Hidden Markovian Chain based occurrence time prediction for GSM-R down fault

GSM for Railways (GSM-R) is a wireless mobile network used for train-ground communication. Down is one of the main faults in the routine of train running. It is a procedure inherently composed of several degradation levels, which was formulated as a Hidden Markovian Chain (HMC). In the framework of HMC, a scheme based on viterbi algorithm and absorption probability was proposed to predict occurrence time of GSM-R down fault. Distinguished to its static version, there was a correction mechanism based on current GSM-R quality, which modified the prediction dynamically. The effectiveness of the scheme was testified by simulations, revealing that it possesses more accuracy than its static version and the deflection is restrained in a rational level.