During my diploma thesis, I designed and implemented a pothole detector system based on the analysis of inertial sensor data.
I studied the currently available sensor data analysis based methods, and solutions for pothole detection, and the models that those methods were based on. Based on these information, I created an advanced model, and a more robust detection algorithm. The algorithm identifies the unusual patterns in the sensors acceleration and angular rate data by monitoring its standard deviation. This detection algorithm was tested on several data sets form different sensors.
Using a NXP LPC1227 microcontroller and a ST LSM330DLC inertial sensor, I created a two-component hardware configuration, and I implemented an embedded c realization of the pothole detection algorithm, to build up the complete pothole detector system.
The optimal parameters for the algorithm were determined during a measurement session, where real-life traffic conditions, and situations were simulated. The system’s proper operation was verified with measurement.