adelaideauto-idlab / distributed-limitedfov-mot Goto Github PK

This repository is provided as part of the our TSP paper:

H. V. Nguyen, H. Rezatofighi, B.-N. Vo, D. Ranasinghe, "Distributed Multi-Object Tracking Under Limited Field of View Sensors". IEEE Transactions on Signal Processing, vol. 69, pp. 5329-5344, 2021, doi: 10.1109/TSP.2021.3103125. Paper.

If you use our provided codes, please consider cite our paper using:

@article{nguyen2021distributed,
	title={Distributed Multi-Object Tracking Under Limited Field of View Sensors},
	author={Nguyen, Hoa Van and Rezatofighi, Hamid and Vo, Ba-Ngu and Ranasinghe, Damith C},
	journal = {{IEEE Transactions on Signal Processing}},
	year={2021
	volume={69},
	number={},
	pages={5329-5344},
	doi={10.1109/TSP.2021.3103125}
}

1. All objects move follow a 4D Constant Velocity (x/y position and velocity) with 2D observations (position only).

2. Each node runs an LMB filter with joint prediction and update step, with a measurement-based Adaptive Birth Procedure (from data of previous measurements).

3. Call demo.mto run it, with the following settings for different scenarios in the Paper:

   • Scenario 1 by a pair of (property, value) of 'case_id',1 in gen_settings:

     settings =  gen_settings('case_id',1,'sel_pd',0.98);  

   • Scenario 2 by a pair of (property, value) of 'case_id',2 in gen_settings:

     settings =  gen_settings('case_id',2,'sel_pd',0.98);  
     

   • Scenario 3 by a pair of (property, value) of 'case_id',3 in gen_settings:

     settings =  gen_settings('case_id',3,'sel_pd',0.98);  
     

4. We only publish our proposed Track Consensus method:

   • Track consensus using OSPA2TC-OSPA2 by setting a pair of (property, value) to 'metric_type','ospa_union' in gen_model:

     model = gen_model(settings,'meas_sigma',10,'lambda_c',10,'track_threshold',0.001,'metric_type','ospa_union'); 

   • Track consensus using Wasserstein TC-WASS by setting a pair of (property, value) to 'metric_type','wasserstein':

     model = gen_model(settings,'meas_sigma',10,'lambda_c',10,'track_threshold',0.001,'metric_type','wasserstein');  

5. Codes are provided for academic research purposes only. For a commercial license, please contact me at [email protected].