Research Analyzer
← Back ICRA 2026

Feedback-MPPI: Fast Sampling-Based MPC Via Rollout Differentiation � Adios Low-Level Controllers

Tommaso Belvedere, Michael Ziegltrum, Giulio Turrisi, Valerio Modugno

PDF

AI summary

Key figure (auto-extracted from paper)
Augmenting sampling-based MPPI with rollout-derived feedback gains enables high-frequency, stable closed-loop control without full re-optimization.
Model Predictive Path Integral sampling-based control rollout differentiation real-time MPC dynamic robotics

Problem

Standard MPPI control suffers from high computational latency, making it unsuitable for real-time, high-frequency control of fast, nonlinear robotic systems that require rapid state corrections.

Approach

The method computes local linear feedback gains by differentiating MPPI rollout costs with respect to the initial state, enabling rapid closed-loop corrections around the current state without re-solving the full optimization problem.

Key results

  • Novel rollout differentiation framework for computing local MPPI feedback gains
  • Improved control accuracy and action smoothness via high-frequency feedback
  • Successful quadruped locomotion over rough terrain in simulation
  • Real-world aggressive quadrotor maneuvers with onboard computation

Why it matters

Bridges sampling-based and gradient-based MPC, enabling robust, high-frequency control on resource-constrained embedded hardware for complex dynamic robots.

Abstract

Model Predictive Path Integral control is a powerful sampling-based approach suitable for complex robotic tasks due to its flexibility in handling nonlinear dynamics and non-convex costs. However, its applicability in real-time, high-frequency robotic con- trol scenarios is limited by computational demands. This paper introduces Feedback-MPPI (F-MPPI), a novel framework that augments standard MPPI by computing local linear feedback gains derived from sensitivity analysis inspired by Riccati-based feedback used in gradient-based MPC. These gains allow for rapid closed-loop corrections around the current state without requiring full re-optimization at each timestep. We demonstrate the effective- ness of F-MPPI through simulations and real-world experiments on two robotic platforms: a quadrupedal robot performing dynamic locomotion on uneven terrain and a quadrotor executing aggressive maneuverswithonboardcomputation.Resultsillustratethatincor- porating local feedback significantly improves control performance and stability, enabling robust, high-frequency operation suitable for complex robotic systems. IndexTerms—Optimizationandoptimalcontrol,motioncontrol, legged robots, model predictive control.

Index terms

Optimization and Optimal Control Motion Control Legged Robots

Related papers