Millimeter-wave radar holds significant potential for dynamic gesture recognition in contactless human-computer interaction, particularly in the Internet of Things and consumer electronics applications. However, a considerable challenge persists in filtering vast amounts of extraneous data from millimeter-wave radar echoes to isolate meaningful gesture features. We present a novel approach based on sparse representation principles to address this. We first generate a range-Doppler map of gestures using a two-dimensional (2-D) fast Fourier transform, then construct a Doppler-Time trajectory from aggregated data across multiple frames. Capitalizing on the intrinsic sparsity in the Doppler-time domain, we employ the orthogonal matching pursuit algorithm to refine a multidimensional feature sequence across time, Doppler, and range dimensions. Central to our approach is a dual-stream Transformer network that explores complex 2-D correlations in feature sequences via multihead self-attention mechanisms. This technique significantly improves gesture feature extraction efficiency and reduces data redundancy. The experimental results show that our model has an average recognition accuracy of 99.17% and a size of 0.17M, which is very suitable for application in embedded devices.

Dynamic gesture recognition using millimeter-wave radar has a broad application prospect in the industrial Internet of Things (IoT) field. However, the existing methods in the random dynamic interference environment, such as throwing objects and waving and easily cause wrong recognition. This article proposes a dynamic gesture recognition method based on a convolutional neural network (CNN)-Transformer network to solve this problem. First, we reshape the original echoes acquired by the frequency-modulated continuous-wave (FMCW) millimeter-wave radar into 3-D data blocks in terms of Chirps$\times $Samples$\times $Frames. And we employ the mean elimination method to eliminate the static interference. Second, we extract dynamic gestures’ distance and Doppler information with the 2-D fast Fourier transform and obtain the range-time map and Doppler-time maps. And we employ the coherent accumulation method to improve the signal-to-noise ratio (SNR). Third, we construct the CNN-Transformer network model for dynamic gesture recognition. The CNN is used to extract the local features of gestures, and multiple Transformer modules are stacked to extract deeper effective features. Finally, we build a data set for gesture recognition, including six kinds of dynamic gestures and two kinds of random interference. The experimental results show that the proposed method has a gesture recognition accuracy of more than 98% and 96% in the noninterference scene and the random dynamic interference scene, respectively, which are superior to the conventional recognition methods.

Due to the large path loss in millimeter wave (mmWave) band, the transmission path between transmitter (Tx) and intelligent reflecting surface (IRS) is considered as a Rayleigh fading channel, and a physics-based channel model is proposed for IRS-assisted mmWave communication system in urban scenario. Also, the horizontal and vertical rotation angles of IRS and the relationship between the scattering gain of IRS reflecting unit and its effective aperture in the incident direction and the desired reflection direction are considered in the proposed model. For the considered communication scenario, the existing reflection phases, which are designed to align the virtual line-of-sight (VLoS) components among Tx, IRS, and receiver (Rx) with the LoS components between Tx and Rx, are not the appropriate reflection phases. Based on the proposed model, we first obtain the statistical phases of the virtual scattering components within a cluster by minimizing phase differences between different IRS reflection units, and then obtain the reflection phases by minimizing the phase differences of the derived statistical phases for all clusters. By comparing with the existing reflection phases, the designed reflection phases can significantly enhance the system performance gains of mmWave communications. Using the designed reflection phases, the expressions of received signal power and upper bound of ergodic sum capacity are derived in this paper, which are validated by using Monte-Carlo simulation results. Numerical results show that the proposed mmWave channel model could accurately simulate the propagation characteristics of IRS. Also, numerical results show that the performance gains of IRS-assisted systems are equivalent to that of large-scale communication systems without using IRS.

Dynamic gesture recognition using millimeter-wave radar is a promising contactless mode of human–computer interaction with wide-ranging applications in various fields, such as intelligent homes, automatic driving, and sign language translation. However, the existing models have too many parameters and are unsuitable for embedded devices. To address this issue, we propose a dynamic gesture recognition method (named “Gesture-mmWAVE”) using millimeter-wave radar based on the multilevel feature fusion (MLFF) and transformer model. We first arrange each frame of the original echo collected by the frequency-modulated continuously modulated millimeter-wave radar in the Chirps × Samples format. Then, we use a 2-D fast Fourier transform to obtain the range-time map and Doppler-time map of gestures while improving the echo signal-to-noise ratio by coherent accumulation. Furthermore, we build an MLFF-transformer network for dynamic gesture recognition. The MLFF-transformer network comprises an MLFF module and a transformer module. The MLFF module employs the residual strategies to fuse the shallow, middle, and deep features and reduce the parameter size of the model using depthwise-separable convolution. The transformer module captures the global features of dynamic gestures and focuses on essential features using the multihead attention mechanism. The experimental results demonstrate that our proposed model achieves an average recognition accuracy of 99.11% on a dataset with 10% random interference. The scale of the proposed model is only 0.42M, which is 25% of that of the MobileNet V3-samll model. Thus, this method has excellent potential for application in embedded devices due to its small parameter size and high recognition accuracy.

Abstract Frequency diversity array (FDA) radar can provide full spatial coverage with stable gains within a pulse duration. Based on the FDA, the integrated radar‐communication system can perform multi‐directional communication and whole‐space detection. However, the embedded communication bits disrupt the correlation of the transmitting waveform of each element. Correspondingly, the range sidelobe level (SLL) of the multi‐dimensional ambiguity function increases significantly. To address this issue, a low‐sidelobe waveform for integrated radar‐communication systems based on the FDA was designed. Two techniques based on the subarray time delay are employed to reduce the SLL in range dimension. Both methods, however, lower the angular resolution. Thus, a tangent FM signal as the baseband waveform to improve the angular resolution was selected. Simultaneously, the received signal processing methods of radar and communication was designed. The performances of the designed waveform are verified by analysing the multi‐dimensional ambiguity function and the bit error rate. The simulation results reveal that the proposed method can maintain a good radar target detection capability and satisfy the communication function.

Intelligent reflecting surface (IRS) composed of a large number of low-cost passive reflecting elements has attracted significant attention from communication communities because of its ability to substantially improve the communication performance. In this paper, the aperture area and radiation pattern of the IRS reflecting elements are considered, and a novel geometry-based three-dimensional (3-D) wideband channel model is proposed for IRS-assisted unmanned aerial vehicle (UAV) communication systems. In the proposed model, the reflection phase is designed by jointly considering the aperture area of the IRS reflecting element and the propagation phases among UAV, IRS, and receiver (Rx), each IRS reflecting element is modeled as an anomalous reflector instead of a specular reflector, and large-scale IRS reflecting elements can jointly beamform the signal in a desired direction. Based on the proposed model, the effects of arbitrary trajectory of UAV and the number and the size of passive reflecting elements on channel statistical characteristics are considered, and the average received signal power and the ergodic sum capacity, which consider the impacts of the number and the size of passive reflecting elements, are also investigated. Furthermore, the path loss of the IRS-assisted link, which is in inverse proportion to the square of aperture area of IRS reflecting elements, is derived, and it coincides with the measured results in real outdoor scenarios. Analysis shows that the communication performance can be enhanced by increasing the number and the size of IRS reflecting elements, and it is validated by numerical results and Monte-Carlo simulation results.

Abstract In order to achieve precise localization in the underwater environment, an underwater localization method based on hybrid invasive weed algorithm is proposed. First, considering the insufficient local search capability of the invasive weed algorithm, the salp swarm algorithm is used to pre‐process the initial population, the position and velocity update formulas are employed to further improve the search depth, and the hybrid invasive weed algorithm is obtained. Then, the localization error is designed as the objective function to obtain the initial value of the source, and a new localization model is built using the initial values and interference parameters. Finally, an improved two‐step weighted least square method is used to obtain the precise location of the source. The performance of the algorithm is also verified by comparisons with Cramer‐Rao Lower Bound. Simulation experiments show that the proposed algorithm has better localization accuracy than other classic methods.