Research on Multi-Agent State Estimation Algorithm under Intermittent Observation

In this paper, state estimation based on multi-agent theory under intermittent observation is studied. In practice, multi-agent systems often suffer from loss of measurement data. For the case of episodic data loss, the statistical properties of the data before and after the loss are used to detect the loss and reconstruct the measurement function to complete the estimation under intermittent observations; for the case of long-term data loss, the time-varying weighting method is used to reduce the importance of the problem nodes and minimize the global estimation error. The effectiveness of the algorithm is demonstrated by experiments on the dynamic estimation of the target object.