Search Results

Motivated by applying artificial intelligence technology to the automobile industry, reinforcement learning is becoming more and more popular in the community of intelligent driving research. The reward function is one of the critical factors which affecting reinforcement learning. Its design principle is highly dependent on the features of the agent. The agent studied in this paper can do perception, decision-making, and motion-control, which aims to be the assistant or substitute for human driving in the latest future. Therefore, this paper analyzes the characteristics of excellent human driving behavior based on the six-layer model of driving scenarios and constructs it into a human knowledge graph. Furthermore, for highway pilot driving, the expert demo data is created, and the reward function is self-learned via inverse reinforcement learning. The reward function design method proposed in this paper has been verified in the Unity ML-Agent environment.

In recent years, electric vehicle and hybrid vehicle are either on the market or under intensive research and development (R&D). Since the concept of auxiliary power unit (APU) was brought into the automotive industry, the range-extended electric vehicle (ReEV) has become the favor of the worldwide manufacturers. Normally, the APU starts and stops more frequently in response to the control strategy compared with traditional vehicles, which will affect the ride comfort of passengers. Thus, APU start-stop NVH refinement is an important aspect of ReEV R&D. In this paper, a subjective evaluation on a ReEV was performed to quickly diagnose NVH issues firstly. Based on subjective results, the NVH experiment in a semi-anechoic room was carried out to troubleshoot these issues. The accelerations of the APU mounts, the seat track and the steering wheel as well as interior noise level were acquired and analyzed.