Local Manifold Approximation and Projection for Manifold-Aware Diffusion Planning
This work addresses safety-critical applications in robotics and autonomous systems by improving the reliability of diffusion-based planning, though it is incremental as it builds on existing diffusion methods.
The paper tackles the problem of unreliable trajectory generation in diffusion-based planning for long-horizon tasks by addressing inaccurate guidance during sampling, proposing LoMAP to project samples onto a low-rank subspace to prevent infeasible trajectories, and validates it on offline RL benchmarks with performance enhancements.
Recent advances in diffusion-based generative modeling have demonstrated significant promise in tackling long-horizon, sparse-reward tasks by leveraging offline datasets. While these approaches have achieved promising results, their reliability remains inconsistent due to the inherent stochastic risk of producing infeasible trajectories, limiting their applicability in safety-critical applications. We identify that the primary cause of these failures is inaccurate guidance during the sampling procedure, and demonstrate the existence of manifold deviation by deriving a lower bound on the guidance gap. To address this challenge, we propose Local Manifold Approximation and Projection (LoMAP), a training-free method that projects the guided sample onto a low-rank subspace approximated from offline datasets, preventing infeasible trajectory generation. We validate our approach on standard offline reinforcement learning benchmarks that involve challenging long-horizon planning. Furthermore, we show that, as a standalone module, LoMAP can be incorporated into the hierarchical diffusion planner, providing further performance enhancements.