Home // International Journal On Advances in Intelligent Systems, volume 10, numbers 1 and 2, 2017 // View article
Authors:
Yuto Tamura
Ryohei Murabayashi
Masafumi Hashimoto
Kazuhiko Takahashi
Keywords: moving-object tracking; cooperative tracking; centralized and hierarchical methods; laser scanner; mobile robot.
Abstract:
This paper presents a tracking (estimation of the pose and size) of moving objects such as pedestrians, cars, motorcycles, and bicycles, using multiple mobile robots as sensor nodes. In this cooperative-tracking method, nearby sensor nodes share their tracking information, enabling the tracking of objects that are invisible or partially visible to an individual sensor node. The cooperative-tacking method can then make the tracking system more reliable than the conventional individual tracking method by a single robot. We previously presented a centralized architecture of cooperative tracking. Each sensor node detected moving objects in its own laser-scanned images captured by a single-layer laser scanner. It then sent measurement information related to the moving objects to a central server. The central server estimated the objects’ poses (positions and velocities) and sizes from the measurement information using Bayesian filter. However, such a centralized method might have poor dependability and impose a computational burden upon the central server. To address this problem, this paper presents hierarchical architecture of cooperative tracking. Each sensor node locally estimates the poses and sizes of moving objects and then sends these estimates to the central server, which then merges the pose and size estimates. Experimental results using two sensor nodes in outdoor environments show that the proposed hierarchical cooperative-tracking method provides slightly inferior tracking accuracy and has a smaller computational cost in the central server than a previous centralized method.
Pages: 90 to 101
Copyright: Copyright (c) to authors, 2017. Used with permission.
Publication date: June 30, 2017
Published in: journal
ISSN: 1942-2679