Research on the influence of autonomous controller on human–machine shared control system in obstacle avoidance scenario

• Introducing a multi-dimensional controller evaluation framework. • Proposing controller aggressiveness as a key compatibility factor. • Shifting authority allocation from driver traits to controller features. Although controller design plays a pivotal role in the human–machine shared control system, a systematic comparison of different control paradigms is lacking. This study addresses this issue by developing a comprehensive evaluation framework to investigate the impact of different types of controllers on system performance. First, six categories of mainstream controllers are taxonomically classified based on their control logic and implementation principles. Representative controller prototypes are then developed for each category. Next, a human–machine shared control system framework incorporating multi-dimensional performance evaluation metrics is established. A human–machine cooperative driving experimental platform is then used to evaluate these controllers in obstacle avoidance scenarios. The experimental results reveal three key findings: (1) Controller type significantly affects driver operation behavior and system performance metrics; (2) Controller aggressiveness is a critical factor in system compatibility; and (3) Authority allocation outcomes are more influenced by controller type than driver experience. These findings challenge conventional design assumptions and provide an evidence-based methodology for selecting and optimizing controllers to achieve the desired performance characteristics of a human–machine shared control system.