Model Predictive Control of Robotic Grinding Based on Deep Belief Network

Abstract

Considering the influence of rigid-flexible dynamics on robotic grinding process, a model predictive control approach based on deep belief network (DBN) is proposed to control robotic grinding deformation. The rigid-flexible coupling dynamics of robotic grinding is first established, on the basis of which a robotic grinding prediction model is constructed to predict the change of robotic grinding status and perform feed-forward control. A rolling optimization formula derived from the energy function is also established to optimize control output in real time and perform feedback control. As the accurately model parameters are hard to obtain, a deep belief network is constructed to obtain the parameters of robotic grinding predictive model. Simulation and experimental results indicate that the proposed model predictive control approach can predict abrupt change of robotic grinding status caused by deformation and perform a feed-forward and feedback based combination control, reducing control overflow and system oscillation caused by inaccurate feedback control.

An online human–robot collaborative grinding state recognition approach based on contact dynamics and LSTM

“…Human-robot collaborative grinding Human-robot interaction force occur, contact force changes non-linearly its first and second derivatives; u is motor voltage; is the contact force caused by motor voltage and grinding depth; F r is external disturbance force; and F r is robotic grinding force, which can be expressed as (Chen et al, 2019).…”