Human-Robot Cooperative Navigation And Planning Of The Brain-Controlled Quadrupled Robot

Human-robot interaction and cooperation are critical technologies in the research field of intelligent robots. The "Brain-Computer Interface" technology acquires and recognizes human brain's activity information directly and has become an important developing orientation for human-robot interaction and cooperation. However, the limited categories of EEG signals that can be obtained by the brain-computer interface, the long time it takes to generate and extract EEG signals, and the signal decoding error limit the application of brain-computer interface greatly in human-robot interaction. Now the research on robot application using brain-computer interface is still at the exploratory stage.Facing the technological limitations of the uncertainty existing in EEG signal commands, how to perform the robot's brain-controlled cooperation tasks is the fundamental and common problem that needs urgent solution in the research field of brain-controlled robot. Using quadruped robot as research object, and the quadruped robot's navigation based on non-invasive brain-computer interface as research objective, this dissertation deeply discusses problems about human-robot interaction and collaboration. This dissertation, on one hand, has proposed the systematic method for brain-controlled robot's navigation and the control system of brain-controlled robots to perform complex tasks, which are the foundations for the research on the future applications of brain-controlled robots. On the other hand, this dissertation has analyzed the influence of the decoding accuracy of EEG signals on the brain-controlled navigation performance, which provides a basis for the decoding algorithm optimization of the EEG signals.This dissertation proposes human-robot collaborative navigation strategy based on brain-computer interface. Aiming at the technological limitations of non-invasive brain-computer interface, this dissertation proposes several kinds of control modes and brain-computer interface protocols, so as to complete human-robot interactions and collaborations under different task levels and realize the logical extension of limited EEG signal instructions. The locomotion efficiency of brain-controlled robots is enhanced by the control strategies based on feed-forward strategy and task synchronization. The influence of the EEG signal's decoding error is weakened by the intention recognition and fault-tolerance mechanism. Based on the methods stated above, this dissertation proposes human-robot collaborative navigation control strategy aimed at efficient, flexible and easy motion control in 2-D environment, as well as human-robot collaborative navigation control strategy aimed at safety and reliability, reasonable response, and remaining the efficiency in 3-D terrain. These human-robot collaborative methods and strategies making great use of robot's intelligence provide new approaches to the research and development of brain-controlled robot.This dissertation proposes the system design method for brain-controlled robot. Facing the development trend of the brain-controlled robot's application to perform complex tasks in the future, learning the higher animals' behavioral control characteristics, this dissertation proposes brain-controlled quadruped robot's behavioral control system structure, and different robot control methods including the brain-computer interaction are incorporated into a unified system. The generic steady-state gait generation method is put forward. All the achivements built a basic platform for future control system development and method research on the brain-controlled robots that can perform complex tasks, and can be extended to the application of brain-controlled robot in the future.The brain-controlled robot based on brain-computer interface is a new research focus in the interdisciplinary among intelligent robots, biomimetic movement and control subjects. This dissertation has conducted fundamental and forward-looking researches which have many potential applications.