Determination of road surface conditions using mobile handset

Determination of road surface conditions using mobile handset based accelerometers Student: Conor Molloy Supervisor: Liam Kilmartin

Presentation Overview • Android • Project Overview. • Accelerometer • GPS • Mat. Lab

Android • Android was bought by Google in 2003 • Android is used in 54% of smartphones. • Its open source which allows anyone to develop applications for it. • It uses the Java programming language.

Project Overview • Using Android add-on for Java, which allows developers to write applications for Android smartphones. • Access the smartphone’s sensors, specifically GPS and the Accelerometer. Using permissions. • Write the data from the sensors to a file. • Import the file to Mat. Lab for analysis and visualization.

Accelerometer • Measures the forces experienced by the smartphone on 3 axis. • Can be used to determine the orientation of the phone.

Accelerometer • The phone will most likely not be perfectly vertical so the region between the Z and Y axis is where the vibrations will occur. • Using Eular’s angles the orientation of the phone can be used to correct the data gathered so that the vibrations will be measured on the true vertical axis.

GPS • Use satellites to get the phones position on earth. • Network of 24 satellites • Phone requires connection to 4 of these to get its position.

Code used Accelerometer Code GPS Code

Writing Data to File • Use Java’s built in function to write and create files. • The file is created and then written to using this method. • Writes the accelerometer and location data to a. csv file so it can be used in Mat. Lab.

Start Matlab Import file • Used for mathematical analysis. • Can import. txt files, but need to be. csv for matrix manipulation. • Will be used to calculate the variance of the accelerometer data, which will be divided into frames. Divide data into frames Calculate variance of each frame All frames done Yes Plot variance No

Future Work • Compensate for speed of vehicle. • System to compare with known road surface conditions • Implement server side infrastructure for recording data from multiple devices. • Implement Android application as a service with an unobtrusive mechanism for activation while driving • Investigation of variation of results across handsets and vehicles and implementation in Matlab of methodology for aggregation of data (including rejection of outlier data) on server side

Questions?