Force Adaptation in Contact Tasks with Dynamical Systems

Abstract

In many tasks such as finishing operations, achieving accurate force tracking is essential. However, uncertainties in the robot dynamics and the environment limit the force tracking accuracy. Learning a compensation model for these uncertainties to reduce the force error is an effective approach to overcome this limitation. However, this approach requires an adaptive and robust framework for motion and force generation. In this paper, we use the time-invariant Dynamical System (DS) framework for force adaptation in contact tasks. We propose to improve force tracking accuracy through online adaptation of a state-dependent force correction model encoded with Radial Basis Functions (RBFs). We evaluate our method with a KUKA LWR IV+ robotic arm. We show its efficiency to reduce the force error to a negligible amount with different target forces and robot velocities. Furthermore, we study the effect of the hyper-parameters and provide a guideline for their selection. We showcase a collaborative cleaning task with a human by integrating our method to previous works to achieve force, motion, and task adaptation at the same time. Thereby, we highlight the benefits of using adaptive force control in real-world environments where we need reactive and adaptive behaviours in response to interactions with the environment.

A Survey of Robot Manipulation in Contact

“…Many tasks where manipulation in contact is required also involve tools. They can either be rigidly attached to the robot arm or can be grasped by the robot for use; in this survey we do not differentiate between these cases, except by noting that grasping a tool always creates uncertainty regarding Wiping or polishing [18,19,20,21,22,23,24,25,26,27] Grinding or similar [28,29,30,31,32,33,34,35,36,37] Scooping [38,39] Peg-in-hole variants [40, 41, 42, 43, 44, 45, 17, 46, 47, 48, 49, 50] [51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64] [65, 66 Massage [95], velcro peeling [96] engraving [97] the location of the tooltip, which increases the need for compliance when in contact. There are of course methods to alleviate this uncertainty if enough information of the tool has been properly measured (for example, [98]).…”

“…The simplest tasks in this region is perhaps wiping [18], which has been revisited upon multiple papers with different methods [19,20,21,22,23]; in this task the original material is not affected, since wiping mostly refers to cleaning the material. Another term used for, practically, the same task, is polishing [24,25,26,27]. These tasks have been completed with either implicit or explicit control of the forces; implicit is often enough since an important part is keeping contact with the surface.…”

“…Whereas the original version is again for position only, there have been extensions that allow the usage in in-contact tasks. Dynamical systems have been shown to manage variable impedance control [161,38,95] for in-contact tasks, and also recently to provide accurate force control on unknown surfaces [26,27].…”

Model Predictive Robot-Environment Interaction Control for Mobile Manipulation Tasks

“…Impedance control allows tuning the controller's stiffness and damping based on the desired task, but its performance degrades when the environment's model is unknown or time-varying. Recent approaches propose to learn an error force term to compensate for the uncertainties in the robot and the environment model [18]. Another possibility is to change the controller gains based on the force tracking error [19], or an online estimate of the environment model [20], [21].…”

Learning Variable Impedance Control for Aerial Sliding on Uneven Heterogeneous Surfaces by Proprioceptive and Tactile Sensing

“…Literature on adaptive sliding on partially-known and uneven surfaces exists mainly in the manipulation community, using passivity [8] or adaptive force control [9], [10], but they do not directly transfer to aerial robots. Adaptive force control is often used to modify the interaction force along the direction normal to the surface according to geometry changes, while keeping the force constant along the sliding direction.…”