Smart detection of vehicle accidents using object identification sensors with artificial intelligent systems

Abstract

According to Government of India, around 1,46,000 people [20] lost their lives in five hundred thousand of accidents, where 7% of lives could have been saved if they would have got medical attention before-hand [21]. This can be achieved by intimating the accident to the nearby emergency unit in minimal time by using artificial intelligence based on the severity of accidents. In existing methods, the accidents are detected using On-based unit, and transmitted to the control unit using nearby antennas, where the severity of the accidents are classified using data-mining. Then the fetched data is compared with existing accident dataset which it is retrieved from previous accidents, the analysed results are then transmitted to the nearby emergency unit [1], [2]. This will lead to ambiguous prediction of data because if the data doesn’t exist in the database, intensity of accident must be analysed manually in which it leads to increase in time complexity for transmitting the data to the nearby emergency unit due to intermediate infrastructure. To overcome these drawbacks, in this proposed system the accidents are detected using sensors and the severity of accident will be calculated using machine learning algorithms like k-means clustering and Support vector machine (SVM) classification under reinforcement learning with help of force and impact obtain while vehicle crashes, then the values are transmitted to the nearby emergency unit using Breadth-first-search in the form of A* Search algorithm.