Content Details | NUML Online Research Repository

Content Details

Back to Department Listing

Title MULTI OBJECT TRACKING USING NON-LINEAR FILTERING TECHNIQUES

Author(s) Ali Raza Malik

Abstract In this research, the problem of redundant detections of tracks in distributed sensor networks for multi-object tracking is addressed. Measurement noise and network-induced errors often result in multiple detections of the same object, complicating the accurate estimation of object count and position. This makes it challenging to accurately determine the true number of objects and their locations. Track-to-track association algorithms help address this issue. Many such algorithms have been developed and can be broadly categorized into two types: statistical algorithms and clustering-based algorithms. A key clustering-based approach is the fuzzy track-to-track association algorithm, which is the focus of this research. A variation of this algorithm is tested on data generated from a model simulating a multi-sensor, multi-target environment. In real-world sensors, errors typically arise in azimuth, elevation, and range, so this thesis proposes an error model based on these parameters. The association algorithm’s resolutions are also grounded in this realistic error model. Additionally, time synchronization is critical before performing track association. This thesis employs a linear predictor to synchronize tracks before association, and the performance of the algorithm is analyzed under these conditions.

Type Thesis/Dissertation MS

Faculty Engineering and Computer Science

Department Engineering

Language English

Publication Date 2025-06-23

Subject

Publisher

Contributor(s)

Format

Identifier

Source

Relation

Coverage

Rights