NGHIÊN CỨU THUẬT TOÁN TỰ ĐỘNG NHẬN DẠNG TÍN HIỆU RA-ĐA SỬ DỤNG PHÂN TÍCH WAVELET KẾT HỢP MẠNG ALEXNET=AUTOMATIC RADAR SIGNAL RECOGNITION USING THE ANALYTIC WAVELET TRANSFORM AND ALEXNET

0 Người đánh giá. Xếp hạng trung bình 0

Tác giả: Xuân Tùng Vũ

Ngôn ngữ: vie

Ký hiệu phân loại:

Thông tin xuất bản: Tạp chí Khoa học và Công nghệ - Đại học Thái Nguyên, 2024

Mô tả vật lý: tr.222 - 230

Bộ sưu tập: Metadata

ID: 484662

Thêm vào giỏ

Bài báo này đề xuất một cách tiếp cận khác để nhận dạng tín hiệu ra-đa sử dụng mạng AlexNet để nâng cao độ chính xác và rút ngắn thời gian huấn luyện phục vụ cho mục đích trinh sát điện tử. Thuật toán đề xuất gồm có 02 bước: bước 1 sử dụng để trích xuất đặc trưng tín hiệu bằng kỹ thuật phân tích Wavelet trên cả hai miền thời gian-tần số. Bước 2 sử dụng mạng AlexNet để nhận dạng các tín hiệu nói trên. Hiệu quả của thuật toán được đánh giá bằng các tín hiệu mô phỏng trên MATLAB. Ngoài ra, thuật toán đề xuất được đánh giá qua 2 trường hợp. Trường hợp 1 đánh giá ảnh hưởng của các hàm Wavelet: Morse, Cauchy and Bessel. Kết quả mô phỏng cho thấy, hàm Morse Wavelet cho độ chính xác cao nhất so với các hàm Cauchy và Bessel. Trường hợp 2 so sánh hiệu quả của thuật toán đề xuất với các mạng khác như: GoogleNet, ResNet và VGG-16. Kết quả mô phỏng cho thấy, thuật toán đề xuất cho độ chính xác nhận dạng cao nhất (85%) trong khi đó các phương pháp là <
80%. Ngoài ra, thuật toán đề xuất cho thời gian huấn luyện giảm đi ½ so với các cấu trúc mạng đã được đề xuất, huấn luyện trong MATLAB.This paper proposes an alternative approach to radar signal modulation recognition using the combination of the analytic wavelet transform and the AlexNet network to improve accuracy and time training. The proposed algorithm includes two steps. The first step is used to extract signal features using wavelet analysis techniques in both time-frequency domains. The second step uses the AlexNet network to identify the above signals. The algorithm's effectiveness is evaluated by using simulated radar signals in a MATLAB environment. In addition, the proposed method is evaluated in two stages. The first stage involves analyzing wavelet functions such as Morse, Cauchy and Bessel on effectiveness of the proposed method. The simulation results showed that the Morse wavelet function provided the highest accuracy in comparison with Cauchy and Bessel. The second stage provides comparisons with other networks such as GoogleNet, ResNet, and VGG-16. Simulation results show that the proposed algorithm has the highest recognition accuracy (85%), while other methods are lower than 80%, and the network training time is reduced ½ in comparison with other networks.

1. Đọc trực tuyến

1. Đọc trực tuyến

Tạo bộ sưu tập với mã QR