With the growth of lightweight robots in industry handling assistance becomes more and more important for solving industrial tasks.

We present an adaptive compliance control mode for industrial robots based on a learned equilibrium model.

This enables us to cope with changing manufacturing environments, attached grippers as well as devices with inaccurate dynamic models, e.g. stiff tubes, wires, protection shields or hose packages.

A further feature of the proposed method is the expandability by an additional parameterization that allows to deal with task variability.

In this work we evaluate our approach using the example of a task that incorporates variable payloads. All experiments are conducted in a simulation framework to evaluate the feasibility of the proposed approach for industrial robot scenarios.