Estimation Methods Research On Systematic Biases And Unknown Input Under Complex Systems

In the multi-sensor cooperative detection,tracking and recognition system,systematic bias registration is one of the key issues to be solved in order to obtain accurate target positioning.With the changing of strongly complex environmental conditions and offensive and defensive technologies,it is difficult to establish above-mentioned system's movement model and system deviation model.Especially,the system deviation presents unknown,random,and mutability,while the system is disturbed by unknown inputs from the outside.Therefore,this paper focuses on systematic bias estimation,unknown input estimation,and other issues.The main research innovations are as follows:Firstly,considering the influence of random input on the system,unknown interference action items are introduced in the traditional sensor measurement model,so that the measurement model mainly includes target motion state,systematic bias,unknown input and measurement noise,etc.However,it is difficult to establish the target motion model under complex factors.In order to obtain the system deviation pseudo-measurement model independent of the target state,the measurement in the state space is projected for the system deviation space.The above theoretical derivation and conclusions provide guarantees for the systematic bias and unknown input estimation methods proposed in this paper.Secondly,for the problem of systematic bias estimation under the condition that the target motion model is difficult to establish,the traditional method of augmented push is no longer applicable.Therefore,there is an unbiased estimator design method has been proposed for system bias under unknown inputs in this paper.At first,the systematic bias pseudo-measurement model is derived under the influence of unknown interference.At second,based on the systematic bias pseudo-measurement model and the known system bias dynamic model,an unbiased estimator of system bias is designed.At third,the three different scenes are analyzed,which are the same disturbance,different disturbances and disturbance degradation.Simulation results show that the systematic bias estimation method proposed in this paper is more feasible and universal than the traditional method.Finally,the measurement error caused by unexpected input from the outside cannot be solved by strategies including systematic bias registration and filtering.Therefore,an original method has been proposed for estimating the effect of systematic bias interference in this paper.On one hand,based on the theoretical derivation of this paper,the pseudo-measurement model of system bias is derived,and then the dynamic system of system deviation is constructed.On the other hand,a pseudo-measurement model of unknown disturbance is established based on the unbiased estimator designed in this paper.The difference between the predicted value and the measured value is compared with the set threshold value to determine the magnitude and duration of the unknown disturbance that is encountered in each stage of the measurement process.The simulation results show that the proposed algorithm can accurately determine the existence of unknown disturbances,besides estimate its size and determine its duration.Moreover,measurement registration has been further improved under the influence of unknown interference.