AutoProfile: an Intelligent Profile Switching System for Smartphones
Wei Yang,Sheng Chen,Zhengyang Wang,Yang Xu,Liusheng Huang
DOI: https://doi.org/10.1109/tmc.2022.3141205
IF: 6.075
2023-01-01
IEEE Transactions on Mobile Computing
Abstract:Smartphones have been the necessities for us due to their advanced computing capabilities and ubiquitous connectivity to our daily lives. However, they also produce many negative influences, such as ring noise, nuisance calls, which interrupt people’s attention when working. It would be user-friendly if smartphones can automatically sense the surroundings and dynamically work at an appropriate profile to prevent their ringing from disturbing people in some special circumstances. To address this issue, in this paper, we propose a novel smartphone profile switching system, called AutoProfile, which combines the techniques of acoustic sensing, walk detection and machine learning to automatically and dynamically change smartphones’ profiles in different scenarios. We develop a new compact ambient sound scheme for feature extraction, named DWT & MFCC fingerprint, which can effectively distinguish between different social scenarios and outperforms the existing method. To evaluate the performance of AutoProfile, we conduct experiments in 8 scenarios and take multiple influence factors into consideration. The results demonstrate that AutoProfile can realize overall recognition accuracies of $91.4 \;\%$ and $90.6 \;\%$ when using Random Forest and $k$ -nearest Neighbors classifiers, respectively. Moreover, since AutoProfile senses the ambient sound passively, it does not create additional noise compared with some active acoustic sensing schemes. In addition, the power consumption of AutoProfile is acceptable, and thus AutoProfile can be tailored as a background service of smartphones to make them become “smarter”.