Adaptive Finite Impulse Response Filter Based on Time Delay Neural Network

This paper discusses the design of an adaptive system for filtering primary navigation information obtained from MEMS sensors based on the approximation of a non-causal finite impulse response filter using a time delay neural network. Basic principles of finite impulse response filters are represented. Characteristics of the most widespread filters acceptable for processing data in MEMS devices are given. The approach to design of the adaptive non-casual filter using a time delay neural network are represented. Simulation results is given.