Wireless Broadband Signal Detection Scheme Based on Quantum Machine Learning

‍ ‍Wireless communication develops to B5G/6G. With the birth of terahertz communication and other technologies， the modulation order of signals and the complexity of channel information increase， and the difficulty of receiving signal detection is also rising. At present， most researches combine wireless communication field with machine learning， but classical machine learning schemes are likely to encounter computational bottlenecks. Quantum machine learning， as a cross-discipline between quantum computing and machine learning， may become a potential technology of 6G. This paper applies quantum machine learning to signal detection in wireless broadband communication for the first time， and designs a quantum-classical hybrid machine learning model based on tensor network. The hybrid model is composed of quantum circuits with parameters and neural networks. The signal data is mapped to the Hilbert space in the quantum circuit for training. It has the advantages of fast convergence， low cost and high performance. The neural network can play a nonlinear correction effect on the output result of the quantum circuit. The simulation in this paper is carried out on Google's Tensorflow-Quantum platform. The results show that the quantum-classical hybrid machine learning model proposed in this paper takes the bit error rate as the quantitative index in signal detection problems， which is about 10 percentage points lower than the existing pure quantum algorithm and neural network algorithm.