Wi-Fi-based human activity recognition has attracted significant attention. Deep learning methods are widely used to achieve feature representation and activity sensing. While more learnable parameters in the neural networks model lead to richer feature extraction, it results in significant resource consumption, rendering the model unsuitable for lightweight Internet of Things (IoT) devices. Furthermore, the sensing performance heavily relies on the quality and quantity of data, which is a time-consuming and labor-intensive task. Therefore, there is a need to explore methods that reduce the dependence on the quality and quantity of the dataset while ensuring recognition performance and decreasing model complexity to adapt to ubiquitous lightweight IoT devices. In this paper, we propose a novel Lightweight-Complex Temporal Convolution Network (L-CTCN) for human activity recognition. Specifically, this approach effectively combines complex convolution with a Temporal Convolution Network (TCN). Complex convolution can extract richer information from limited raw complex data, reducing the reliance on the quality and quantity of training samples. Based on the designed TCN framework with 1D convolution and residual blocks, the proposed model can achieve lightweight human activity recognition. Extensive experiments verify the effectiveness of the proposed method. We can achieve an average recognition accuracy of 96.6% with only 0.17 M parameter size. This method performs well under conditions of low sampling rates and a low number of subcarriers and samples.
With the explosive growth of unmanned aerial vehicle (UAV) applications in numerous fields, low-altitude networks face formidable challenges in monitoring. In this context, integrated sensing and communication (ISAC) networks through three-dimensional (3D) wide-area sensing have emerged as the key solution. However, the differences in networking mechanisms between communication and sensing, along with the transition from two-dimensional (2D) to 3D networking, complicate the realization of seamless 3D sensing. We aimed to address these challenges by analyzing the sensing capabilities of a single base station and the disparities between communication and sensing. Based on this, an innovative 3D sensing model for ISAC single base stations was proposed, defining the sensing boundaries and providing a foundation for designing the key parameters of ISAC base stations. Additionally, a multi-base station (multi-BS) low-altitude networked 3D sensing cellular-like architecture was proposed, overcoming the limitations of traditional 2D networks and achieving seamless 3D sensing. To validate the effectiveness of the model, comprehensive tests were conducted in both controlled laboratory conditions and real-world commercial network environments. The results show that the model successfully achieved stable and continuous sensing with the expected coverage and accuracy in networked environments.
1 January 2026 in IEEE Transactions on Communications
In this paper, we propose an optimization algorithm for passive reflecting unit (PRU) group allocation (PGA) aimed at enhancing the minimum average rate (MAR) of the intelligent reflecting surface (IRS)-assisted unmanned aerial vehicle (UAV)-enabled non-orthogonal multiple access (NOMA) communication system. First, we propose a novel IRS allocation scheme and the phase control policy, i.e., the PRUs of the IRS are grouped for allocation. Then, the MAR is maximized by jointly designing the PRU group allocation and UAV power allocation. To address the multi-variable non-convex problem, we employ the block coordinate descent (BCD) method to decompose it into two subproblems. Subsequently, the subproblems are transformed into convex problems using the successive convex approximation (SCA) technique. Simulation results demonstrate that the PGA algorithm outperforms other traditional algorithms, conventional IRS allocation (CIA) with OMA, NOMA.