In this paper

some signal processing technique for fault diagnosis and its application in gear systems are discussed. These technique includes: stationary signal processing

non-stationary signal processing

non-Gauss and colored noise processing

non-linear signal processing

blind signal separation

singular value decomposition and intelligent diagnosis technique. The characteristic

range of application

and research progress are analyzed and compared

included some developing direction in the future. These results could offer theoretical foundation for fault diagnosis and on-line monitoring of gear system

which are useful to secure running of gear system.