Toward Single-Step MPPI via Differentiable Predictive Control
This addresses real-time implementation issues in model predictive control for robotics or autonomous systems, though it is an incremental improvement over existing MPPI methods.
The paper tackled the computational and tuning challenges of Model Predictive Path Integral (MPPI) control by proposing Step-MPPI, a framework that learns a neural sampling distribution for single-step lookahead, achieving millisecond-level latency while maintaining long-horizon foresight across multiple challenging tasks.
Model predictive path integral (MPPI) is a sampling-based method for solving complex model predictive control (MPC) problems, but its real-time implementation faces two key challenges: the computational cost and sample requirements grow with the prediction horizon, and manually tuning the sampling covariance requires balancing exploration and noise. To address these issues, we propose Step-MPPI, a framework that learns a sampling distribution for efficient single-step lookahead MPPI implementation. Specifically, we use a neural network to parameterize the MPPI proposal distribution at each time step, and train it in a self-supervised manner over a long horizon using the MPC cost, constraint penalties, and a maximum-entropy regularization term. By embedding long-horizon objectives into training the neural distribution policy, Step-MPPI achieves the foresight of a multi-step optimizer with the millisecond-level latency of single-step lookahead. We demonstrate the efficiency of Step-MPPI across multiple challenging tasks in which MPPI suffers from high dimensionality and/or long control horizons.