We propose SAIL (Speed-Adaptive Imitation Learning), a framework for enabling faster-than- demonstration execution of policies by addressing key technical challenges in robot dynamics and state-action distribution shifts. Offline Imitation Learning (IL) methods such as Behavior Cloning (BC) are a simple and effective way to acquire complex robotic manipulation skills. However, existing IL-trained policies are confined to execute the task at the same speed as shown in the demonstration. This limits the task throughput of a robotic system, a critical requirement for applications such as industrial automation. SAIL features four tightly-connected components: High-gain control to enable high-fidelity tracking of IL policy trajectories, consistency-preserving trajectory generation to ensure smoother robot motion, adaptive speed modulation that dynamically adjusts execution speed based on motion complexity, and action scheduling to handle real-world system latencies. Experimental validation on six robotic manipulation tasks shows that SAIL achieves up to a 4× speedup over demonstration speed in simulation and up to 3.2× speedup on physical robot.
Stacking Cups to Paramid Shape
Wiping Board
Baking: pick up bowl and put them in the oven
Folding Cloth
(a) Policy Level: Starting with synchronized observations (Obs. sync) from robot state
and camera inputs, the system generates (1) temporally-consistent action predictions through action-conditioned CFG and (2) time-varying action interval δt.
(b) Controller Level: The predicted actions are scheduled for execution while accounting for sensing-inference delays, with outdated actions being discarded. The
scheduled actions are executed using a high-gain controller with velocity feedforward (Vel FF) terms to track trajectory at the specified time parametrization.
Challenge 1: Divergence of Consecutive Action Prediction Diffusion policy would produce diverging action predictions, which is harmful for high-speed execution. We propose a novel action-conditioned CFG to generate temporally-consistent action predictions.
Challenge 2: Adaptive Speed Modulation Motion segments that are nonlinear and require precise control should not be sped up. Based on the motion complexity, we adjust the execution speed adaptively.
Challenge 3: Controller Behavior Shift
Challenge 4: System Latency in Control Loop System latencies caused by communication and control loop can lead to out-of-distribution inputs to the policy and time-misaligned action commands to the controller. To address this, we propose a novel action scheduling mechanism to handle real-world system latencies.
TODO: add:
TODO: add:
Adaptive speed modulation detects the motion complexity and adjusts the execution speed accordingly. Policy execution is only sped up when the motion is linear and smooth. Execution is slowed down for complex motions.
TODO: add simulation introduction, put tables and numbers of real and sim experiments here
TODO: add limitations
TODO: add acknowledgement
placeholder 7