Smart Cities | IEEE INFOCOM 2017
Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing Architectures for Smart Cities Andrea Capponi
University of Luxembourg
Claudio Fiandrino
Imdea Networks Institute
Dzmitry Kliazovich
ExaMotive
Pascal Bouvry Stefano Giordano
University of Luxembourg Università di Pisa May 1, 2017
Smart Cities: Introduction I
50% of worldwide population lives in cities
I
Cities account for I
80% of worldwide gas consumption
I
75% of global energy consumption
I
60% of residential water use
N. B. Grimm, S. H. Faeth, N. E. Golubiewski, C. L. Redman, J. Wu, X. Bai, and J. M. Briggs, “Global change and the ecology of cities,” in Science, vol. 319, no. 5864, 2008, pp. 756–760.
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
1 of 14
Sensing as a Service (S2 aaS) for IoT Sensors
Consumers
Cloud Providers
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
2 of 14
Mobile Crowdsensing I
Appealing paradigm for sensing and collecting data I
Monitoring phenomena in smart cities
I
Sensing as a Service (S2 aaS) business model
I
Sensors commonly available in mobile and IoT devices
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
3 of 14
Mobile Crowdsensing I
Appealing paradigm for sensing and collecting data I
Monitoring phenomena in smart cities
I
Sensing as a Service (S2 aaS) business model
I
Sensors commonly available in mobile and IoT devices
Figure: Diffusion of mobile devices Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
3 of 14
Mobile Crowdsensing I
Appealing paradigm for sensing and collecting data I
Monitoring phenomena in smart cities
I
Sensing as a Service (S2 aaS) business model
I
Sensors commonly available in mobile and IoT devices
Figure: Diffusion of mobile devices Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
3 of 14
Sensing Paradigms
I
Two sensing paradigms:
Participatory I
Opportunistic
Active user engagement
I
Minimal user involvement
I
Sensing and reporting user-driven
I
Sensing and reporting application- or device-driven
I
Centralized, direct task assignment
I
Distributed, no direct task assignment
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
4 of 14
Mobile Crowdsensing Scenario Environmental Context Cs
S2 aaS Cloud Collector
Crowd iFi W
ta Da
Data Collection Utility dsa
E LT
IoT Devices
Threshold
Accelerometer
Gyroscope
Microphone
Dual Camera
Temperature
Smartphone Sensing Potential sps
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
5 of 14
Mobile Crowdsensing Scenario Environmental Context Cs
S2 aaS Cloud Collector
Crowd iFi W
ta Da
Data Collection Utility dsa
E LT
IoT Devices
Threshold
Accelerometer
Gyroscope
Microphone
Dual Camera
Temperature
Smartphone Sensing Potential sps
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
5 of 14
Proposed Opportunistic Framework
Sensing Decision Policy Cs · [ · sps + (1
) · dsa ] >
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
6 of 14
Proposed Opportunistic Framework
Data Collection Utility
Threshold
Sensing Decision Policy Cs · [ · sps + (1
Environmental Context
) · dsa ] >
Smartphone Sensing Potential
Balancing Coefficient
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
6 of 14
Proposed Opportunistic Framework I
Data Collection Utility dsa I I
I
Defines utility for the cloud collector Based on number of samples already received
Smartphone Sensing Potential, sps I I
Based on number of samples already received Function of locally spent energy Es for sensing and reporting Es = Esc + Esr
I
Environmental Context Cs I
I
Binary value which depends on the context awareness
Threshold I
Depends on the level of battery of the devices (B) and the amount of reported data that devices have already contributed (D)
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
7 of 14
CrowdSenSim I
Custom simulator for crowdsensing activities I
Access and download: http://crowdsensim.gforge.uni.lu
I
Contact:
[email protected]
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
8 of 14
CrowdSenSim: Features and Architecture I
Large scale (time-space)
I
Realistic urban environments City Layout
List of Events
CrowdSensing Inputs
User Mobility
S IMULATOR
Results
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
9 of 14
Evaluation Settings
I
3 sensors: accelerometer, temperature, pressure
I
Users: 20 000
I
Walking speed: U [1
I
Walking period: U [10
I
Simulation period: 8 AM - 2 PM
I
Digipoint to provide street-level maps
I
City centers of Luxembourg, Madrid and Trento
1.5] m/s 20] min
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
10 of 14
Performance Evaluation
Traces
0-50 50-100 100-150 150-200 200-250 250-300 300-350 350-400 400-450 450-500 500-550 550-600 600-650 650-700 700-750 750-800 800-850 850-900
Uniform
Current drain (µAh)
(a) Sensing Cost
3 500 3 000 2 500 2 000 1 500 1 000 500 0
Uniform
Traces
0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60-65 65-70 70-75 75-80 80-85
3 500 3 000 2 500 2 000 1 500 1 000 500 0
Num. of Participants
Energy spent for sensing and communications
Num. of Participants
I
Energy (J)
(b) Communication Cost
I
Sensing and communication same distribution but different scales
I
Real-world traces are the results of a study on pedestrian mobility and are public available on Crawdad (ostermalm_dense_run2) 1 1 http://crawdad.org/kth/walkers/20140505
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
11 of 14
Performance Evaluation
I
Normalized distribution of amount of collected data for the different cities over the time period 8:00 AM - 2:00 PM
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
(a) Luxembourg
1
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
(b) Trento
1
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
(c) Madrid
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
12 of 14
Performance Evaluation New metric called Sample Distribution (SD): SD = Nsa |t /
=
Pn
i,j i j
d(i, j)
n(n
1)
Area 1
Area 3
Area 4
Area 5
150
150
150
150
150
100
100
100
100
100
50
50
50
50
50
0
0
0
0
0
Time interval
Time interval
8: 00 9: -9:0 00 0 10 -10 :0 :00 0 11 -11 :0 :00 0 12 -12 :0 :00 0 13 -13 :0 :00 014 :0 0
200
8: 00 9: -9:0 00 0 10 -10 :0 :00 0 11 -11 :0 :00 0 12 -12 :0 :00 0 13 -13 :0 :00 014 :0 0
200
8: 00 9: -9:0 00 0 10 -10 :0 :00 0 11 -11 :0 :00 0 12 -12 :0 :00 0 13 -13 :0 :00 014 :0 0
200
8: 00 9: -9:0 00 0 10 -10 :0 :00 0 11 -11 :0 :00 0 12 -12 :0 :00 0 13 -13 :0 :00 014 :0 0
200
Time interval
I
Area 2
200
9: -9:0 00 0 10 -10 :0 :00 0 11 -11 :0 :00 0 12 -12 :0 :00 0 13 -13 :0 :00 014 :0 0
Sample Distribution (sample/m)
2
8: 00
I
Time interval
Time interval
Sample distribution (SD) for the different areas over the time period 8:00 AM - 2:00 PM
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
13 of 14
Conclusion The proposed framework I
MCS systems for S2 aaS applications in smart cities
I
Data collection in opportunistic MCS systems
I
Energy-efficient and cost effective
CrowdSenSim I
Performance evaluation in large realistic urban environments
I
Custom simulator for crowdsensing activities I
Access and download: http://crowdsensim.gforge.uni.lu
I
Contact:
[email protected]
Andrea Capponi | Smart Cities | IEEE INFOCOM 2017 | Energy Efficient Data Collection in Opportunistic Mobile Crowdsensing
14 of 14
Thank You! Andrea Capponi
