Low Observable Moving Target Tracking Based on Modified PDA-AI
- PDF / 230,009 Bytes
- 12 Pages / 439.37 x 666.142 pts Page_size
- 10 Downloads / 220 Views
Low Observable Moving Target Tracking Based on Modified PDA-AI Zhengzhou Li & Guoping Li & Ju Tan & Fengcun Tian & Gang Jin & Youcheng Ren
Received: 23 June 2009 / Accepted: 13 July 2010 / Published online: 1 August 2010 # Springer Science+Business Media, LLC 2010
Abstract During tracking low observable moving target in electro-optical (EO) imaging tracking system, multiple false alarms resulted from low signal-to-noise rate (SNR) would seriously debase the performance of target tracking. Probabilistic data association with amplitude information (PDA-AI) assumes that amplitude of target is not correlative among different sampling instants and larger amplitude is, greater the probability of validated measurement being target of interest would be. In EO imaging tracking system, amplitude and motion of target of interest are consistent and highly correlative. A modified PDA-AI is discussed and developed to resolve the inconsistency between PDA-AI and EO tracking system in this paper, which analyzes target motion by means of modeling amplitude and motion as well as their consistency as stationary random signal. The theory analysis with Cramer-Rao lower bound (CRLB) and experiments results show that the performance of low observable target tracking of the modified PDA-AI would be more accurate and more reliable than that of the original PDA-AI. Keywords Low observable moving target tracking . Target motion analysis . Probabilistic data association . Amplitude information . Cramer-Rao lower bound
1 Introduction In recent years, low observable moving target tracking has been widely studied in these systems such as infrared detector, radar and sonar [1–3]. Multiple measurements would be detected in validation region for low detection probability and high false alarm probability. These false alarms would largely increase the uncertainty of target recognition, and greatly debase the reliability and precision of target tracking. Z. Li (*) : G. Li : J. Tan : F. Tian Communication Engineering College of Chongqing University, Chongqing 400044, China e-mail: [email protected] J. Tan Chongqing University of Arts and Sciences, Chongqing 402160, China G. Jin : Y. Ren China Aerodynamics Research and Development Center, Mianyang 621000, China
1246
J Infrared Milli Terahz Waves (2010) 31:1245–1256
During tracking such low observable moving target with less-than-unity probability of detection and high probability of false alarm, data association—deciding which of the received multiple measurements to use to update each track—is fairly crucial. A number of algorithms have been developed to track such target, and they could be mainly classed as three kinds so far, namely, nearest-neighbor association, splitting tracking association [4–6] and probabilistic data association (PDA) [7, 8]. The original PDA takes all of the validated measurements into account with different weights, and it could overcome the deficiency, such as low association efficiency and combination blast caused by nearest-neighbor association and splitting track
Data Loading...
