The invention relates to the technical field of
wireless communication
signal processing, in particular to a non-cooperative
signal whole-process
processing and multi-
station cooperative positioning method based on
deep learning. According to the technical scheme, the method comprises the steps of
frequency spectrum monitoring and
signal detection, non-cooperative signal
demodulation,
radiation source identification, special signal protocol analysis and multi-
station cooperative positioning. According to the invention, the limitation of link-by-link isolated
processing in the traditional technology is broken through, and the integrated and high-robustness processing of the non-cooperative signal from detection to positioning is realized; core pain points such as difficulty in
weak signal detection, influence of
phase distortion on
demodulation, restriction of low signal-to-
noise ratio on recognition, contradiction between data privacy and positioning precision and difficulty in
adaptation of a model to environment change in a complex
electromagnetic environment are solved, and non-cooperative
signal processing efficiency and positioning precision are integrally improved through collaborative optimization of all links. The method can stably adapt to actual requirements of key scenes such as
frequency spectrum monitoring,
electronic countermeasure and public safety prevention and control, and has remarkable practical value and technical advantages.