Recognition of Lane Changing Maneuvers for Vehicle Driving Safety

The increasing number of vehicles has caused traffic conditions to become increasingly complicated in terms safety. Emerging autonomous (AVs) have the potential significantly reduce crashes. advanced driver assistance system (ADAS) received widespread attention. Lane keeping and lane changing are two basic driving maneuvers on highways. It is very important for ADAS technology identify them effectively. maneuver recognition been used study safety many years. Different models proposed. With development technology, machine learning introduced this field with effective results. However, which require a lot physical data as input unaffordable sensors lead high cost AV platforms. This impedes AVs. proposes model based distinct set data. scenario from natural vehicle trajectory dataset (i.e., HighD) learning. Acceleration velocity extracted labeled normalized features then into k-nearest neighbor (KNN) classification model. trained was applied another good results show that acceleration features, accuracy (LK), left (LCL) right (LCR) 100%, 97.89% 96.19%.