Research On The State Fusion Estimation Algorithm For Multi-rate Sensor Systems

With the rapid development and widespread application of computer networks,the traditional physical transmission mode can no longer meet the growing demand of system functions.Therefore,it is urgent to develop more efficient transmission modes.The Internet of Things(Io T)based on information perception connects components with different functions through the internet.It can realize the information transmission and interaction between things and people,things and things,and ultimately achieve a high degree of cognition and intelligent control of the physical world.To estimate the state of the target accurately,Io T relies on multiple autonomous sensors distributed in space to observe the target at different sampling rates.The information collected by the sensors is correlated,combined and integrated for further situation assessment,decision-making and control.Meanwhile,when the data is transmitted through the internet,the limited network carrying capacity and communication bandwidth will cause many networked uncertainties,such as data loss,time delay,communication interference,etc.These uncertainties will reduce the estimation accuracy and make many assumptions in traditional estimation theory difficult to be applied to multi-rate sensor systems.Therefore,it is very crucial to investigate the state fusion estimation algorithm of multi-rate sensor systems.In this thesis,the sensor system in Io T has been selected to be the research object.The state estimation under multi-rate sampling and networked uncertainties is chosen to be the research target.The fusion architecture is set to the main research threat.The state fusion estimation algorithm of multi-rate sensor systems and systems with uncertainties are studied.The research content mainly includes the following aspects.(1)For the multi-rate sensor system in which multiple sensors sample at different rates in Io T,a state space model is established according to the sampling mechanism.According to the characteristics that the measurements arrive at the fusion center in turn,the fusion architecture of the measurements in the fusion estimation is studied.Based on the linear Kalman filtering,a sequential fusion estimation algorithm in terms of linear minimum variance is proposed.The example of integrated navigation system verifies the good performance of the algorithm in state estimation of multi-ate sensor system,and the superiority in estimation accuracy and computational burden is demonstrated.(2)On the basis of multi-rate sampling,the randomly missing of measurements is considered,where measurements from each sensor are missing stochastically with certain probabilities.A multi-rate sampling system model with random parameters is established.The processing method of missing measurements is investigated.Based on the Kalman filtering,a sequential fusion estimation algorithm in terms of linear minimum variance is proposed.An example of an integrated navigation system where the measurements are randomly missing demonstrates the effectiveness and feasibility of the proposed sequential fusion algorithm.(3)On the basis of multi-rate sampling,the random delay of measurements is considered.The time-delay variable and random parameters are introduced into the system model,so the system model with random delay is established.The method of model reduction to eliminate time-delay variables is studied,and the noise correlation introduced by model reduction is analyzed.A sequential fusion estimation algorithm in terms of linear minimum variance is proposed.Finally,the simulation results of an integrated navigation system with random delay prove that the proposed sequential fusion algorithm can deal with the state fusion estimation problem of random delay of multi-rate sensor systems effectively.