A target group tracking algorithm based on a hybrid sensor network

Abstract

Traditional tracking algorithms based on static sensors have several problems. First, the targets only occur in a part of the interested area; however, a large number of static sensors are distributed in the area to guarantee entire coverage, which leads to wastage of sensor resources. Second, many static sensors have to remain in active mode to track the targets, which causes an increase of energy consumption. To solve these problems, a target group tracking algorithm based on a hybrid sensor network is proposed in this paper, which includes static sensors and mobile sensors. First, an estimation algorithm is proposed to estimate the objective region by static sensors, which work in low-power sensing mode. Second, a movement algorithm based on sliding windows is proposed for mobile sensors to obtain the destinations. Simulation results show that this algorithm can reduce the number of mobile sensors participating in the tracking task and prolong the network lifetime.

Dual-Template Siamese Network with Attention Feature Fusion for Object Tracking

“…With the continuous development of deep learning, object tracking algorithms based on them have become the mainstream due to their excellent performance, and they can be divided into filtering algorithms combined with deep features and Siamese network based algorithms [3].…”