Vehicle Data Collection for Cloud Processing
Kang Shin, PhD
Kevin and Nancy O’Conner Professor of Computer Science; Founding Director, Real-time Computing Laboratory, Electrical Engineering and Computer Science, U-M College of Engineering
May 2016 – April 2018
Graduate Student Research Assistant, Electrical Engineering and Computer Science, U-M College of Engineering
Build a flexible and extensible vehicular test bed called CarLab, which:
- Collects data from the CAN bus, installed sensors, and smartphones on each car
- Transfers data to, and processes it in the cloud
CarLab is designed to collect multi-faceted sensory data from consumer vehicles. This is in contrast to existing test beds and will therefore lead to unique insights. Furthermore, it is designed to return value to test bed participants through useful applications, thereby attracting more participants for the test bed.
Our approach consists of an in-vehicle hub (Raspberry Pi or ARTIK 5/10), OBD scanner, and a smartphone app. The cloud components involve a server to collect the data and a visualization website. Additionally, we will build several data analytics applications on top of the platform to demonstrate the test bed and provide utility back to the users.
We implemented and released the following:
- CarLab Android app
- CarLab-plugin-development library
- Web-based dashboard to visualize and download data
- Web-based data collection campaign management
Additionally, we are releasing the source code of CarLab on Github for these components, and a website was written that documents the use of CarLab.