Target state estimation method and system

Abstract

The invention discloses a target state estimation method and system. The method comprises the following steps: acquiring a communication topological graph of a distributed sensor network; obtaining anobservation model of a nonlinear system and prediction information of a sensor at the previous moment; calculating interaction information of the sensor at the current moment according to the observation model and the prediction information of the sensor at the previous moment, and enabling the interaction information of the sensor and the interaction information of a neighbor node sensor to interact with each other; according to the interaction information of the sensor and the interaction information of the neighbor node sensor, performing fusion by a weighted average consistency algorithmto obtain state estimation parameters of the sensor at the current moment, wherein the state estimation parameters of the sensor comprise a state estimation value and an error covariance estimation matrix of the sensor; and determining the state estimation parameter of the sensor at the current moment as the state estimation parameter of the target at the current moment. The method can reduce thecommunication burden of the network, guarantees that the filtering result cannot be diverged, improves the network communication efficiency, and improves the estimation precision of the target state.

Description

technical field [0001] The invention relates to the field of sensor networks, in particular to a target state estimation method and system. Background technique [0002] At present, for the problem of determining the state of the target, the more common solution is the centralized filtering method including the central node. This method uses the main sensor as the central node, and uses the central node to obtain the global information of the entire sensor network to realize the target state. estimate. In this centralized target state estimation method, when the number of sensor network nodes increases and the structure becomes complex, the amount of information acquired and processed by the central node becomes huge, which brings great challenges to communication and may lead to the loss of target estimation. Problems such as poor real-time performance, low efficiency, and even filtering divergence have great limitations in practical applications. [0003] In the past yea...

AI Technical Summary

Problems solved by technology

However, due to the limitation of filtering conditions, the consistency method cannot be directly applied to the distributed filtering of nonlinear systems.

Some scholars try to reconstruct the pseudo-measurement matrix, however, since the pseudo-measurement matrix is ​​obtained by statistical linear regression method, which is an approximate process, the result is not ideal

But if the approximation error is directly ignored, it may lead to different filter divergence

Images