Mobile users can open program in new tab for better viewing.

Open program in new tab

Day 1 03/12/2020
Room #1

An opening speech by General Chair 13:00 - 13:05

Rui José

A welcome message by EAI 13:00 - 13:05

Elena Davydova

A welcome message by EAI Community Manager 13:05 - 13:10

Michal Dudic

EAI Urb-IoT 2020 Presentations 13:10 - 15:40

13:10 - 13:40
A Crowd-sourced Obstacle Detection and Navigation App for Visually Impaired

Abstract. Individuals with sight impairments rely heavily on various types of travel-aid when navigating their ways across their neighborhoods. Recently, there have been many breakthrough technologies that focus on the visually impaired by providing solutions such as wearable bands and optical wearable devices. However, such technologies are costly and not suited for the general market. Others have started investigating smartphone applications as a much more widely available solution but with limited applicability on outdoor barriers and obstacles that these groups of people face in their day to day journeys. In this work, we propose GeoNotify, a smartphone application which is tailored to detect unexpected temporary obstacles that could cause injury to visually impaired people. We present how advances in Convolutional Neural Networks merged with crowd-sourcing methodologies could be used to build more accurate models capable of recognizing wide representations of the real-world obstacles.
Authors: Edward Kim (University of Washington), Joshua Sterner (University of Washington), Afra Mashhadi (University of Washington),
Hide Authors & Abstract

Show Authors & Abstract
13:40 - 14:15
An ecosystem approach to the design of sensing systems for bicycles

Bicycles equipped with sensors, processing capacity and communications can be a promising source of data about the personal and the collective reality of urban cycling. While this concept has been attracting considerable interest, the key usual assumption is the design of a closed system where a uniform set of sensing bicycles, with a concrete set of sensors, is used to support a specific service. The core challenge, however, is how to generalise sensing approaches so that they can be collectively supported by many heterogeneous bicycles, owned by a multitude of entities, and integrated into a common ecosystem of urban data. In this work, we provide a comprehensive analysis of the design space for onbike sensing. We consider a diverse set of sensing alternatives, the potential value propositions associated with their data, and the collective perspective of how to optimise sensing by exploring the complementarities between heterogeneous bicycles. This broader perspective should inform the design of more effective sensing strategies that can maximise the overall value generated by bicycles in smart cycling ecosystems and enable new cycling services.
Authors: Rui José (University of Minho), Ricardo Cabral (Centro Algoritmi, University of Minho, Portugal), Eduardo Peixoto (Centro Algoritmi, University of Minho, Portugal), Carlos Carvalho (Centro Algoritmi, University of Minho, Portugal),
Hide Authors & Abstract

Show Authors & Abstract
14:15 - 14:40
Calibration of Low-cost Particulate Matter Sensors with Elastic Weight Consolidation (EWC) as an Incremental Deep Learning Method

Urban air quality is an important problem of our time. Since high precision monitoring stations cannot capture the temporal and spatial dynamics in the urban atmosphere, low-cost sensors must be used to setup dense measurement grids. However, low-cost sensors are imprecise, biased and susceptible to environmental influences. While neural networks have been explored for their calibration, issues include the amount of data needed for training, requiring sensors to be co-located with reference stations for extensive periods of time. Also re-calibrating them with new data can lead to catastrophic forgetting. We propose using Elastic Weight Consolidation (EWC) as an incremental calibration method. By exploiting the Fisher-Information-Matrix it enables the network to compensate for different sources of error, both pertaining to the sensor itself, as well as caused by varying environmental conditions. Models are pre-calibrated with data of 40 hours measurement on a low-cost SDS011 PM sensor and then re-calibrated on another SDS011 sensor. Our evaluation on 1.5 years of real world data shows that a model using EWC with a time period of data of 6 hours for re-calibration is more precise than models without EWC, even those with longer re-calibration periods. This demonstrates that EWC is suitable for on-the-fly collaborative calibration of low-cost sensors.
Authors: Rainer Schlund (Karlsruhe Institute of Technology (KIT), TECO), Johannes Riesterer (Karlsruhe Institute of Technology (KIT), TECO), Marcel Köpke (Karlsruhe Institute of Technology (KIT), TECO), Michal Kowalski (Helmholtz Zentrum München, German Research Center for Environmental Health (HMGU)), Paul Tremper (Karlsruhe Institute of Technology (KIT), TECO), Matthias Budde (Karlsruhe Institute of Technology (KIT), TECO), Michael Beigl (Karlsruhe Institute of Technology (KIT), TECO),
Hide Authors & Abstract

Show Authors & Abstract
14:40 - 15:10
Person-Flow Estimation with Preserving Privacy using Multiple 3D People Counters

The spread of mobile phones made it easy to estimate person-flow for corporate marketing, crowd analysis, and countermeasures for disaster and disease. However, due to recent privacy concerns, regulations have been tightened around the world and most smartphone operating systems have increased privacy protection. To solve this, in this study, we propose the person-flow estimation technique with preserving privacy. We use 3D People Counter which can record only the time and direction of passing people, a person's height, and walking speed, therefore it preserves privacy from the moment of collecting data. To estimate people's in-out data, we propose four methods and they use some of the sensor data above in different combinations. We compared these methods and the height-based method could estimate about 79% of the sensor data as in-out data. Additionally, we also created a system to interpolate in-out data into person-flow data and to visualize it. By using this method, we believe that it can be used for the purposes described in the beginning.
Authors: Yoshiteru Nagata (Nagoya University, Nagoya, Aichi, Japan), Takuro Yonezawa (Nagoya University, Nagoya, Aichi, Japan), Nobuo Kawaguchi (Nagoya University, Nagoya, Aichi, Japan),
Hide Authors & Abstract

Show Authors & Abstract
15:10 - 15:40
Quality and Reliability Metrics for IoT Systems: A Consolidated View

Quality and reliability metrics play an important role in the evaluation of the state of a system during the development and testing phases, and serve as tools to optimize the testing process or to define the exit or acceptance criteria of the system. This study provides a consolidated view on the available quality and reliability metrics applicable to Internet of Things (IoT) systems, as no comprehensive study has provided such a view specific to these systems. The quality and reliability metrics categorized and discussed in this paper are divided into three categories: metrics assessing the quality of an IoT system or service, metrics for assessing the effectiveness of the testing process, and metrics that can be universally applied in both cases. In the discussion, advises to proper usage of discussed metrics in a testing process are then given.
Authors: Matej Klima (Czech Technical University in Prague), Vaclav Rechtberger (Czech Technical University in Prague), Miroslav Bures (Czech Technical University in Prague), Xavier Bellekens (University of Strathclyde), Hanan Hindy (Abertay University), Bestoun Ahmed (Karlstad University),
Hide Authors & Abstract

Show Authors & Abstract

A closing speech by General Chair 15:40 - 15:45

Rui José