21 August 2023
Fully exploiting soft robots' capabilities requires devising strategies that can accurately control their movements with the limited amount of control sources available. Existing approaches attack this challenge by means of mostly feedforward learning-based methods or with feedback model-based approaches. The former preserves the robot's elasticity without requiring a precise description of the model. Still, it is time-expensive and does not allow drawing any theoretical conclusion on the system's physical properties or behavior. Conversely, the latter leads to high performance while compensating for disturbances at the cost of stiffening up the robot's behavior and relying on a precise system description. Thus, there is no controller available that can preserve the robot's elasticity while achieving good performance.
To tackle this challenge, EMERGE partners from Delft University of Technology and collaborators propose a controller that relies upon the intersection of learning-based and model-based methods: a purely feedforward iterative learning control algorithm that refines the torque action by leveraging both the knowledge of the model and data obtained from past experiences.
After presenting a 3D polynomial description of soft robots, the authors study their intrinsic properties, e.g., input-to-state stability, and we prove the convergence of the controller coping with locally Lipschitz nonlinearities. Finally, they validate the proposed approach through simulations and experiments involving multiple systems, trajectories, and in the case of external disturbances and model mismatches.
Source: M. Pierallini et al., "A Provably Stable Iterative Learning Controller for Continuum Soft Robots," in IEEE Robotics and Automation Letters, vol. 8, no. 10, pp. 6427-6434, Oct. 2023, doi: 10.1109/LRA.2023.3307007.
Access EMERGE publications in the link below.
More Information
