Human activity recognition (HAR) on wearable and mobile devices is constrained by memory footprint and computational budget, yet competitive accuracy must be maintained across heterogeneous sensor configurations. Selective state space models (SSMs) offer linear time sequence processing with input dependent gating, presenting a compelling alternative to quadratic complexity attention mechanisms. However, the design space for deploying SSMs in the TinyML regime remains largely unexplored. In this paper, BabyMamba-HAR is introduced, a framework comprising two novel lightweight Mamba inspired architectures optimized for resource constrained HAR: (1) CI-BabyMamba-HAR, using a channel independent stem that processes each sensor channel through shared weight, but instance independent transformations to prevent cross channel noise propagation, and (2) Crossover-BiDir-BabyMamba-HAR, using an early fusion stem that achieves channel count independent computational complexity. Both variants incorporate weight tied bidirectional scanning and lightweight temporal attention pooling. Through evaluation across eight diverse benchmarks, it is demonstrated that Crossover-BiDir-BabyMamba-HAR achieves 86.52% average macro F1-score with approximately 27K parameters and 2.21M MACs, matching TinyHAR (86.16%) while requiring 11x fewer MACs on high channel datasets. Systematic ablation studies reveal that bidirectional scanning contributes up to 8.42% F1-score improvement, and gated temporal attention provides up to 8.94% F1-score gain over mean pooling. These findings establish practical design principles for deploying selective state space models as efficient TinyML backbones for HAR.

翻译：在可穿戴和移动设备上进行人体活动识别受限于内存占用和计算预算，同时必须在异构传感器配置下保持有竞争力的准确率。选择性状态空间模型通过输入依赖的门控机制提供线性时间序列处理能力，为二次复杂度的注意力机制提供了一个引人注目的替代方案。然而，在TinyML领域部署SSM的设计空间在很大程度上仍未得到探索。本文提出了BabyMamba-HAR，这是一个包含两种新颖的、受Mamba启发的轻量级架构的框架，专为资源受限的HAR任务优化：(1) CI-BabyMamba-HAR，采用通道独立主干，通过共享权重但实例独立的变换处理每个传感器通道，以防止跨通道噪声传播；(2) Crossover-BiDir-BabyMamba-HAR，采用早期融合主干，实现了与通道数量无关的计算复杂度。两种变体都采用了权重绑定的双向扫描和轻量级时序注意力池化。通过在八个不同基准数据集上的评估，结果表明Crossover-BiDir-BabyMamba-HAR以约27K参数和2.21M MACs实现了86.52%的平均宏F1分数，与TinyHAR（86.16%）性能相当，同时在高通道数据集上所需的MACs减少了11倍。系统的消融研究表明，双向扫描可带来高达8.42%的F1分数提升，而门控时序注意力池化相比平均池化可提供高达8.94%的F1分数增益。这些发现为将选择性状态空间模型部署为HAR的高效TinyML骨干网络确立了实用的设计原则。