A study of track initiation and data association for multiple targetting tracking

PDF

Authors
  1. Chen, J.
  2. Lo, T.
  3. Litva, J.
Corporate Authors
McMaster Univ, Hamilton ONT (CAN) Communications Research Lab;Defence Research Establishment Ottawa, Ottawa ONT (CAN)
Abstract
Multiple target tracking (MTT) is an esential requirement for surveillance system employing one or more sensor, together with computer subsystems, to interpret an environment that includes both true targets and clutters. With the recent proliferation and the increasing sophistication of surveillance systems, designers of radar and other systems are recognizing the importance of MTT. Therefore, a large amount of literature emerged in recent years. However, in most of the literature, simulated data were used to the MTT algorithms. Thus simplifications, such as that tracks already exist and that the dynamics are known, are assumed. In a real system where the only known information is the measurement returns, the capability of these algorithm becomes questionable. In the report, a MTT algorithm is developed. Following the discussion of the problem of multiple target tracking, a new track initiation algorithm is proposed and implemented in the system. This new initiation algorithm combines the idea of multiple hypothesis tracking with the standard Hough transform for track initiation. Only a few scans of radar returns are required.
Keywords
MTT (Multiple Target Tracking);Track initiation;Hough Transform;Data association;Nearest neighbor;Direction probability;Hopfield neural network;Bolxmann machine
Report Number
CRL-271;DREO-CR-93-654 — Technical Report
Date of publication
28 Jun 1993
Number of Pages
82
DSTKIM No
95-00863
CANDIS No
148588
Format(s):
Hardcopy;Document Image stored on Optical Disk

Permanent link

Document 1 of 1

Date modified: