Multi-sensor data fusion based on electro-optical tracking system

Abstract

The modern electro-optical (EO) measurement and tracing system usually has different kind of EO sensors that can get measurement of the target at the same time. The traditional tracking system of electro-optical measurement device mainly relies on single sensor that has better precision, and switches to other sensors while tracking is unstable. However, it brings on unstable tracking performance and dramatic decrease of tracking precision while switching among different sensors that have different measurement accuracy. To overcome the disadvantage of the conventional tracking method, a federated Kalman filter algorithm based on multi-sensor fusion of several sensors in EO tracking system is proposed. The local filters deal with the target track from each sensor, then the synthetic target estimation is made according to the sensor fusion of local estimation. Then the weight factor is assigned to each local filter using the fuzzy logic technique by analyzing the covariance of target estimation. The weight factor represents the degree of performance of the local estimation and shows the relationship between each sensor. The simulations with Monte Carlo methods show that the proposed algorithm is effective in optoelectronic tracking system especially in the situation of switching among different sensors, and the precision is also superior to the traditional single sensor tracking algorithm.