Vehicle Operation Classification

Title Vehicle Operation Classification
Summary Classify modes of operation of Volvo vehicles based on on-board data
Keywords Data Mining, Machine Learning, Ubiquitous Knowledge Discovery
TimeFrame Winter 2016 / Spring 2017
References Time series classification

Unsupervised and semi-supervised clustering


Prerequisites Artificial Intelligence and Learning Systems courses
Author Karthik Bangalore Girijeswara
Supervisor Sławomir Nowaczyk, Yuantao Fan, Mohamed-Rafik Bouguelia
Level Master
Status Ongoing

Generate PDF template

With rapid development and growth of interconnected devices, more physical systems are integrated with computer-based systems, e.g. sensors and actuators can be sensed and controlled remotely across the network. It is enticing to analyze sensor data streaming from devices (on large scale), discover interesting patterns and knowledge.

In the ReDi2Service project we have collected approximately 1TB of data from Volvo buses in normal operation. It is interesting to analyse this data from the usage point of view, and come up with good description and/or classification (possibly hierarchical or even an ontology) of vehicle operation.

The data contains both GPS positions, as well as on-board signals such as vehicle speed or engine torque. The goal of the project is to provide a framework for organising the data in a way that will facilitate future access to interesting portions of this data according to multiple criteria.

We are interested in both low-level information (for example, detecting workshop visits, splitting the data into trips from turning the engine on to turning it off, distinguishing between highway and in-city operation, or finding uphill and downhill driving) as well as in more abstract description (is the bus in regular line traffic or performing some other duty, automatically finding similarities and differences between missions, etc).

More details to come...