DynaFlow: Dynamics-embedded Flow Matching for Physically Consistent Motion Generation from State-only Demonstrations
This work addresses the challenge of bridging kinematic data to real-world execution for robotics, enabling deployable motion generation from state-only demonstrations, though it is incremental as it builds on existing flow matching and simulation techniques.
The paper tackles the problem of generating physically consistent motion from state-only demonstrations by introducing DynaFlow, a framework that embeds a differentiable simulator into flow matching to ensure dynamic feasibility, and demonstrates its effectiveness by deploying generated actions on a Go1 quadruped robot to reproduce gaits and execute long-horizon motions.
This paper introduces DynaFlow, a novel framework that embeds a differentiable simulator directly into a flow matching model. By generating trajectories in the action space and mapping them to dynamically feasible state trajectories via the simulator, DynaFlow ensures all outputs are physically consistent by construction. This end-to-end differentiable architecture enables training on state-only demonstrations, allowing the model to simultaneously generate physically consistent state trajectories while inferring the underlying action sequences required to produce them. We demonstrate the effectiveness of our approach through quantitative evaluations and showcase its real-world applicability by deploying the generated actions onto a physical Go1 quadruped robot. The robot successfully reproduces diverse gait present in the dataset, executes long-horizon motions in open-loop control and translates infeasible kinematic demonstrations into dynamically executable, stylistic behaviors. These hardware experiments validate that DynaFlow produces deployable, highly effective motions on real-world hardware from state-only demonstrations, effectively bridging the gap between kinematic data and real-world execution.