ABPolicy: Asynchronous B-Spline Flow Policy for Real-Time and Smooth Robotic Manipulation

TL;DR

ABPolicy introduces an asynchronous B-spline flow policy for robotic manipulation that enhances motion smoothness and responsiveness by operating in a control-point space, ensuring intra- and inter-chunk continuity with real-time updates.

Contribution

The paper presents a novel asynchronous flow-matching policy using B-spline control points, improving smoothness and responsiveness in robotic manipulation tasks.

Findings

Reduces trajectory jerk for smoother motion

Achieves real-time, continuous updates in manipulation

Improves performance across diverse tasks

Abstract

Robotic manipulation requires policies that are smooth and responsive to evolving observations. However, synchronous inference in the raw action space introduces several challenges, including intra-chunk jitter, inter-chunk discontinuities, and stop-and-go execution. These issues undermine a policy's smoothness and its responsiveness to environmental changes. We propose ABPolicy, an asynchronous flow-matching policy that operates in a B-spline control-point action space. First, the B-spline representation ensures intra-chunk smoothness. Second, we introduce bidirectional action prediction coupled with refitting optimization to enforce inter-chunk continuity. Finally, by leveraging asynchronous inference, ABPolicy delivers real-time, continuous updates. We evaluate ABPolicy across seven tasks encompassing both static settings and dynamic settings with moving objects. Empirical results indicate that ABPolicy reduces trajectory jerk, leading to smoother motion and improved performance. Project website: https://teee000.github.io/ABPolicy/.