Parameterized Grinding Planning of Robotic Arms Based on Optimization Experiences Generalization for Tubular Workpieces

In order to make industrial robotic arms adaptively accomplish the uniform grinding tasks for tubular workpieces of different sizes, spatial orientations and surface characteristics, a parameterized planning framework is proposed in this paper. In this framework, Moving-Circle-Determination (MCD)-based geometric parameter calculation is designed for two types of tubular workpieces (both elbow and tee tube). Then the Adaptive Uniform Grouping (AUG) mechanism is designed and applied in path cost calculation module to make all points evenly grouped, and results in a minimal computational complexity of optimization algorithm in the framework. And then two-stage Progressive Optimization (TPO) algorithm, containing both Experience Generalization Ant Colony System (EGACS) and Double Deep Q-Network (DDQN), is designed to calculate the optimal grinding order in the maximal convergence rate, where optimization capabilities of both EGACS and DDQN are enhanced with accumulation of uniform grinding task based on optimization experiences generalization mechanism applied in TPO algorithm. Finally, Parameterized Velocity Vector Summation (PVVS) is designed and applied in the path planning and tracking control module to drive robotic arms to accomplish uniform grinding tasks. Compared to the state-of-the-art methods and algorithms, performance advantages of the planning framework proposed are at least 16.94%, 21.60%, 28.78%, 4.59% and 48.68% in terms of size estimation, spatial orientation estimation, convergence rate, solution quality and computational complexity respectively. Note to Practitioners—This paper was motivated by the grinding path planning problem in uniform grinding tasks for tubular workpieces, containing both tee tubes and elbows. In the existing related researches, grinding planning of elbows is rarely focused, in addition, experience accumulated in grinding process has not been effectively utilized in new grinding tasks. Therefore, in this paper, a new planning method is suggested to not only generate grinding paths for both tee tubes and elbows but also make all grinding experience accumulated till now used in new grinding tasks by combining deep reinforcement learning and the heuristic optimization algorithm. Based on this design, the ability of this suggested planning method will gradually enhance with increase of completed grinding tasks, just like humans gradually become proficient through continuous practice. In this paper, details of this planning method are provided in the form of mathematical formulas and algorithm processes, then performance advantages of the suggested method are proved by the theoretical proofs and comparative experiments, and the latter demonstrates efficiency and effectiveness of suggested method in practical grinding tasks. In our future research, complexity of current planning method will be simplified by introducing reusable modular approach for both overall architecture and detailed designs, while application scenario of current planning method will be further expanded to more types of workpieces, such as a tubular workpiece with an elliptical cross-section.