Distributed Fusion Estimation for Multi-Sensor with Bounded Noise

In practical applications, sensors may encounter disturbances that result in the measurement of outliers, leading to abnormal data. Therefore, when performing distributed fusion estimation for time-varying discrete systems with multiple sensors, it is assumed that the noise is bounded and contains outliers, and the goal is to estimate the true state of the system. The improved set-membership filtering algorithm is used to filter the bounded noise with outliers. The design of local filters involves transmitting the local estimates from each sensor to the fusion center. Considering the integrity and accuracy of information, an improved Minkowski sum algorithm is adopted to achieve global estimation. Finally, the effectiveness of the proposed multi-sensor robust distributed fusion is verified through a simulation example.