# ConnectedCars – A hobbyist design

This is a 3 min read

# ConnectedCar – A Social Innovation

Now we know the concept and intent. Connected cars, a fledgling of the IoT innovations needs an ecosystem to see the light of the day. Here is the list of top 10 obstacles for connected cars according to Forbes.com. It predominantly talks about how players of this ecosystem like Car manufacturers, OEMs and MNPs will have a hard time collaborating to make this happen.

# Experimental Design

While above all are true and are to be considered. Here is a much simplified, cheaper and a toned down version of connected car targeting only very few use cases when compared to original concept.   This is  ONLY to demonstrate the facets of a connected car as a hobbyist. This design can target ONLY basic use cases like reporting car’s vital parameters, reporting Co2 emissions, vehicle-to-Mobile app interaction, location broadcasting and tracking. This inexpensive and less intrusive design won’t include a imbedded OS for the cars.  The major implications on the inner workings are, all the communications between car and mobile app is routed through the cloud and/or push messaging mechanisms.

Architecture components

• In-vehicle module (Gemalto concept board  + Arduino UNO compatible GSM Shield)
• OBD II connector cable for Arduino UNO.
• Mobile Companion App (Android App to talk to in-vehicle module)
• IoT data platform – we got different options here
• ThingSpeak – Free
• ThingFabric by 2lemtery.com – $$• Carriots –$$
• custom big data platform

Lets see how these will connect and work together.

The use case explained above is very basic and it can be extended to make other simple use cases work.  By applying some formulae on the vehicle OBD II parameters we could calculate the vehicle’s Co2 emission status. We could design a configurable  system lets say an API call to IoT platform will be made if the level of Co2 emissions are beyond the threshold.

As for as the platform the important takeaway is while we could certainly design a custom data platform using any Hadoop flavours available today I would strongly recommend IoT centric data platforms like 2lemetry.com’s ThingFarbic.  Here are some of the reasons,

• They have taken the basic horizontally scalable parallel processing Big data platforms to the next level with the addition of some real time rule engines and Complex event processors (CEP). The bells and whistles these IoT data platforms comes with could save a lot of time for a IoT centric project like this.
• The protocols they support for event submission from connected devices are IoT centric and device optimised for bandwidth like MQTT , STOMP etc..

Hope this all make sense.