Accelerated gradient descent for high frequency Model Predictive Control

TL;DR

This work studies the potential effectiveness of first-order methods and shows on a torque controlled manipulator that they can equal the performances of second-order methods.

Abstract

The recent promises of Model Predictive Control in robotics have motivated the development of tailored second-order methods to solve optimal control problems efficiently. While those methods benefit from strong convergence properties, tailored efficient implementations are challenging to derive. In this work, we study the potential effectiveness of first-order methods and show on a torque controlled manipulator that they can equal the performances of second-order methods.

Figures (3)

• Figure 1: Circle tracking in the end effector space under external disturbance
• Figure 2: Comparison of the running cost for the first-order (AGD) and second-order method (DDP)
• Figure 3: End-effector positions of the robot for the first-order (AGD) and second-order method (DDP)