Research On Grinding Control Of Collaborative Robot Based On RBF Force/Poistion Mixing

Aluminum alloy as a common metal material,because of its light weight,corrosion resistance,durable and other good characteristics,is widely used in furniture,electrical appliances and communication equipment manufacturing.At present,the surface treatment of aluminum alloy products polishing process is still mainly artificial,but artificial grinding efficiency is low,high cost,not conducive to the production of aluminum alloy products,and to replace artificial aluminum alloy products surface polishing robot can improve the production efficiency and quality of enterprises,reduce the production cost.Because of the advantages of easy deployment,low price and force control,cooperative robot is an ideal tool for surface grinding of aluminum alloy products.Aiming at the problem of surface grinding of aluminum alloy products,this paper studies force control and pose planning of surface grinding of aluminum alloy products based on UR5 e cooperative robot.The main work contents and achievements are as follows:(1)Establish kinematics,dynamics,terminal contact force and pose planning models of robot grinding system.Firstly,kinematics and dynamics equations of the robot are calculated according to the parameters of connecting rod and joint.Then,the contact force between the robot’s end grinding tool and the flat and curved aluminum alloy surfaces was analyzed respectively.At the same time,in order to realize the constant control of the force in the tracking process,the stiffness model was established and the plane grinding and radial basis(RBF)neural network model were applied to the surface grinding.Finally,the attitude of the desired trajectory of the robot is planned when the surface is polished.The equal spacing method is used to divide the path points on the grinding trajectory,and the grinding trajectory is divided into several path points.Then,according to the different curvature of the surface between the two adjacent path points,the path is divided into several micro surfaces.(2)Aiming at the coupling problem of force and pose during robot grinding,a hybrid control algorithm based on RBF neural network was designed to decouple force and pose.On the position and posture controller,the RBF neural network was used to approximate and adjust the systematic errors caused by the robot touching the workpiece in the process of grinding,so as to realize the accurate position,posture and trajectory control of the robot end tool,and the correctness of the algorithm was verified by the Lyapunov function.On the force controller,PID control algorithm is used to compensate the gravity and mechanical vibration of the robot,and reduce the force control error when the robot is grinding.(3)Matlab and Gazebo were used for co-simulation to verify the effectiveness of the proposed control algorithm.Firstly,the robot based on RBF neural network algorithm can track the joint trajectory and the terminal position trajectory,and the stable tracking effect of the terminal trajectory can be achieved after 2 s.Then,the expected normal force is set as 25 N and 45 N respectively,and the workpiece stiffness is 25000.The constant force tracking simulation test is carried out on the robot end normal force.The test results show that the maximum error between the expected normal force and the actual normal force is ±3 N.(4)Build a robot grinding experiment platform to verify the performance of the designed controller.The expected grinding force of the robot was set to 30 N,and the actual grinding experiments were carried out on the flat aluminum alloy plate,curved aluminum alloy plate and aluminum alloy automobile wheel hub,and the grinding effects of position control,PID force/bit hybrid control and RBF neural network force/bit hybrid control were compared and analyzed.Experimental results show that when the force/bit hybrid control based on RBF neural network is adopted,the minimum error between the expected normal force and the actual normal force at the end of the robot is ±3 N.Meanwhile,compared with the surface roughness of the polished aluminum alloy,the mean roughness is reduced by 0.407 um,and the standard deviation is reduced by 0.348 um.The surface roughness of aluminum alloy is the lowest.The uniformity and consistency of the surface are the highest.It is concluded that the control algorithm proposed in this paper has better grinding effect.