Soft Actor-Critic Reinforcement Learning for Dual Robotic Arms Transporting Long Objects

With the rapid advancement of robotic technologies, robotic arms have become indispensable core components in industrial automation. However, collaborative control of multiple robotic arms in complex tasks remains a challenging issue. Among these tasks, long-object transportation is a typical collaborative scenario that requires robotic arms to maintain smooth object handling while optimizing temporal efficiency during operation. To address this issue, this paper proposes a reinforcement learning algorithm—Master-Slave Soft Actor-Critic (MSSAC)—which integrates a master-slave control architecture. The master arm makes real-time decisions based on global states, while the slave arm dynamically follows the master arm's guidance. Simulation results show that, compared to the conventional SAC algorithm, the MSSAC algorithm demonstrates superior performance in dual-arm collaborative transportation tasks, providing an innovative solution for multi-arm coordinated operations in complex logistics scenarios.