Detection of Road Artefacts Using Fuzzy Adaptive Thresholding

In this paper the authors are proposing approximate method for road artefacts detection and their location by analyzing acceleration values recorded in the car during driving over the road fragment using the smartphone mounted in the car. The new method called F-THRESH has been introduced, which is adaptively adjusting threshold for road artefacts detection by the fuzzy system means, allowing for outlier detection in chaotic time streams. First, the road quality is being calculated, then the difference between the current data point and mean acceleration is calculated and those two values are used as the input for the fuzzy system, which is calculating threshold to classify data point as an outlier. The proposed method has been compared to the previously implemented method and has an accuracy over 94% with 1.3% of False Positive Rate for the same problem which makes it a great candidate to be implemented in the IoT Edge scenarios, for reducing amount of data being sent to the cloud analyzing system.