Towards AI-driven task and motion planning of robotic assembly operations

Abstract

Nowadays, the integration of artificial intelligence (AI) technologies is playing an increasingly important role in robotic assembly operations enhancing aspects related to dynamic planning as well as human-robot collaboration. Under this scope, the proposed manuscript is focused on the creation of an AI-based framework that can coordinate both task and motion planning under assembly operations. This framework includes an AI Task Planner for dynamic task allocation to the available resources and a Neural Network Motion Planner Selector (NN-MPS) for effective decision-making related to robot movements. Thus, the correct motion strategies for robots are being established in order to achieve smooth and accurate operations in dynamic conditions. The proposed solution is being validated under a real case study derived from the automotive industry, where both robust and accurate planning is required towards the assembly of high payload parts.