arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.00152cs.CVcs.AIeess.SP
收藏

边缘微控制器上的实时人类活动识别:基于多光谱传感器融合的动态分层推理

Real-Time Human Activity Recognition on Edge Microcontrollers: Dynamic Hierarchical Inference with Multi-Spectral Sensor Fusion

Nanyang Technological University(南洋理工大学);Henan University(河南大学)

Boyu Li, Kuangji Zuo, Lincong Li, Yonghui Wu

AI总结:

HPPI-Net通过多光谱传感器融合和分层推理实现边缘微控制器上的实时人类活动识别,提升准确率并降低资源消耗。

摘要:

对边缘应用中准确的设备端模式识别需求日益增加,但现有方法难以在准确性与计算约束之间取得平衡。为解决这一挑战,提出了一种基于多光谱融合和可解释模块的资源感知分层网络,即分层并行伪图像增强融合网络(HPPI-Net),用于实时、设备端的人类活动识别(HAR)。该网络部署在ARM Cortex-M4微控制器上,以低功耗实时推断为目标,优化后仅占用22.3 KiB的RAM和439.5 KiB的ROM,实现了96.70%的准确率。HPPI-Net采用双层架构。第一层利用快速傅里叶变换(FFT)频谱图提取初步特征,第二层则选择性激活用于静止活动识别的专用模块或用于动态状态的并行LSTM-MobileNet网络(PLMN)。PLMN通过三个并行LSTM编码器融合FFT、小波和Gabor频谱图,并通过高效通道注意力(ECA)和深度可分离卷积(DSC)对拼接特征进行细化,从而在通道层面提供可解释性的同时大幅减少乘法-累加操作。与MobileNetV3相比,HPPI-Net在准确率上提高1.22%,RAM使用减少71.2%,ROM使用减少42.1%。这些结果表明,HPPI-Net在准确率与效率之间取得了良好的平衡,并提供了可解释的预测,为在内存受限的边缘平台上实现可穿戴、工业和智能家居的HAR提供了实用解决方案。

英文摘要:

The demand for accurate on-device pattern recognition in edge applications is intensifying, yet existing approaches struggle to reconcile accuracy with computational constraints. To address this challenge, a resource-aware hierarchical network based on multi-spectral fusion and interpretable modules, namely the Hierarchical Parallel Pseudo-image Enhancement Fusion Network (HPPI-Net), is proposed for real-time, on-device Human Activity Recognition (HAR). Deployed on an ARM Cortex-M4 microcontroller for low-power real-time inference, HPPI-Net achieves 96.70% accuracy while utilizing only 22.3 KiB of RAM and 439.5 KiB of ROM after optimization. HPPI-Net employs a two-layer architecture. The first layer extracts preliminary features using Fast Fourier Transform (FFT) spectrograms, while the second layer selectively activates either a dedicated module for stationary activity recognition or a parallel LSTM-MobileNet network (PLMN) for dynamic states. PLMN fuses FFT, Wavelet, and Gabor spectrograms through three parallel LSTM encoders and refines the concatenated features using Efficient Channel Attention (ECA) and Depthwise Separable Convolution (DSC), thereby offering channel-level interpretability while substantially reducing multiply-accumulate operations. Compared with MobileNetV3, HPPI-Net improves accuracy by 1.22% and reduces RAM usage by 71.2% and ROM usage by 42.1%. These results demonstrate that HPPI-Net achieves a favorable accuracy-efficiency trade-off and provides explainable predictions, establishing a practical solution for wearable, industrial, and smart home HAR on memory-constrained edge platforms.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑