The University of Science and Technology of China has made important progress in multi-frequency microwave wireless sensing based on Rydberg atoms

Source: HKUST News Network

The team of Academician Guo Guangcan of China University of Science and Technology has made new progress in multi-frequency microwave sensing. The team of Shi Baosen and Ding Dongsheng used artificial intelligence methods to achieve precision detection based on Rydberg atomic multi-frequency microwave, The relevant results were published in the internationally renowned academic journal Nature communications on April 14 under the title of "Deep learning enhanced Rydberg multifrequency microwave recognition".

The Metrology Development Plan (2021-2035) issued by The State Council recently proposed to build a national modern advanced measurement system in 2035 with quantum metrology as the core, scientific and technological level professional, in line with the development needs of The Times and the trend of international development. Because Rydberg atom has a large electric dipole moment, it can produce a strong response to weak electric field, so as a very promising microwave measurement system is favored by people, and has made rapid development. However, there are still many scientific problems to be solved in the field of microwave measurement based on Rydberg atoms, and multi-frequency microwave reception is one of them: this is because multi-frequency microwave can cause complex interference patterns in atoms, which seriously interferes with signal reception and recognition.

In recent years, the scientific research team led by Shi Baosen and Ding Dongsheng has made important progress by using the Rydberg atomic system to focus on quantum simulation and quantum precision measurement science. In this work, the team successfully detected phase-modulated multifrequency microwave fields (frequency division multiplexing binary phase-shift keying signals, a widely used signal transmission method in digital communications) using Rydberg atoms as microwave antennas and modems (as shown in Figure 1) based on a rubidium atom system at room temperature. The received modulated signal is analyzed by deep learning neural network, and the high-fidelity demodulation of multi-frequency microwave signal is realized, and the robustness of the experimental scheme against microwave noise is further tested.

Figure 1 (a) Atomic energy level diagram. (b) Drawings of experimental equipment. (c-e) is the schematic diagram of the neural network layer. (c) is a one-dimensional convolutional layer, (d) is a bidirectional long short-term memory layer, and (e) is a fully connected layer.

The work effectively decodes an FDM phase-shift keying signal with a noisy QR code (as shown in Figure 2) with an accuracy of up to 99.32%. The research results show that the Rydberg microwave receiver based on deep learning enhancement can allow direct decoding of 20 frequency division multiplexing (FDM) signals at once, without the need for multiple bandpass filters and other complex circuits. The innovation of this work is that a scheme is proposed and implemented to effectively detect multifrequency microwave electric fields without solving the main equation, taking advantage of the sensitivity of Rydberg atoms while reducing the impact of noise. This work combines atomic sensing with deep learning, and provides an important reference for the cross combination of precision measurement and neural network. In addition, the results can also be applied to detect multiple targets simultaneously.

Figure 2 shows the machine learning decoding results. (a-c) is the recovery result of the deep learning model to the transmitted signal when the training time is different.

The work was highly praised by the reviewers: "The results presented in this work are very useful to other researchers in the field of atomic and molecular photophysics, as it shows the future application of deep learning in quantum-enhanced sensing of atomic systems." (“The results presented here are very useful to other researchers in the field of AMO physics,  because they can guide and inform future applications of deep learning to quantum-enhanced sensing with atomic  systems.”)

Liu Zongkai, PhD candidate at the Key Laboratory of Quantum Information of the Chinese Academy of Sciences, is the first author of this paper, and Professors Ding Dongsheng and Shi Baosen are co-corresponding authors. This work was supported by the Ministry of Science and Technology, the National Foundation, the Chinese Academy of Sciences, the Anhui Provincial Major Science and Technology Project, and the University of Science and Technology of China.

The article links: https://www.nature.com/articles/s41467-022-29686-7



(Key Laboratory of Quantum Information, Academy of Quantum Information and Quantum Technology Innovation, Scientific Research Department, Chinese Academy of Sciences)