Railway Signal's De-noising Using Homomorphism Technology

With the advent of the railway transportation,high-speed and high-density train operations need faster and more reliable information-receiver. However,railway signal transmission channel is very bad,so the signal will inevitably be mixed with a variety of noise.We always assumed that these noises are additive.So far,the field of railway signals has not yet taken any effective measures to deal with non-additive noise.In this paper,according to homomorphic principle,we choose three realizable algorithms to wipe out the power-frequency noise from FSK signal.They are named: component-adding method;sub-frequency method;interval-overturn method.Further more,different algorithm needs different filter,and filter is designed in strictest way in interval-overturn method.About filter design,two filtering methods are chosen.One is the traditional FIR filter,the other is zero-setting filter.Each algorithm is discussed in three aspects,that is:signal noise added,multiple noises added,close to signal noise added.These algorithms are confirmed with MATLAB tool.The results show that they are effective to remove multiplicative noise,and can also achieve high accuracy.