Frequency Matching in Spiking Neural Networks for mmWave Sensing

Millimeter-wave (mmWave) sensing enables privacy-preserving, always-on edge perception, but its measurements are often sparse, temporally irregular, and corrupted by high-frequency noise. Existing mmWave pipelines predominantly rely on artificial neural networks (ANNs), which achieve robustness through extensive preprocessing or deep architectures, thereby limiting their efficiency on edge devices. In this work, we study spiking neural networks (SNNs) for mmWave sensing from a mechanism–data alignment perspective. By leveraging the low-pass filtering behavior of leaky integrate-and-fire (LIF) dynamics, we analyze how their implicit temporal filtering interacts with the frequency structure of mmWave signals. Our analysis shows that when discriminative information resides in low-to-mid frequencies, LIF dynamics can inherently suppress high-frequency noise, clarifying when and why SNNs outperform ANNs. Based on this insight, we derive a principled criterion for configuring the membrane decay factor by matching the effective bandwidth of LIF dynamics to the data’s discriminative spectral content. Experimental results across four widely used mmWave datasets validate the proposed frequency-matching hypothesis, yielding an average test-accuracy improvement of 6.22% and a 3.64× reduction in theoretical energy consumption relative to ANN baselines, under a unified evaluation protocol.

Millimeter-wave (mmWave) sensing enables privacy-friendly applications such as human activity recognition and smart environments, but its signals are often noisy, sparse, and irregular. Existing solutions mainly rely on conventional deep neural networks, which usually require large models or heavy preprocessing, limiting their deployment on low-power edge devices. In this work, we study spiking neural networks (SNNs), a brain-inspired computing paradigm, for mmWave sensing. We show that the temporal dynamics of spiking neurons naturally act as a low-pass filter, allowing SNNs to suppress high-frequency noise in mmWave signals. Based on this insight, we develop a principled method to match neuron dynamics with the frequency characteristics of the sensing data. Experiments on four widely used mmWave datasets show that our approach improves recognition accuracy by 6.22% on average while reducing theoretical energy consumption by 3.64× compared with conventional neural networks. Our findings suggest that SNNs can provide a more efficient and robust solution for future always-on edge sensing applications.