Real-time MPC with Mobility-enhanced Feature

Real-time Control Mobility Legged Robots MPC

Real-time NMPC scheme with a mobility-based cost function for terrain adaptation in legged robots.

Project Overview

This work introduces a Real-Time Nonlinear Model Predictive Control (NMPC) specifically designed for legged robots to achieve dynamic locomotion across various terrains.

Key contributions:

Mobility-Based Criterion: Maximizes leg mobility and improves adaptation to terrain features.
Real-Time Capability: Online re-planning at 25 Hz with a 2-second prediction horizon.
Efficient Model: Uses a single rigid body dynamic model for computational efficiency.

Read Paper