A Universal Control Scheme of Human-Like Steering in Multiple Driving Scenarios

Autor: Jian Song, Shengnan Fang, Shuai Cheng
Rok vydání: 2021
Předmět:
Zdroj: IEEE Transactions on Intelligent Transportation Systems. 22:3135-3145
ISSN: 1558-0016
1524-9050
DOI: 10.1109/tits.2020.2982002
Popis: To incorporate the inherent superiorities of an experienced, proficient driver, a human-like control scheme for automated steering systems is proposed in this paper based on the concept of “staying within the safety zone”. Designed in the precognitive architecture, the novel control scheme takes into account key characteristics of human steering behaviors in high-speed conditions by determining the steering activation and the steering intensity based on the inversed-time-to-lane-crossing (iTLC), generating smooth, moderate steering actions to maintain the vehicle within a predefined safety zone. The proposed control scheme provides a universal solution for steering control in multiple driving scenarios by accordingly allocating the safety zone to ensure the driving safety therein. A simulator experiment was conducted to test both the fully autonomous and the driver-in-the-loop performance of the proposed control scheme in lane keeping and lane changing. A traditional error-minimizing control algorithm was implemented as a comparison. Results indicate that the novel scheme outperforms the conventional control schemes by yielding higher steering performance, lower steering effort, and higher robustness, proving the benefits of imitating the patterns of human control behaviors in the control algorithm of an automated steering system.
Databáze: OpenAIRE