorithm with different reward functions were simulated and verified. The final experimental results show that the PER-DDPG algorithm using APF achieves the best results in performance, indicating that improved sampling methods and artificial potential field-based reward functions can improve the performance of the algorithm to a certain extent.
- Citation
- YK, S., Wu, J., He, R., and song, S., "Research on Autonomous Driving Decision Based on Improved Deep Deterministic Policy Algorithm," SAE Technical Paper 2022-01-0178, 2022, .