STOCHASTIC FORCE CONTROL IN A ROBOT ARM

Abstract Stochastic force control, a new technique of force sensing and control in robot arms is discussed. Force sensing in industrial robots is usually an elaborate task requiring both hardware and software modification of the original design of the robot arm. This new approach does not employ any sensor and the force information is derived from the variation of the current in the joint actuators. The current waveforms of the joints are modelled using ARIMA (Auto-Regressive Integrated Moving-Average) models. The error between observation and the prediction is utilised to modify the robot position and remove external forces acting on the robot. The technique is being developed considering a number of applications such as edge following and peg-hole insertion. The latest progress in the project is given.