Machine learning approach for robot diagnostic system

This study introduces the machine learning (ML) approaches for robot diagnostic system (RDS). The feature algorithm is implemented on robot diagnostic system using an acoustic filtering technique. It is performed on the environment of industrial embedded compact-RIO (ECRIO). The fault diagnostic process is based on the machine learning algorithms by using acoustic feature information for dynamic arm of robot. Acoustic signals are acquired from array microphone of industry. Experimental investigations were carried out for a practical robot arm system. The effectiveness of the proposed system on the fault features is classified by using the multiple layer perceptron (MLP). A smart fault prediction for robot arm was carried out so that it can distinguish the feature differences of our proposed conditions. For classification and identification of robot, we can adopt the MLP method of ML algorithm. Implementation results on GPU processors can be discussed.