Grant: Collaborative Science, Technology and Innovation Program - Collaborative R&D Initiatives

In manufacturing projects that require a high degree of adaptability such as shipbuilding or energy, robots often fail to provide utility to the welder as they require significant time to program, are often stationary, and for safety must be isolated from nearby workers. To address these issues, this Project will introduce cobots to the large-scale welding job workflow through the adoption of vision systems combined with machine learning to identify the potential weld seams and adapt programming accordingly, providing a simple interface for the welding operator to assign tasks to the cobot. In addition, the cobot motion will be optimized by creating a library of robust parametric or micro-motions, collision avoidance planning, and resolved redundancy while avoiding singularities or cable wrapping. Through a simulation module, the operator will verify the planned trajectory and make any necessary corrections before the cobot executes the task.