Research On The Key Technologies Of Human Machine Cooperative Safety Protection For Robot Assembly

Human-robot collaboration and human-machine interaction are currently one of the important contents of research in the field of robotics.Humans and robots have complementary advantages when performing assembly operations.In the working environment,humans are very good at sensing tasks and are very suitable for narrow spaces.Perform complex operations.The robot is good at picking and placing tasks with high repeatability,and the robot can perform tasks at high speed and maintain accuracy.Human-robot cooperation can make better use of the advantages of humans and robots.Therefore,it is of great significance to develop human-robot cooperation safety protection technology for robot assembly.This paper proposes a multi-eye vision-based human-machine collaboration safety protection technology.It uses multiple RGB-D sensors to sense working environment information and detect the position of human joint points.In order to ensure real-time performance and improve processing speed,mean value drift is used to human joints.Point position is tracked,and then according to different cooperation tasks,the robot adopts different safety protection controls,including robot speed control and trajectory control.In this paper,the ABB IR1200 robot and multiple LeEco RGB-D vision sensors are used as experimental platforms,and the safety protection technology is studied by combining visual perception and robot control.The research contents are as follows:First of all,in the research of the perception part,in order to make the sensor completely cover the working space as much as possible,this paper proposes a sensor layout method that uses three sensors to sense the working environment and reduce the dead zone.The OpenPose algorithm was used to detect the coordinates of 18 joint points of the human body in the working environment,and a tracking algorithm was added to improve the detection frame rate.Secondly,according to the relative position relationship between different joint points of the human body and the robot and different human-robot collaboration tasks,the robot will adopt different security strategies.Under loose cooperation,the robot will use speed scaling control.When the human approaches the robot,the robot will Slow down until it stops.When humans and robots cooperate closely,that is,the human-robot space is shared,the joint points of the human body will generate a virtual force,and the robot’s trajectory is controlled according to the dynamics.Finally,the simulation system and the physical platform of the robot are introduced respectively,and the stability,reliability and real-time performance of the technology in this paper are tested under the simulation and the physical platform.The actual performance of the entire technology is understood,and the research results of this paper are carried out.Summarize,point out the shortcomings of the technology and propose improvement directions.