Download PDFOpen PDF in browser

Multi-Target Tracking and Detection, fusing RADAR and AIS Signals using Poisson Multi-Bernoulli Mixture Tracking, in support of Autonomous Sailing

EasyChair Preprint 4259

13 pagesDate: September 25, 2020

Abstract

To sail safely, an autonomous vessel should be able to keep track of the position and motion of other vessels and obstacles, which refers to the multi-target tracking problem. Furthermore, RADAR and automatic identification system (AIS) are two sensors commonly used onboard for tracking maritime targets. The fusion of these two sensors, utilizing complementary information and handling the conflicting data, gets increasingly important during autonomous sailing. However, due to the immaturity of multi-target tracking methods, the fusion was hardly systematically discussed, when there are missed detections from certain single sensors and conflicts between two sensors. As the new multi-target tracking methods have been proposed, this paper first presents a sequential measurement-level fusion approach of RADAR and AIS based on the newest random finite set (RFS)-based filter — Poisson multi-Bernoulli mixture (PMBM) filter. The comparison of the performance both using sequential fusion and using the sensor information individually is presented in this article. Then the proposed sequential fusion of RADAR and AIS based on PMBM filter was applied to a real maritime case. The tracking results are given and the performance is analyzed.

Keyphrases: Heterogeneous Sensor Fusion, PMBM filter, Radar/AIS fusion, data association, multi-target tracking

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
@booklet{EasyChair:4259,
  author    = {Tianlei Miao and Ehab El Amam and Peter Slaets and Davy Pissoort},
  title     = {Multi-Target Tracking and Detection, fusing RADAR and AIS Signals using Poisson Multi-Bernoulli Mixture Tracking, in support of Autonomous Sailing},
  howpublished = {EasyChair Preprint 4259},
  year      = {EasyChair, 2020}}
Download PDFOpen PDF in browser