Terminal Force Soft Sensing of Hydraulic Manipulator Based on Joint Torque Compensation

To meet the requirements of the high-precision operation of the hydraulic manipulator in a complex environment, it is necessary to equip it with an accurate force measurement system. Due to the large and complex end-effector contact force against the working environment, the force sensor is vulnerable to damage. Therefore, focusing on the accurate measurements of the terminal force with force sensorless, this paper proposed a force soft-sensing method of hydraulic manipulator based on joint torque compensation, taking 3-D of hydraulic manipulator as the research object. A finite Fourier series model was used to design the excitation trajectory, and the recursive least squares method was used to identify the dynamics parameters of the manipulator. The nonlinear friction torque model was used to replace the Coulomb viscous friction model to improve the precision of manipulator dynamics model. A neural network joint torque compensation model was established to reduce the influence of uncertain factors on the precision of the dynamics model. The AMESim/Simulink co-simulation model was built to design the triangular trajectory of the hydraulic manipulator end-effector, which verified the high accuracy of the dynamic model proposed in this research. A constant force load of 500 N and a variable force load of 0~500 N were applied to the horizontal and vertical direction of the end-effector respectively. Under the constant force load, the soft sensing accuracy of the end force in the horizontal and vertical directions is 4.84% and 2.79%, respectively. Under the variable force load, the accuracy of the soft sensor can reach 5.73% in the horizontal direction and 4.81% in the vertical direction. By comparing the end force soft sensor accuracy with that before optimization, it is verified that the force soft-sensing model of hydraulic manipulator based on joint torque compensation can effectively improve the accuracy of end-effector contact force.