Suboptimal Fusion Filter for Dynamic Systems

R. Minhas and V. Shin (Korea)


Dynamic system, Kalman filter, Suboptimalfilter, Data fusion1. NTRODUCTIONdata fusion system is needed to estimate the state of system,based on information from sensor measurements. Theultimate target of well designed system is large gain ofinformation with minimized and real-time processing. Theintegration and fusion of information is used in design ofhigh-accuracy control systems. Multiple sensors in a systemare used to attain the higher level of accuracy in real-timeenviro


The filter robustness in multi-sensor dynamic system is mainly dependent upon communication, synchronization, data association, correlation, and estimation errors. Estimation errors can also become worse in the presence of bad initialization, modeling errors and assumptions made. A new reduced order robust suboptimal filter is proposed for multi-sensor dynamic systems which reduces the computational cost for state estimation. The new filter has parallel structure and is very suitable for parallel processing of observation which can also be helpful to minimize the computation time. The examples demonstrate the fusion of state estimates based on observations through several sensors with the effect of common process noise.

Important Links:

Go Back