A Smart User Authentication Approach using Sensing Seat

The support of smart technologies has greatly contributed to the emergence of smart office with various personalized services and applications. User authentication is one of the critical technologies enabling sectional services according to individual demands. This paper proposed a smart approach to identifying users based on a sensing chair, which is convenient and intuitive without extra actions from users. Our proposed method shows an overall classification accuracy of 92.2 % in a 33 subjects experiment using KNN algorithm, which uses a set of designed features to present the sitting pressure distribution. We also investigated the effects of density of uniform sensor distribution on user authentication accuracy. The original sensor placement (1024 uniformly-distributed sensors) was reduced and evaluated accordingly. The results indicated that a reasonably acceptable performance can be achieved using much fewer sensors (an overall classification accuracy of 87.7% with 64 uniformly-distributed sensors).