Force control
Force control is essential to precisely perform tasks involving contact, which are ubiquitous in robotics. Yet, current Model Predictive Control (MPC) approaches fail to handle this complex sensory input. The main reason for this is that most state-based formulations consider the robot’s joint position and velocity as states and forces as control inputs, thus preventing effective use of the force sensor.
During my PhD, I collaborated with Sébastien Kleff to propose solutions to this challenging problem. We explored two different approaches:
The first was naive and consisted of using an estimation module to estimate and compensate the model’s inaccuracies. Despite its simplicity, this approach led to surprisingly good results in force tracking tasks [1].
The second one investigated the use of a visco-elastic model of the contact force and achieved state-of-the-art results [2].
Selected publications
[1] Armand Jordana★, Sébastien Kleff★, Justin Carpentier, Nicolas Mansard, and Ludovic Righetti. “Force feedback model-predictive control via online estimation.” In 2024 IEEE International Conference on Robotics and Automation (ICRA).
Paper Video Abstract
Nonlinear model-predictive control has recently shown its practicability in robotics. However it remains limited in contact interaction tasks due to its inability to leverage sensed efforts. In this work, we propose a novel model-predictive control approach that incorporates direct feedback from force sensors while circumventing explicit modeling of the contact force evolution. Our approach is based on the online estimation of the discrepancy between the force predicted by the dynamics model and force measurements, combined with high-frequency nonlinear model-predictive control. We report an experimental validation on a torque-controlled manipulator in challenging tasks for which accurate force tracking is necessary. We show that a simple reformulation of the optimal control problem combined with standard estimation tools enables to achieve state-of-the-art performance in force control while preserving the benefits of model-predictive control, thereby outperforming traditional force control techniques. This work paves the way toward a more systematic integration of force sensors in model predictive control.
[2] Sébastien Kleff, Armand Jordana, Nicolas Mansard, and Ludovic Righetti. “Force Feedback in Model Predictive Control: A Soft Contact Approach.” preprint (2024).
Paper Video Abstract
Model-predictive control is an appealing framework to control robots due to its ability to exploit both sensory information and model predictions. But its performance remains fundamentally limited in tasks involving contact with the environment, in part because optimal control policies do not reason over force measurements. In this article, we propose a first complete answer to this issue by introducing a novel approach to perform force feedback in model-predictive control. We propose to augment the state-space with a visco-elastic model of the contact force in the task space in order to systematically include measured efforts into the optimal control loop. We derive a complete predictive controller with an efficient formulation whose implementation is released in open-source. We demonstrate through simulation studies and hardware experiments that our approach enables to combine the benefits of force control and model-predictive control within a single architecture, thereby outperforming existing approaches in challenging contact tasks.
