Judo: Sampling-Based Model Predictive Control Simulator

This paper introduces Judo, a simulator for sampling based MPC (Model Predictive Control) for robotic control planning. The main takeaways here are that Judo provides standardization for MPC based simulation, built on-top of Mujoco. Additionally it offers implementations of several MPC algorithms including: - Predictive Sampling - CEM (Cross-Entropy Method) - MPPI (Model Predictive Path Integral) In general, it's important to note that a sampling-based Model Predictive Controller (MPC) is a type of MPC that solves the optimal control problem by generating and evaluating multiple sampled trajectories (or control sequences) rather than using gradient-based optimization or solving a quadratic program directly.