Compensate undesired force and torque measurements using parametric regression methods

Haptics as well as force and torque measurements are increasingly gaining more attention in the fields of kinesthetic learning and robot Learning from Demonstration (LfD). For such learning techniques, it is essential to obtain accurate force and torque measurements in order to enable accurate control. However, force and torque measurements using a 6- Axis force and torque sensor mounted at the end-effector of an industrial robot are known to be corrupted due to the robot's internal forces, gravity, un-modelled dynamics and nonlinear effects. Non-parametric regression is used to alleviate the negative impact of these factors on the measurements. However, non-parametric regression requires data to be available on-line which increases the system latency. In this paper, parametric regression will be used to estimate the undesired forces at the end effector for a pre-defined trajectory with limited speed. The parametric regression requires low computational complexity without intensive training over the operational space under the given assumptions. In addition, parametric regression does not need data to be available online. In this work, two compensation methods, namely linear regression and Random Forest Regression are experimentally evaluated and their relative performance is established in comparison to each other. These methods are experimentally validated using Motoman SDA10D dual- Arm industrial robot controlled by Robot Operating System (ROS). The experiments showed that force and torque compensation based on linear regression and random forests has tangentially close performance.