Research On Filtering Algorithm Based On Event Driven

Filtering strategies play an important role in estimation theory,and are used to extract knowledge of the true states typically from noisy measurements or observations made of the system.The name ‘filter'is appropriate since it removes unwanted noise from the signal.The state estimation of nonlinear systems has caught the focus of many researchers,and becomes a hot research field with great theoritical value and application field.However,to achieve the requirement of high estimation accuracy,the sampling points need to be large,and the calculation amount increases correspondingly.According to this problem,we need to accord the selected event to choose some points which contain useful information to update the state.The selected step is becoming a hot and efficient method to study the estimation problems for the current and future.Based on the previous work of other researchers,by studing the limitations of the Kalman filter and particle filter,we implement some researches and improvement in some issues that we are facing in the field of state estimation.The main contributions of this thesis are described as follows:(1)A square-based sampling algorithm and a modified Kalman filter based on this sampling strategy have been presented.The square-based sampling,which can get more important information with less sampling points,is more effective than the periodic sampling.The stability of the algorithm is analyzed in the last.The effectiveness of the algorithm is illustrated through the theoretical proof and simulation.(2)Aiming at the out-of-sequence measurements(OOSMs)problem,an algorithm based on information filter was proposed.By introducing a selective threshold(ST)to determine the value of the OOSMs,the estimate only update with the useful OOSMs.By the simulation,we show that selective fusion can reduce computational costs while maintaining a good performance.(3)An improved particle filter based on pearson correlation coefficient(PPC)is proposed to reduce the particle degeneracy and sample impoverishment.The PPC is adopted to determine whether the particles close to the true states.Finally,the stability of the algorithm is analyzed and some simulations are provided to illustrate the effectiveness of the proposed filter.(4)A new particle filter based on smooth variable structure(SVSF-PF)was introduced.This method makes use of an existence subspace as a smoothing boundary layer to keep the particles bounded within a region of the true state trajectory.This creates a robust and stable estimation strategy,which can improve the effect of particle degeneracy and sample impoverishment.At last,the simulation results demonstrate the better performances of the proposed robust filter.