Real-time cutter position-toolpath matching algorithm based on Hidden Markov Model for intelligent applications of machining process

Real-time monitoring of the actual position of the cutter during machining allows for the analysis of temporal variations in material removal rates, machining quality, and other factors. However, due to the influence of machine tool vibrations and control errors, the actual cutter position often deviates randomly from the theoretical toolpath. Simply using the nearest point search method makes it difficult to accurately find the correct matching toolpath segment, which affects the accuracy of the aforementioned analysis. In this paper, we propose an inference and matching algorithm based on the Hidden Markov Model (HMM) for matching the cutter's position with the corresponding toolpath segment. This algorithm functions as a navigation system, enabling real-time inference of the specific toolpath segment where the cutter should be positioned at any given moment during the machining process. We evaluated the accuracy of our algorithm using over 1.3 million cutter positions and 740,000 toolpath segments from real machining cases, achieving a high matching accuracy of 96.48%. Even at higher sampling intervals(1000-1500ms), the experimental results show that the algorithm's predictions remain reliable, with an accuracy exceeding 92%. Several potential intelligent applications based on this algorithm are also introduced.