LOW-COST CUTTING TOOL DIAGNOSIS BASED ON SENSOR-FUSION

AbstractA monitoring system for the cutting tool condition for an industrial machining center is proposed. Four matching patterns and stochastic modelling approaches (Artificial Neural Network, Learning Vector Quantization, Support Vector Machine, and Hidden Markov Model) are compared for the diagnosis step. Integration of several sensor signals into a single fused estimation is considered. Several performance indexes such as binary and multiple classification, false alarm and false fault rate, and operating costs are considered for the comparison. Early results show that Hidden Markov Model-based approach fusing three sensors outperforms other techniques and exhibits 98% efficiency.