Research On Human-Robot Interaction And Compliance Control Based On Touch/Force Fusion Information

With the continuous development of intelligent robot technology,human-robot interaction has become a new research field.Empowering the robot with powerful interaction has become an important prerequisite for improving the performance of human-robot interaction.Especially in the physical human-robot cooperation,giving the robot a strong intention to understand has become the only way to improve the flexibility of physical human-robot cooperation.However,in the physical human-robot cooperation,due to the immature research of the current flexible skin touch sensor,coupled with the shortcomings of the Neural electrical signal(including electroencephalogram and EMG signal)weak signal,poor anti-interference,contact person intentional understanding in the machine has become a key problem that restricts robot compliance in physical human-robot interaction.Therefore,in view of the common problems in the current physical human-robot interaction in terms of intention understanding and compliance control,this article is based on the collaboration between humans and robots to complete the trajectory following,the research on human-robot interaction and compliance control based on the fusion of tactile and force-sensing information is studied.The research contents mainly include:(1)Contact type analysis and perception system design of everyone's collaborative intentions.Taking everyone's cooperation as the starting point,we study the mechanism of cooperation among everyone,analyze the way of contact-oriented collaboration,and solve the contact person from the perspective of multi-modal interaction.The intention of the machine cooperation solves the problem;At the same time,the design of the tactile and force perception system is completed,and the performance test and packaging of the tactile transmitter are performed.(2)Operational intent description based on touch and force fusion information.Definition and decision generation are comprehensive description and definition of the operator's operation intention in physical human-robot cooperation.Based on traditional six-dimensional force sensing information,increase The flexible array of tactile pressure sensors independently developed by the laboratory collects the image of the hand during the operation process,and indirectly reflects the human operation intention according to the pressure distribution and size of the sensor.First,multi-modal information fusion technology is used to obtain Describe and define the operator's intentions;then use the convolutional neural network to identify the human's operational intent pattern.Finally,the decision generation is based on the description and defined operational intent and recognition results.(3)Contact-based human-robot cooperative compliance control method based on touch and force fusion information Based on the definition of operation intention description and decision generation,a contact-based human-robot cooperative compliance control based on touch and force fusion information is proposed.The strategy introduces the variable that can be used to switch the intention mode to the control system.A variable admittance controller based on haptic feedback is designed to adjust the virtual damping parameter of the robot in real time in response to the author's control intention.The feedback variable admittance control experimental system platform was carried out,and the physical human-robot cooperation follow-up experiment was carried out.The experimental results show that the variable admittance control strategy based on haptic feedback can better solve the human operation intention and greatly improve the flexibility of physical human-robot cooperation.