Search papers, labs, and topics across Lattice.
This paper introduces a hybrid framework for multi-point object pushing that combines a dynamic contact model with residual reinforcement learning (RL). The dynamic model incorporates unilateral constraints and a box friction model to represent contact nonlinearities, and is extended to handle multiple simultaneous contact points. The RL component acts as a residual module to improve the performance of a model-based controller, enabling more accurate trajectory following and stable contact interactions.
Augmenting model-based controllers with residual RL can significantly improve the robustness and accuracy of multi-point object pushing compared to traditional PD control.
Robotic contact manipulation involves applying controlled forces at contact points to guide an object along a desired trajectory while respecting the underlying physical interactions. This letter presents a novel framework that integrates dynamic modeling and Reinforcement Learning (RL) to achieve robust object pushing with a redundant robotic arm. First, a comprehensive dynamic contact model is formulated, incorporating unilateral constraints and a box friction model to capture the nonlinearities present in real-world contact dynamics. Second, the model is extended to handle multiple simultaneous point contacts, enabling effective trajectory planning and tracking for a redundant robotic arm in multi-contact pushing tasks. Third, an RL strategy is introduced as a residual module that augments a model-based controller to improve pushing performance. Simulation and real-world experiments with a Kinova Gen2 arm demonstrate that the proposed method achieves accurate trajectory following and stable contact interactions, significantly outperforming traditional PD control strategies in dynamic pushing scenarios.
