Hardware Development for Joint Sparse Decentralized Heterogeneous Data Fusion for Target Estimation

For surveillance applications, users are typically interested in detecting and tracking targets of interest. In many surveillance systems, heterogeneous sensor data are collected by sensors of varying sensing modalities with different dimensionalities. However, individual sensing system collects only one kind of signal, leading to a lack of accuracy to characterize the target. One of the most popular approaches to overcome single sensor assessment is data fusion. In our recent work, a new joint-sparse data-level fusion (JSDLF) approach to integrate heterogeneous sensor data for target discovery is developed in the matter of using a decentralized architecture algorithm. Several decentralized implementations of the data-level fusion approach based on the JSDLF approach were developed and investigated in the software aspect. In this paper, the hardware development of this algorithm will be displayed. Several unmanned aerial vehicles (UAVs) were utilized as nodes for image or RF SIGINT data processing. Meanwhile, a computer with RF SIGINT data processing communicated with these nodes as decentralized architecture as the hardware development of this algorithm. The performance of the decentralized hardware development is represented in this paper and the results are comparable with only software development in the algorithm.