In this paper, we review a variety of deep learning algorithms and models for modulation recognition and classification of wireless communication signals. Specifically, deep learning (DL) has shown overwhelming advantages in computer vision, robotics, and voice recognition. Recently, DL has been proposed to apply to wireless communications for signal detection and classification in order to better learn the active users for electromagnetic spectrum sharing purposes. Therefore, we aim to provide a survey on the most recent techniques which use DL for recognizing and classifying a wireless signal. We focus on the most widely used DL models, emphasize the advantages and limitations, and discuss the challenges as well as future directions. In addition, we also apply a DL algorithm, convolutional neural network (CNN), to demonstrate the feasibility of using CNN to recognize and classify the over-the-air wireless signals using Mathworks DL toolbox with PlutoSDR and Universal Software Radio Peripheral (USRP), respectively.

The detection of deep‐seated lesions is of great significance for biomedical applications. However, due to the strong photon absorption and scattering of biological tissues, it is challenging to realize in vivo deep optical detections, particularly for those using the safe laser irradiance below clinical maximum permissible exposure (MPE). In this work, the combination of ultra‐bright surface‐enhanced Raman scattering (SERS) nanotags and transmission Raman spectroscopy (TRS) is reported to achieve the non‐invasive and photosafe detection of “phantom” lesions deeply hidden in biological tissues, under the guidance of theoretical calculations showing the importance of SERS nanotags’ brightness and the expansion of laser beam size. Using a home‐built TRS system with a laser power density of 0.264 W cm−2 (below the MPE criteria), we successfully demonstrated the detection of SERS nanotags through up to 14‐cm‐thick ex vivo porcine tissues, as well as in vivo imaging of “phantom” lesions labeled by SERS nanotags in a 1.5‐cm‐thick unshaved mouse under MPE. This work highlights the potential of transmission Raman‐guided identification and non‐invasive imaging toward clinically photosafe cancer diagnoses.

With the rapid development and wide deployment of wireless technology, Wi-Fi signals have no longer been confined to the Internet as a communication medium. Wi-Fi signals will be modulated again by human actions when propagating indoors, carrying rich human body state information. Therefore, a novel wireless sensing technology is gradually emerging that can realize gesture recognition, human daily activity detection, identification, indoor localization and human body tracking, vital signs detection, imaging, and emotional recognition by extracting effective feature information about human actions from Wi-Fi signals. Researchers mainly use channel state information or frequency modulated carrier wave in their current implementation schemes of wireless sensing technology, called "Walls have eyes", and these schemes cover radio-frequency technology, signal processing technology, and machine learning. These available wireless sensing systems can be used in many applications such as smart home, medical health care, search-and-rescue, security, and with the high precision and passively device-free through-wall detection function. This paper elaborates the research actuality and summarizes each system structure and the basic principles of various wireless sensing applications in detail. Meanwhile, two popular implementation schemes are analyzed. In addition, the future diversely application prospects of wireless sensing systems are presented.

Steganography, also known as “invisible” communication, refers to the technique of hiding information into another medium such as video, audio, image, and text. Surface‐enhanced Raman scattering (SERS) nanotags are well suited to information encoding owing to their nanoscale dimensions, fingerprint optical spectral features, and remarkable multiplexing capability. Herein, Raman ink, fabricated by doping a new type of SERS nanotags (gap‐enhanced Raman tags, GERTs) into commercial ink, is demonstrated as security ink for multiplexing steganography. A stego‐text is written using two types of Raman ink containing different GERTs with diverse Raman signals, imaged via a confocal Raman system, and processed by the classical least squares method to extract the hidden message. The Raman ink generates distinct spectral profiles with low background from pure ink upon near‐infrared laser irradiation. A multiplexing combination of seven kinds of messages extracted from the written stego‐text is demonstrated, which adds enhanced safety and flexibility to information encoding. In addition, Raman ink exhibits good photostability and long‐term stability. Therefore, GERTs‐based Raman ink is promising for steganographic use in information security.