Frequency-Modulated Continuous-Wave (FMCW) radar is extensively used in human activity recognition (HAR) due to its non-intrusive operation, all-weather performance, and anti-interference ability. However, FMCW signals predominantly relies on the Discrete Fourier Transform (DFT) method, which is limited by its fixed basis functions and restricted time-frequency resolution (TFR). Therefore, this paper proposes a cross-attention augmented end-to-end architecture for FMCW radar HAR called CASA-SWIN. The preprocessing module of CASA-SWIN learns the potential Range-Time (R-T), Range-Doppler (R-D), Dopper-Time(D-T) features of rawdata through multiple complex fully connected layers with preset parameters. Furthermore, it utilizes a three-branch cross-attention mechanism (TBCAM) module to extract intrinsic related information for R-T, D-T, R-D branches. For the D-T branch, proposed multi-window filtering module (MWFM) comprises multiple 1D complex convolutions with varying kernel sizes, emulating different window lengths in the Short-Time Fourier Transform (STFT). Finally, a three-branch swin-transformer is used as the HAR classifier. Moreover, proposed CASA-SWIN is evaluated on self-collected HAR, Gait recognition datasets, and open-source gesture recognition datasets. Compared to 12 advanced HAR algorithms, a series of experiments including comparative experiments, cross-environment experiments, open-source datasets experiments, and ablation experiments verify the effectiveness of the proposed module and CASA-SWIN.

In recent years, Doppler through-wall radar (TWR) have shown its potential in Internet of Things (IoT) applications like target positioning, smart security, and health monitoring. However, traditional time-frequency analysis (TFA) struggles with close-range or crossed instantaneous frequency (IF) from multiple targets, reducing positioning accuracy. To address this, we propose a novel algorithm, local maximum reassignment chirplet basis transform (LMRCBT). Firstly, the time-frequency characteristics of the target signal are analyzed using Short-Time Fourier Transform (STFT) and the time-frequency distribution (TFD) is chunked to ensure local consistency. Next, we introduce the chirplet basis function combined with a kurtosis-based local optimal Chirp domain theory to capture the optimal TFD and enhance signal energy concentration. Finally, we design a timefrequency reassignment operator (RO) to optimize the redistribution of time frequency coefficients, improving resolution and boosting noise suppression. The main contribution of LMRCBT is its effective handling of non-stationary signals, especially those with close-range, non-proportional, or crossed IF, improving target localization accuracy and suppressing timefrequency aliasing, as shown in experiments.

Disassembly is a critical step in the remanufacturing of end-of-life products. High labor costs and the limited ability of robots to perform intricate disassembly tasks have led to the increasing use of human‒robot collaboration (HRC) for disassembly. This paper addresses a challenge in HRC-based disassembly, i.e., the inherent human uncertainty during disassembly. The uncertainty is that the disassembly time for a task and the task sequence selection by a human during execution might differ from the pre-defined disassembly plan so that dynamic disassembly planning for subsequent tasks is necessary. Stackelberg equilibrium-enabled disassembly task assignment policies are designed to meet the above purpose efficiently and safely. The human leader's policy is to choose tasks that maximize the efficiency-related return value based on the robot's optimal response to the human's choice. As the follower, the robot selects tasks that maximize the safety-related return value for each human task choice. To identify the optimal values of the policies to ensure the safety and efficiency of the entire HRC-based disassembly process, an improved multi-agent proximal policy optimization (i-MAPPO) algorithm is designed. Finally, a case study for disassembling an electric vehicle battery is used to verify that the proposed approach can adapt to human uncertainty with a high success rate while ensuring that the disassembly time remains short and the human-robot distance remains within the safety threshold throughout the disassembly process.

Doppler through-wall radar (TWR) is a promising device for the Internet of Things (IoT), effective for indoor tracking, health monitoring, and smart homes. However, employing it to estimate the trajectories of multiple targets presents challenges associated with time-frequency analysis (TFA). In this article, a multimodal network called MWVD is proposed, which eliminates the crossterm problem of Wigner-Ville distribution (WVD) and improves the accuracy of instantaneous frequency (IF) extraction to obtain accurate localization. In the MWVD, both the WVD spectrogram and the 1-D complex signals are used as inputs to the network. The complex signals are passed through the proposed multiwindow short-time filtering (MWSTF) module followed by an adaptive wavelet attention fusion (AWAF) module to simulate the wavelet transform. Subsequently, the enhanced WVD spectrogram is obtained by the energy compression module. As a result, comprehensive experiments, including simulated signal tests, module ablation studies, fusion mode ablation analyses, and real TWR target tracking, are conducted to demonstrate the proposed algorithm’s excellence, which will be combined with more IoT applications in the future.

Through-wall radar-based target localization algorithm has great promise in the Internet of Things (IoT), such as indoor positioning, health monitoring, and surveillance. However, when locating multiple targets, time-frequency aliasing occurs in the echo’s time-frequency distribution (TFD), which makes it challenging to accurately extract the target instantaneous frequency (IF) curve from the TFD, ultimately hindering the achievement of high-precision positioning. In this article, we propose a joint algorithm based on the improved linear chirplet transform (ILCT) and singular value decomposition (SVD) for target tracking, aiming to improve the localization precision of TWR. We design an ILCT algorithm to increase the time-frequency energy concentration of the target component of interest. Then, we use the SVD algorithm, based on the ILCT results, to accurately separate the target signal of interest from the echo signal. Finally, the motion path of the tracked target is synthesized based on the estimated target IF curves. Experimental results of target tracking demonstrated that the proposed method not only improves the target localization precision but also effectively suppresses the time-frequency aliasing phenomenon.

This letter proposes a target localization algorithm for Doppler through-wall radar (TWR) based on the Hough transform and instantaneous frequency (IF) correction. This algorithm first employs the extended Bézier curve model, which is based on Hough transform, to fit the interested target IF. Then, the fitted IF curve is used to demodulate the signal, and the IF error of the tracked target is obtained by approximating the target demodulated signal. Finally, the obtained IF error is used to correct the results of the extended Bézier fit, and the motion trajectory of the target is synthesized according to the corrected IF curve. The proposed method combines curve fitting with IF correction, resulting in improved positioning accuracy of the TWR and effective suppression of frequency aliasing in the time-frequency distribution (TFD). Experimental results demonstrate the effectiveness of the proposed algorithm.

In target tracking applications by Doppler through-wall radar (TWR), the frequency ambiguity issue is a notable drawback, which can severely degrade the target localization performance. To overcome this drawback, a convolutional neural networks (CNNs) based target tracking algorithm is proposed in this letter. First, the short-time Fourier transform (STFT) is used to acquire an echo spectrogram. Then, the spectrogram resolution is enhanced by introducing a modified CNN module. Compared with the traditional CNN module, an additional convolutional weight block with a residual structure is designed to help extract deeper time–frequency features. Finally, the target instantaneous frequency (IF) curves are extracted from the enhanced spectrogram and the target positions are obtained. Experimental results prove that the proposed method can effectively enhance the spectrogram resolution and suppress the frequency ambiguity issue, which would lead to higher target localization accuracy and better robustness than traditional target tracking approaches.

The micro-Doppler (m-D) modulations to radar backscattering introduced by the flexible body articulations and complicated movement patterns of human movers can provide valuable information for activity classification and help to identify the interested targets. In particular, the m-D signal of limbs, as a highly distinctive feature of human activities, can be used as an effective clue to discriminate between the armed and unarmed persons, as well as the humans against other small animals. In this article, a novel theoretical method is proposed to extract the target m-D trajectories through an integrated application of modified Viterbi algorithm and Hough transform. Through this method, multiple components corresponding to various target scattering parts and their respective m-D trajectories can be accurately extracted and estimated, even in the overlapping regions of different scattering parts in the time–frequency (TF) distribution. The employed search method enhances upon the traditional ones and improves the efficiency of finding the optimal paths considerably. Finally, a series of experiments is conducted to illustrate the validity and performance of the proposed techniques. Compared to short-time Fourier transform (STFT) peak detection and traditional Viterbi algorithm, the average error of m-D frequency estimated by the proposed algorithm is reduced by 82.1% and 71.8%, respectively. Besides, the processing time is reduced by 36.4% compared to the traditional Viterbi algorithm.