This research is part of the ’LINC project’. This project is co-financed by the European Regional and Development Fund through the Urban Innovative Actions Initiative. The LINC project consists of a larger consortium led by Gate21. The consortium includes Municipality of Albertslund, Municipality of Gladsaxe, Nobina Danmark A/S, IBM Danmark ApS, Roskilde University (RUC) and The Technical University of Denmark (DTU). The project is funded by project partners and the EU programme Urban Innovative Actions (UIA), which is supporting the project with 25 million DKK.
L ARGE S CALE PASSENGER D ETECTION WITH S MARTPHONE /B US I MPLICIT I NTERACTION AND M ULTISENSORY U NSUPERVISED C AUSE - EFFECT L EARNING
arXiv:submit/4174401 [cs.LG] 20 Feb 2022
A P REPRINT Valentino Servizi Department of Technology, Management and Economics Technical University of Denmark (DTU) valse@dtu.dk Dan R. Persson Department of Applied Mathematics and Computer Science DTU Per Bækgaard Department of Applied Mathematics and Computer Science DTU
Francisco C. Pereira Department of Technology, Management and Economics DTU
Hannah Villadsen Department of People and Technology Roskilde University Denmark
Jeppe Rich Department of Technology, Management and Economics DTU
Otto A. Nielsen Department of Technology, Management and Economics DTU
February 20, 2022
A BSTRACT Intelligent Transportation Systems (ITS) underpin the concept of Mobility as a Service (MaaS), which requires universal and seamless users’ access across multiple public and private transportation systems while allowing operators’ proportional revenue sharing. Current user sensing technologies such as Walk-in/Walk-out (WIWO) and Check-in/Check-out (CICO) have limited scalability for large-scale deployments. These limitations prevent ITS from supporting analysis, optimization, calculation of revenue sharing, and control of MaaS comfort, safety, and efficiency. We focus on the concept of implicit Be-in/Be-out (BIBO) smartphone-sensing and classification. To close the gap and enhance smartphones towards MaaS, we developed a proprietary smartphonesensing platform collecting contemporary Bluetooth Low Energy (BLE) signals from BLE devices installed on buses and Global Positioning System (GPS) locations of both buses and smartphones. To enable the training of a model based on GPS features against the BLE pseudo-label, we propose the Cause-Effect Multitask Wasserstein Autoencoder (CEMWA). CEMWA combines and extends several frameworks around Wasserstein autoencoders and neural networks. As a dimensionality reduction tool, CEMWA obtains an auto-validated representation of a latent space describing users’ smartphones within the transport system. This representation allows BIBO clustering via DBSCAN. We perform an ablation study of CEMWA’s alternative architectures and benchmark against the best available supervised methods. We analyze performance’s sensitivity to label quality. Under the naïve assumption of accurate ground truth, XGBoost outperforms CEMWA. Although XGBoost and Random Forest prove to be tolerant to label noise, CEMWA is agnostic to label noise by design and provides the best performance with an 88% F1 score. Keywords Device-to-device · Sensor-to-sensor · Ground-truth-validation · Wasserstein-auto-encoders · Autonomousvehicles
1