卡爾曼預估器與多目標追蹤法則

Kalman Predictor and Multitarget Tracking Algorithm

周瑞雄(Jui-Hsiung Chou) ； 陳梧桐(Wu-Ton Chen) ； 陳春明(Chuen-Ming Chen) ； 孫惠民(Huey-Min Sun)

《中華技術學院學報》 26期 (2003/04) Pp. 22-37

https://doi.org/10.7095/JCIT.200304.0022

摘要

本文目地在於針對多目標追蹤提出一些適合在假信號環境中從事對多目標作起始及追蹤的方法。首先建立一般性的多目標追蹤模型，並以卡爾曼預估器，探究其優劣性，作為發展新演算法之參考。為處理因多目標所造成龐大的計算問題，將整體目標群與量測回訊集合分成數個集串，使龐大的計算問題變成為個別獨立的小問題。對於實際雷達訊號所偵測的飛行物與一般多目標模型之相關性將深入研究並建立新適用模型；根據所建立的新模型，推導出有效的新型動態追蹤演算法。
最後，使得偵測多目標過程中的微小訊號能夠清楚呈現，且能將消除雜訊之預估器及平滑器，建立於新模型成為另一新型演算法則，建立一套系統化的螱目標追蹤模式。

關鍵字

卡爾曼預估器 ； 多目標追蹤法則

並列摘要

The purpose of this paper is to develop Kalman predictor algorithm used to the multitarget tracking problem and track initiation procedure in a cluttered environment. The useful multitarget tracking model is investigated and developed. The existed algorithms are studied to have the well-known techniques in hand. Instead of solving a large problem, the entire set of targets and measurements is divided into several clusters such that a number of smaller problems are solved independently. Moreover, the relationship between the real radar signal and the signal in the general model is given. According to the modified results, a new dynamic model can be erected to fit the real implementation. Furthermore, some new tracking algorithms are developed for the new dynamic model.
Finally, computer simulation results will be given the proposed algorithms that can be usefully applied in the real tracking.

並列關鍵字

Kalman predictor ； Multitarget tracking algorithm