Research On Force Control Of Collaborative Robot Based On Sensorless Estimation Of External Wrench

As the development of robots,people have more and more demands for collaborative robots.In the meantime,security issues and interaction which is between human and computer are increasingly concerned.The traditional speed-position control method has not met people's current needs.Force control gradually replaced the original speed-position control.Force control has applications in various fields,such as assembling products on industrial production lines,helping people to make movements which require greater strength,collaborating with people to complete work,preventing damage due to collisions,and so on.If force control is to be applied to robots,it is necessary for robots to be able to estimate their own external forces.People currently use various sensors to achieve this purpose.Sensors also have their own disadvantages,such as high cost,easy to change the structure of the robot,etc.,so the external force estimation without sensor is necessary.If no sensors are used,accurate robot models will be very important.Dynamic parameter identification is very important to obtain accurate robot's models.In this paper,the parameter identification based on the connected combination method and the estimation method of sensorless external force using Kalman filter are proposed.During the parameter identification,the trajectory of joint needs to be optimized.This article introduces an artificial fish swarm algorithm to optimize the trajectory.Due to the insufficient application of traditional artificial fish swarm algorithm in trajectory optimization,this article has improved three aspects of artificial fish swarm algorithm.They are visual field and step size,optimal individual retention and foraging behavior.The improvement can optimize the trajectory well.The connected combination method is used to simplify the robot model,so that each identification is equivalent to the identification of a two-link robot,which accelerates the identification speed and improves the identification accuracy.After obtaining the robot dynamics parameters,the generalized momentum is introduced to redefine the dynamic equations,determine the noise term,adjust the Kalman filter parameters,and improve the compensation scheme when the joint speed is close to zero.The joint angle,joint current,and joint angular speed are used to estimate the external force on the end of the robotic arm without using other sensors.Compared with the original compensation scheme,the accuracy of the estimation has improved.Based on the method mentioned above,the article implements a scheme for parameter identification and external force estimation without sensors.Through the analysis of the simulation results,the ideas in this paper can realize parameter identification and external force estimation without sensors.