Dynamic Parameter Identification Of Collaborative Robots Based On A Base Force Sensor And Its Application In Kinesthetic Teaching And Collision Detection

The increase in living standards brings about a trend toward individualized consumption.For meeting the consumer’s need,the manufacturing industry is developing towards a small batch and flexible production method.For rapid automation of this production model,using collaborative robots is an excellent choice,because of its flexible deployment and simple operation.Therefore,collaborative robots have developed rapidly in recent years.Flexibility,simple operation and safety,as the three-core competitiveness of collaborative robots,are mainly embodied in kinesthetic teaching and collision detection.This study proposes a scheme based on a base force sensor to achieve high-precision collision detection and flexible kinesthetic teaching.The key point of this scheme is to observe the external forces accurately based on the robot dynamics and respond appropriately.Therefore,this study focuses on four aspects: dynamic modeling,dynamic identification,observation of the external force and responding to the external force.First,this study proposed a method for establishing the dynamic model of the series robot,including a linear dynamic model of the robot joint and a linear dynamic model of the robot base.Then,a numerical analysis method and an analytical analysis method for obtaining the maximum identifiable dynamic parameter set of a robot were introduced.Then,the method of identifying the robot’s dynamic parameters by using the base force information and joint torque information was proposed.Compared with the traditional identification method,it can obtain more accurate joint link dynamic parameters and joint friction parameters.In the identification process,a method of using the robot’s own motion to identify and compensate the signal coupling of the force sensor was also proposed.Simulations and experiments showed that the proposed dynamic modeling and identification method could accurately identify the dynamic parameters.Then,this study proposed a force observer based on the identified dynamics model,which could accurately observe the magnitude of the external force and roughly observe the position of the external force.Observation information was effectively applied to the kinesthetic teaching of collaborative robots,which realized the kinesthetic teaching in work space.Moreover,this study proposed an adaptive algorithm for kinesthetic teaching based on the frequency of teaching force,which realized a rapid suppression of the teaching instability.Furthermore,observation information was effectively applied to the collision detection of collaborative robots,realizing high-precision full-arm collision detection.Experimental results showed that the proposed observation method and response strategy had good human-robot interaction effects,which could enhance the simple operation and safety of collaborative robots.