Abstract:

This work presents a detailed overview of control systems for coordinated locomotion and manipulation—also known as loco-manipulation—in quadruped platforms. The review categorizes control strategies into three main approaches: model-based, optimization-based, and learning-based. It begins with a comprehensive mathematical formulation of leg-arm quadruped systems, including floating-base dynamics and contact constraints. Subsequently, it examines various control frameworks such as inverse dynamics-based whole-body control, model predictive control (MPC), and reinforcement learning (RL)-based techniques. The importance of accurate dynamic modeling is emphasized, particularly in model-based and optimization-centric approaches. Learning-based and hybrid controllers are assessed in terms of generalization, adaptability, and data efficiency. The work concludes by summarizing key computational concepts and practical insights, offering a roadmap for future research in robust and adaptive loco-manipulation.

Speaker Bio:

Mr. Md Hafizur Rahman is currently pursuing his M.S. degree in the Department of Control and Instrumentation Engineering at King Fahd University of Petroleum and Minerals (KFUPM), Saudi Arabia. He received his B.Sc. degree in Mechatronics Engineering from Rajshahi University of Engineering & Technology (RUET), Bangladesh, in 2022. From 2018 to 2022, he was an active member of a Formula Student electric vehicle project, contributing to the electrical subsystem. His current research focuses on control strategies for loco-manipulation in legged robots and autonomous decision-making systems..