Research On Particle Filter Track-Before-Detect Technology Based On Waveform Agile

Today in twenty-first Century, radar has been widely used in military and civil fields. However, radar still exists many problems for weak target detecting and tracking such as stealth targets or targets in heavy clutter. The track-before-detect technology based on particle filter(PF-TBD) can effectively improve the radar performance. But PF-TBD algorithm look for the optimal effect only from the data processing way, and the target tracking accuracy is related with not only the method of data processing, but also the ways of information processing, such as the selection of the signal waveform and its parameters. Therefore, the combination of signal processing and data processing is one of the research directions of the current to improve the accuracy of target tracking. Using waveform agile technology, radar can keep learning from the environment, and adaptively design the waveform of the next time according to the environment information. It makes radar form a closed loop feedback system to improve performance. The dissertation focuses on the problem of weak target detection and tracking, which studied the target tracking technology based on waveform agile. It mainly comprises:1. Established the TBD model based on grey-scale map, and analyzed the single frame data under the models by simulation. Then studied SIR-TBD and ESIR-TBD algorithm respectively, and simulated the two algorithms. The result shows that the ESIR-TBD algorithm is better than SIR-TBD algorithm.2. Aiming at the problem that PF-TBD is not out of the traditional radar framework, studied a waveform agile method based on PF-TBD. This method make ambiguity function as a bridge to establish relationship between tracking accuracy and waveform parameters base on Cramer Rao Lower Bound(CRLB) evaluation criteria to obtain the criterion function. At last used particle filter track-before-detect method to adaptively design the transmitting waveform by minimizing the predicted mean square error(MSE) to promote radar performance.3. Studied a target tracking algorithm based on waveform agile in clutter. First of all, established the dynamic model in the 2-D plane, nonlinear measurement model and clutter model, and then used particle filter and probabilistic data association algorithm for target tracking in clutter. At last, based on unscented transformation and the CRLB of the covariance of measurement noise, minimized the cost function to choose the optimal waveform parameters in order to improve the accuracy of target tracking.4. Aiming at the problem that the performance evaluation criterion for Track-before-Detect(TBD) is not united and comprehensive, a Joint Detecting and Tracking performance evaluation method is proposed. The method can evaluate and compare the performances of PF-TBD algorithm, WD-PF-TBD algorithm and DBT algorithm effectively in the respects of target and trajectory detection performances.