Markov Chain Realization of Multiple Detection Joint Integrated Probabilistic Data Association.

Area of Science:

• * Signal Processing
• * Data Association
• * Target Tracking

Background:

• * Probabilistic Data Association (PDA) algorithms, like MD-JIPDA, are used for multi-target tracking.
• * These methods partition measurements into cells for track-to-measurement association.
• * Computational complexity increases exponentially with measurement cells and tracks, especially in dense or crossing target scenarios.

Purpose of the Study:

• * To introduce a novel algorithm, MD-MC-JIPDA, to address the computational challenges of existing multi-detection data association methods.
• * To reduce the computational cost associated with joint data association events in complex tracking environments.

Main Methods:

• * Proposed the multiple detection Markov chain joint integrated probabilistic data association (MD-MC-JIPDA) algorithm.
• * Utilized a Markov chain to generate random data association sequences, replacing traditional association event enumeration.
• * Focused on reducing computational load while maintaining tracking accuracy.

Main Results:

• * The Markov chain approach significantly reduces computational complexity compared to methods like MD-JIPDA.
• * The proposed MD-MC-JIPDA algorithm demonstrates effectiveness in experimental validation.
• * Achieved preferable tracking performance with substantially fewer generated association sequences.

Conclusions:

• * MD-MC-JIPDA offers an efficient solution for multi-detection target tracking problems.
• * The algorithm effectively mitigates the exponential growth in computational complexity.
• * Experimental results validate the superiority of MD-MC-JIPDA over existing algorithms in complex scenarios.