B-spline Policy: Accelerating Manipulation Policies via B-spline Action Representations

2026-07-10

AI Summary

Traditional robotic manipulation strategies often rely on discrete action primitives, which hinder high-speed and fluid execution. This work proposes the B-spline Policy (BSP), which introduces B-spline curves as a continuous action representation in reinforcement learning for the first time. By directly predicting knots and control points, BSP generates smooth, temporally scalable trajectories that seamlessly integrate into both policy learning and low-level control. Evaluated in both simulation and real-world robotic tasks, the method significantly reduces task completion time compared to baseline approaches, achieving substantially higher execution efficiency while maintaining high success rates.

Abstract

In this work, we present B-spline Policy (BSP), an action representation designed for accelerating robot manipulation policies. Rather than predicting discrete-time action chunks, BSP parameterizes actions as continuous B-spline curves defined by a set of knots and control points. This representation yields smooth, time-continuous trajectories that can be temporally scaled and executed by low-level controllers at higher frequencies and speeds. We show that B-spline-parameterized actions can be seamlessly integrated into standard policy learning pipelines by directly predicting B-spline parameters. Experiments on simulated and real-world tasks demonstrate that BSP significantly reduces task completion time, achieving substantial improvements over baseline methods while maintaining strong success rates. More results: https://b-spline-policy.github.io

Problem

• robot manipulation
• action representation
• task completion time
• policy acceleration
• trajectory execution

Innovation

• B-spline
• action representation
• robot manipulation
• continuous trajectory
• policy acceleration