Abstract | ||
---|---|---|
This poster investigates the discrimination problem between car driver and car passengers using smartphones, which is critical to active safety enhancement and usage based insurance. The proposed system leverages a smartphone without the assistance of extra devices or resources, e.g., car speakers, bluetooth network, Internet or cloud. Specifically, our system builds a unsupervised machine learning platform on Android based smartphones, in which the data of the Inertial Measurement Unit (IMU), i.e., gyro, accelerator, magnetometer, is taken as the input. With the attitude and trajectory data generated from the IMU measurement, the neural network extract the fundamental features of the spacial characteristics of driver and passengers during accelerating, deceleration, turning, starting, stopping, driving, based on which we may further distinguish driver and passengers. The experiments show that the proposed system is able to provide a classification accuracy over 95%, at a low false positive rate. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/PAC.2017.37 | 2017 IEEE Symposium on Privacy-Aware Computing (PAC) |
Keywords | Field | DocType |
egde computing,driver detection,smart phone,neural network | Android (operating system),Computer science,Real-time computing,Unsupervised learning,Inertial measurement unit,Artificial neural network,Active safety,Bluetooth,Cloud computing,The Internet | Conference |
ISBN | Citations | PageRank |
978-1-5386-1028-2 | 0 | 0.34 |
References | Authors | |
2 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Haihua Gong | 1 | 0 | 1.01 |
Xing Kai | 2 | 442 | 28.13 |
Wenwen Du | 3 | 0 | 1.01 |