Machine learning approach for predicting bumps on road

In today´s days, due to increase in number of vehicles the probability of accidents are also increasing. The user should be aware of the road circumstances for safety purpose. Several methods requires installing dedicated hardware in vehicle which are expensive. so we have designed a Smart-phone based method which uses a Accelerometer and GPS sensors to analyze the road conditions. The designed system is called as Bumps Detection System(BDS) which uses Accelerometer for pothole detection and GPS for plotting the location of potholes on Google Map. Drivers will be informed in advance about count of potholes on road. we have assumed some threshold values on z-axis(Experimentally Derived)while designing the system. To justify these threshold values we have used a machine learning approach. The k means clustering algorithm is applied on the training data to build a model. Random forest classifier is used to evaluate this model on the test data for better prediction.