Sensorless force estimation and control of Delta robot with limited access interface

Abstract

Purpose Delta robot is a parallel robot specifically designed for high-speed pick and place tasks. However, sometimes they are asked to perform additional assembling and squeezing actions, which is beyond the capability of position-controlled Delta robots. Force sensors may be expensive and add mass to the system. Therefore, the purpose of this paper is to study sensorless force control of Delta robots using limited access interface. Design/methodology/approach Static force analysis is performed to establish a relation between joint torques and external forces. The joint torques are observed from signals provided by motor drivers. A distributed mass model is proposed to compensate the gravity of upper arms and forearms. To minimize the effect of backlash and nonlinear frictions brought by gearboxes, model parameters are calibrated in two separated modes: “LIFTING” and “LOWERING”. Finally, a hybrid force estimation model is built to deal with both cases simultaneously. Surrogate model-based force control law is proposed to increase the force control loop rate and handle the force control problem for discrete position-controlled Delta robots. Findings The results show that the force estimation model is effective and mode separation can significantly improve the accuracy. The force control laws indeed stabilize the robot in desired states. Originality/value The proposed solution is based on position-controlled commercial Delta robot and requires no additional force sensor. It is able to extend Delta robots’ capability and meet requirements of emerging complex tasks.

A nonlinear robot joint friction compensation method including stick and sliding characteristics

“…It is necessary to adequately establish the robot dynamic model in many cases. An almost accurate dynamic model, contributes to improve transient performance and reduce steady-state tracking errors in high precision closed-loop motion control (Zhang et al , 2022) and torque control algorithms (Cheng and Jiang, 2018; Zhang et al , 2021). It can increase the sensitivity and accuracy of security perception in collision detection (Guo et al , 2018a, 2018b; Zhang et al , 2019).…”