Methods And Applications Of Measurement Fusion In Multi-Radar System

Multi-radar data fusion is a new inter-disciplinary research technique which has been developing rapidly in recent years. Compared with a single radar at target detection, tracking, status identification, situation analysis and threat assessment, etc., multi-radar can effectively enhance the viability and stability of systems, increase the reliability of the system's output result, improve a stabilized and continuous capability of tracking maneuvering or high-speed targets, and also have an advantage of precision, time and space coverage, spatial resolution, information utilization and other aspects.The basic principles and methods of multi-radar data fusion are introduced in paper including the relative knowledge of space aligning, time aligning, target tracking, fusion estimating. The theory of Kalman filter is particularly described with the help of simulation experiments in the target tracking section. The fusion estimation section mainly introduces two ways on track fusion and measurement fusion, and highlights the practical use and its advantages.Firstly, the introduction is made, in brief, for the definition, basic principles, functional model and the integration level of the multi-radar data fusion. The target tracking technology is explained thoroughly. The Kalman filter and some tracking approaches under the nonlinear condition, as well as their simulation work, are discussed in detail. Space aligning is completed by coordinate conversion, converting various radar data into a unified coordinate reference system. What's more, to solve the problem of different boot time, different timeline and different sampling period, data interpolation and curve fitting are carried out to achieve time aligning. Data association aims at confirming the relationship between the measured value and the target set according to the associated door.Secondly, the track fusion technology is introduced, including the basic concept and the relationship between the track fusion and the function model and structural model of data fusion. And also, the structure of track fusion process is mentioned. Then the introduction and simulation work is made on track fusion algorithms, such as simple convex combination, covariance intersection algorithms, federated filter, of which some simulations are implemented.Finally, the measurement fusion technology is emphasized. The method for sampling radar data consists of uniform sampling and non-uniform sampling. After aligning the points, there are two ways to deal with data, serial process and parallel process. Then, some approaches of measurement fusion are explained in detail, including plot serial consolidation method, least squares method, parallel filter, sequential filter. Besides, the comparison is made between parallel filters and sequential filters by simulating. After that, measurement fusion algorithm is validated in measured data, the result of which is compared with the results of track fusion and single radar tracking. It can be concluded that measurement fusion has non-negligible advantages over other methods.