Search papers, labs, and topics across Lattice.
This paper introduces IMPACT, a novel framework for robotic manipulation that separates task planning from internal-model predictive control to handle forceful interactions with various objects. By addressing the limitations of existing imitation learning methods, which struggle with generalization and complexity, IMPACT enhances performance in real-world tasks such as tool use and object transport. The results show that IMPACT significantly improves success rates and generalizes better to new object weights while ensuring greater safety and energy efficiency in robotic operations.
IMPACT achieves higher success rates in robotic manipulation tasks while generalizing effectively to unseen object weights, outperforming traditional methods.
Real-world robotic manipulation tasks often involve forceful interactions with the environment, such as using tools of varying weights, transporting objects with different masses, and performing contact-rich tasks like table wiping. Previous learning-based approaches typically employ imitation learning policies that output target end-effector poses tracked by low-level impedance controllers. In these systems, forceful interactions are either implicitly realized through steady-state tracking errors or explicitly commanded using wrist force/torque or tactile sensors. However, implicit approaches generalize poorly across object weights, while explicit approaches require specialized hardware and increase system complexity. In this work, we propose IMPACT, a framework that decouples these forceful tasks into task-planning and internal-model-based predictive control. Extensive simulation and real-world experiments demonstrate that the proposed framework achieves higher success rates and improved generalization to unseen object weights, as well as better safety and energy efficiency.
