Sparse Representation Combined with Edge Computing in Target Recognition of Live Work Using Target Recognition Algorithm

• Published in: 2024 International Conference on Smart Systems for Electrical, Electronics, Communication and Computer Engineering (ICSSEECC) pp. 763 - 767
• Main Authors: Hussian, Ahmed, Jaaywel, Waleed Sadeq, Alabdeli, Haider, Ahmed, Ibrahem, Ismail, Laith S.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 28.06.2024
• Subjects: Accuracy; Dictionaries; Edge computing; loT; Sparse approximation; sparse representation; Target recognition; Target tracking; Training; Uncertainty
• DOI: 10.1109/ICSSEECC61126.2024.10649530

Signal decomposition is crucial in signal processing, but orthogonal decomposition can be problematic for signals with a broad time-frequency field. To address this issue, sparse representation is used to compress high-resolution range profile data and extract functions. A redundant structural dictionary and a fast sparse representation approach are introduced for radar target detection. This research uses a unique Target Recognition (TR) method with a fast sparse decomposition technique for objectively recognizing SAR images. The study distinguishes malicious loT -Edge Computer (EC) network traffic from compromised loT equipment using the loT botnet attack detection (SRF -loTAD) method with rehabilitation error threshold criteria. The TR-EC method extracts the generalized two-dimensional main element evaluating properties of training samples to construct sub- dictionaries. The orthogonal matching tracking method computes coefficients for sparse displays of test samples across sub-dictionaries. The results of an actual loT -based network dataset were compared to those obtained using an autoencoder approach, revealing the highest analysis, comparison, accuracy, and performance ratios compared to other approaches.