A Multi-perception fusion using shared-control method for brain-mobile robot

For human-robot collaboration, brain-computer interface is promising to express human perception to improve the adaptability of human-robot collaboration in complex environments. In this study, a multi-perception fusion using shared control method (MPF-SC) is proposed to accurately integrate human perception and robot perception. This MPF-SC is applied in brain-controlled mobile robots to accomplish navigation and obstacle avoidance in complex terrain with multiple undetectable obstacles. The MPF-SC establishes a mapping relationship between visual stimulus interface and environment by computer vision, and utilizes a grid costmap to describe the human perception. It integrates EEG and EMG signals with user intent to dynamically adjust the grid costmap, mapping obstacle regions and integrating robot navigation to jointly accomplish driving tasks-with the aim of achieving human-machine shared perception. Sixteen subjects participated in an online obstacle avoidance experiment and compared the performance of the proposed method with two traditional methods. The research results show that the MPF-SC can generate smoother trajectories, achieve a significantly reduced collision rate during navigation, and significantly enhance user comfort. The MPF-SC based on brain-computer interface, fully leverages human anticipation of risks and the robot's perception of obstacle environments, demonstrating that bilateral intelligence is capable of adapting to increasingly complex environments, thereby offering a novel avenue and intuitive avenue for human-machine shared control.