2. UrbanSense Goals
Understand and get aware of environmental and behaviour phenomena
• Impact for the city
• Research platform
• Identify critical urban areas
• Open data
• Detect events in real time and automatically
• Wireless networks testbed
• Evaluate the impact of urban intervention actions
• Data analysis and pattern recognition
• Urban planning
3. Characteristics
Sensor network – static and mobile
• Hundreds of units
• Wireless communications
• 400 Mobile units in buses
• Real-time transmission
• Static units located down town
• Opportunistic communications for delay tolerant data
• Units with heterogeneous set of sensors
• Local processing capacity
• 550 environmental sensors
• Count people and vehicles locally
• 60 Video cameras
• Adaptive sampling rate
• 500 GPS, accelerometers, and On Board Devices
4. Sensors
UrbanSense includes 600 sensor units. Hererogeneous sets of sensors.
Meteorological
• Temperature
Relative Humidity
•
• Stand alone
• VOC
•
50 sensors (mobile and fixed)
75 sensors (mobile and fixed)
• Azote Dioxide
•
• Pluviometer,
Wind Vane
Anemometer
•
10 sensors (fixed)
•
75 sensors (mobile and fixed)
• Solar Radiation
•
10 sensors (fixed)
75 sensors (mobile and fixed)
•
•
•
•
50 sensors (mobile and fixed)
• Carbon Dioxide
•
•
• GPS and Accelerometer
•
50 sensors (mobile and fixed)
500 sensors (mobile)
• OBD – On Board Device
50 sensors (mobile and fixed)
• Carbon Monoxide
25 sensors (fixed)
Mobility
75 sensors (mobile and fixed)
• Particles PM10
High precision
1 sensor (fixed)
• Embebed
• Ozone (O3)
•
• Luminosity
Noise
Air Pollution
Video
• Cameras
•
60 sensors (mobile and fixed)
6. Perceiving people and vehicles anonymously
Contours computed with RPi
Local Processing
=
Anonymity
&
Light Communication
At time 15:01
8 leaved
10 entered
•
•
•
•
No video streaming
No video storage
Images described by statistics (descriptors)
Low bandwidth requirements
7. Perceiving people and vehicles anonymously
Sensing unit protype
What can be done?
• Counting people / vehicles
• Classifying vehicles
• Detecting patterns (e.g. crowding)
Where?
• Streets
• Buildings
• Public transports
8. Buses as City Scanners
Frequency map
Buses have city-wide coverage
• Scanning the city using sensors
• Detecting and predicting traffic jams
• Characterizing mobility
B
A
#vehicles / day
>= 47
40 to 46
34 to 39
27 to 33
21 to 26
14 to 20
8 to 13
1 to 7
(Sample: 108 buses, Wed 27/Nov/2013)