GeneWorker: An end-to-end robotic reinforcement learning approach with collaborative generator and worker networks

Hao Wang, Hengyu Man, Wenxue Cui, Riyu Lu, Chenxin Cai, Xiaopeng Fan

Abstract: Highlights•GeneWorker can achieve a mean success rate of over 90.67% on continuous robotic tasks.•GeneWorker outperforms previous state-of-the-art methods by a minimum of 54% on the pick-and-place task.•GeneWorker can be successfully extended to discrete control tasks with excellent performance.•GeneWorker has the potential to be integrated with emerging techniques such as meta-learning and large language models.