Decision making for multi-robot fixture planning using multi-agent reinforcement learning

Within the realm of flexible manufacturing, fixture layout planning allows manufacturers to rapidly deploy optimal fixturing plans that can reduce surface deformation that leads to crack propagation in components during manufacturing tasks. The role of fixture layout planning has evolved from being performed by experienced engineers to computational methods due to the number of possible configurations for components. Current optimisation methods commonly fall into sub-optimal positions due to the existence of local optima, with data-driven machine learning techniques relying on costly to collect labelled training data. In this paper, we present a framework for multi-agent reinforcement learning with team decision theory to find optimal fixturing plans for manufacturing tasks. We demonstrate our approach on two representative aerospace components with complex geometries across a set of drilling tasks, illustrating the capabilities of our method; we will compare this against state of the art methods to showcase our method’s improvement at finding optimal fixturing plans with 3 times the improvement in deformation control within tolerance bounds. Note to Practitioners —Fixture layout planning is one of the most fundamental manufacturing tasks that must be carried out before production cycles can begin, to ensure that deformation is within tolerances to avoid crack propagation and component damage. However, reliance on humans to generate these plans has led to sub-optimal solutions, leading to manufacturers incurring losses and wanting to explore analytical methods for fixture planning. In this vein, there may be the temptation to find a single solution that can be applied to all problems regardless of complexity. However, in the age of flexible manufacturing, the benefits of building tailored solutions within a framework -referred to as “freedom within a framework” -become more apparent. This paper outlines the framework for multi-agent reinforcement learning for fixture layout planning, and demonstrates the capabilities of this framework to outperform current state of the art methods with a simple algorithm through extensive experiments on representative aerospace components. We provide code implementations of our work on GitHub (Multi-Agent Fixture Planner on GitHub).