CoSS: Co-optimizing Sensor and Sampling Rate for Data-Efficient AI in Human Activity Recognition

TL;DR

This work introduces a pragmatic framework for data-efficient utilization in HAR tasks, considering the optimization of both sensor modalities and sampling rate simultaneously, and tests the framework's effectiveness in optimizing sensor modality and sampling rate selection.

Abstract

Recent advancements in Artificial Neural Networks have significantly improved human activity recognition using multiple time-series sensors. While employing numerous sensors with high-frequency sampling rates usually improves the results, it often leads to data inefficiency and unnecessary expansion of the ANN, posing a challenge for their practical deployment on edge devices. Addressing these issues, our work introduces a pragmatic framework for data-efficient utilization in HAR tasks, considering the optimization of both sensor modalities and sampling rate simultaneously. Central to our approach are the designed trainable parameters, termed 'Weight Scores,' which assess the significance of each sensor modality and sampling rate during the training phase. These scores guide the sensor modalities and sampling rate selection. The pruning method allows users to make a trade-off between computational budgets and performance by selecting the sensor modalities and sampling rates according to the weight score ranking. We tested our framework's effectiveness in optimizing sensor modality and sampling rate selection using three public HAR benchmark datasets. The results show that the sensor and sampling rate combination selected via CoSS achieves similar classification performance to configurations using the highest sampling rate with all sensors but at a reduced hardware cost.

Figures (7)

• Figure 1: Proposed CoSS framework, the framework is based on multi-branch feature fusion architecture, where three additional layers, such as resampling layers, sampling rate, and sensor selection layers, are embedded. SRn is the downsampling data from the resampling layer
• Figure 2: Sensor configurations of the three datasets, every sensor is annotated with an identifying number in this work
• Figure 3: Weight score of sensors after training in different experiments
• Figure 4: Sensor optimization result
• Figure 5: Weight score of sampling rate after training in different experiments