Smartphone location identification and transport mode recognition using an ensemble of generative adversarial networks

Günthermann, Lukas, Simpson, Ivor and Roggen, Daniel (2020) Smartphone location identification and transport mode recognition using an ensemble of generative adversarial networks. UbiComp/ISWC '20, Virtual Event Mexico, September, 2020. Published in: Adjunct Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers. 311-316. ACM, NY, United States. ISBN 9781450380768

[img] PDF - Published Version
Restricted to SRO admin only

Download (632kB)
[img] PDF - Accepted Version
Download (600kB)

Abstract

We present a generative adversarial network (GAN) approach to recognising modes of transportation from smartphone motion sensor data, as part of our contribution to the Sussex-Huawei Locomotion-Transportation (SHL) recognition challenge 2020 as team noname. Our approach identifies the location where the smartphone of the test dataset is carried on the body through heuristics, after which a location-specific model is trained based on the available published data at this location. Performance on the validation data is 0.95, which we expect to be very similar on the test set, if our estimation of the location of the phone on the test set is correct. We are highly confident in this location estimation. If however it were wrong, an accuracy as low as 30% could be expected.

Item Type: Conference Proceedings
Schools and Departments: School of Engineering and Informatics > Informatics
SWORD Depositor: Mx Elements Account
Depositing User: Mx Elements Account
Date Deposited: 14 Oct 2020 06:55
Last Modified: 14 Oct 2020 11:10
URI: http://sro.sussex.ac.uk/id/eprint/94358

View download statistics for this item

📧 Request an update