State Prediction of Industrial Robot Based on Hidden Markov Model

In view of the production process, it is necessary to master the running state of industrial robots in real time. Firstly, the current research status of equipment diagnostic prediction is investigated. Secondly, based on the hidden Markov model, the theoretical model of state prediction of industrial robot is derived. The model includes two calculations. The operation states of industrial robot are divided into three states. The three states are hidden states. The probability of occurrence of the explicit state current and voltage value in each hidden state accords with the mixture Gaussian distribution. Using EM algorithm, the iterative parameters of the observed state Gaussian mixture model of industrial robot are derived, and the observed state Gaussian mixture model is trained by using the historical current and voltage data of industrial robot joints. Finally, according to the current and voltage data collected at different times, the state prediction model of industrial robot is verified.