Secure Finite-Memory Target Tracking in Heterogeneous Sensor Networks Under Cyber Attacks

Heterogeneous sensor networks (HSNs) are networks composed of different types of sensors, which have the advantage of collecting information about different characteristics of the target. This work addresses the problem of tracking the position of target objects in indoor spaces using HSNs. We propose a secure and reliable target tracking algorithm that remains effective even under cyber attacks during the information transmission process of HSNs. The proposed algorithm is designed to have a finite-memory structure that does not continuously remember past errors and can use the Kullback-Leibler divergence to determine the confidence/trust level of sensor information under cyber attacks. We conducted experiments to track the position of a mobile robot in an HSN consisting of cameras and LiDAR, and demonstrated that the proposed algorithm provides secure and reliable tracking performance under cyber attacks.