On Neural Inertial Classification Networks for Pedestrian Activity Recognition
This work addresses the problem of fair comparison and evaluation for researchers in inertial sensing, but it is incremental as it focuses on benchmarking existing techniques rather than introducing new methods.
This paper tackles the lack of a common benchmark for neural inertial classification networks in pedestrian activity recognition by analyzing ten data-driven techniques, finding that data augmentation through rotation and multi-head architecture yields the most significant improvements across four datasets with over 936 minutes of inertial data.
Inertial sensors are crucial for recognizing pedestrian activity. Recent advances in deep learning have greatly improved inertial sensing performance and robustness. Different domains and platforms use deep-learning techniques to enhance network performance, but there is no common benchmark. The latter is crucial for fair comparison and evaluation within a standardized framework. The aim of this paper is to fill this gap by defining and analyzing ten data-driven techniques for improving neural inertial classification networks. In order to accomplish this, we focused on three aspects of neural networks: network architecture, data augmentation, and data preprocessing. The experiments were conducted across four datasets collected from 78 participants. In total, over 936 minutes of inertial data sampled between 50-200Hz were analyzed. Data augmentation through rotation and multi-head architecture consistently yields the most significant improvements. Additionally, this study outlines benchmarking strategies for enhancing neural inertial classification networks.