Gait-based Authentication: Evaluation of Energy Consumption on Commercial Devices

Smartphones and smartwatches, thanks to the sensors they are equipped with, provide a suitable platform for implementing and deploying motion-based behavioral biometrics. In this paper, a gait-based authentication method is designed and implemented for commercially available devices, in order to fill the gap between prototypical and real-world solutions. Particular attention is given to the energy requirements of the system. To this purpose, the impact of the sampling frequency on the accuracy vs consumption trade-off is first analyzed. Then, several architectural configurations are compared to assess the benefits obtained by executing the whole method on the smartwatch, instead of offloading computation to a smartphone. Results show that a smartwatch-only solution is more energy-efficient, as the power consumption due to transferring the data to an external device outweighs the benefits deriving from the decreased computational load.