Skip to content
Snippets Groups Projects
Select Git revision
  • master default protected
1 result
user avatar
update README
Jesse Raitapuro authored
34bd4c4a
History
user avatar 34bd4c4a
History
Name Last commit Last update
Explorer
FrontEnd
MachineLearning
ParkMonitor
TrafficAnalyzer
new_data/crops/unsorted
.gitignore
README.md
Utility.py
__init__.py
README.md

Future City Challenge

The Idea of Future City Challenge competition was to design innovative solutions and applications that make everyday life in cities easier.

Videos for our solutions:

All solutions made for the competition had to include the following parts:

  • Usage of LoRa-network(s)

  • Use of IBM Bluemix (or IBM Cloud) platform

  • Solution had to have at least one device using Microchip's parts and/or components

All participating cities had different challenges to solve, which had to be resolved by the competitors.

Our team decided to pick two of the Jyväskylä's challenges, which were the following ones:

  • Utilisation rate of on-street parking

Is there parking space available in Yliopistonkatu street when a parent should take their child to a speech therapist? Can this be checked somewherein advance ? Is it true that all parking spaces are always occupied?

  • Crowdsourcing - behavior and movement of the people in an urban environment

In what areas of the pedestrian street do people gather, on what day of the week and in what time of the day? How can this be assessed (for example, the telephone network, Bluetooth, cameras)?

  • Utilisation rate of on-street parking

The solution we made for the challenge used Raspberry Pi (+ pi camera), "LoRa-modem (Sodaq Explorer)" and Python programming language with Tensorflow library.

Technical information about the parking utilisation challenge can be found here

  • Crowdsourcing - behavior and movement of the people in an urban environment

For the crowdsourcing challenge, we decided to use a regular desktop computer + LoRa-modem and Python with Tensorflow library. LoRa usage for the challenge wasn't the best choice since it would have worked better with a regular network connection. Basically, the desktop computer receives images from different cameras, analyzes the received images with Tensorflow and calculated objects (vehicles and people) that were detected.

  • Both solutions used Python + Flask application platform which was running inside IBM Bluemix. This application worked as a frontend for end users

Technical information about the crowdsourcing challenge can be found here