Collective Robot Reinforcement Learning with Distributed Asynchronous Guided Policy Search
This work addresses the problem of scaling robot reinforcement learning for real-world manipulation tasks by enabling multiple robots to share experience, though it is incremental as it builds on existing Guided Policy Search methods.
The paper tackles the challenge of training robot policies that generalize across diverse real-world conditions by proposing a distributed and asynchronous version of Guided Policy Search, demonstrating on a vision-based door opening task with four robots that it achieves better generalization, utilization, and training times compared to single-robot methods.
In principle, reinforcement learning and policy search methods can enable robots to learn highly complex and general skills that may allow them to function amid the complexity and diversity of the real world. However, training a policy that generalizes well across a wide range of real-world conditions requires far greater quantity and diversity of experience than is practical to collect with a single robot. Fortunately, it is possible for multiple robots to share their experience with one another, and thereby, learn a policy collectively. In this work, we explore distributed and asynchronous policy learning as a means to achieve generalization and improved training times on challenging, real-world manipulation tasks. We propose a distributed and asynchronous version of Guided Policy Search and use it to demonstrate collective policy learning on a vision-based door opening task using four robots. We show that it achieves better generalization, utilization, and training times than the single robot alternative.