Local LinearWavelet Neural Network Based Unscented Kalman Filter for Vehicle Collision Estimation

Abstract

The prevalence of traf ic accidents is a serious issue on a global scale and has elevated to a substantial public health issue. One of the most frequent reasons for car accidents is driver inattention, which can cause serious injuries or even death. For this reason, advancing the technology behind Advanced Driver Assistance Systems (ADAS) is essential to enhancing traf ic safety measures. Drivers can benefit from ADAS technology in a number of ways, including by getting alerts for lane departure, collision avoidance, and blind spot recognition. To identify possible threats and warn drivers to take action, these systems make use of a variety of sensors and algorithms. In some circumstances, they can even take control of the car to avoid collisions. A significant contributory factor to traf ic accidents is driver inattention. By warning drivers in advance of a probable collision, the number of fatalities and injuries brought on by accidents and careless driving can be significantly reduced. Common automobile sensors used to monitor the area surrounding the car and anticipate crashes include radar, LIDAR, and cameras. However, adverse environmental factors including weather and light radiation might af ect these sensors' accuracy. Furthermore, potential risks may go unreported if they are obstructing the sensor's line of sight or are outside of its limited field of view. To address these problems, a local linear wavelet neural network (LLWNN)- based vehicle collision estimate warning system has been developed. This research presents a vehicle collision estimate warning system based on a local linear wavelet neural network (LLWNN) and an unscented Kalman filter by combining sensor and wireless vehicle communication data. The LLWNN-based unscented Kalman filter, which also provides pertinent collision alerts for drivers, which was used to estimate the trajectory of distant objects. By implementing this method, it is hoped that the number of collisions caused by reckless driving would decrease, resulting in fewer fatalities and injuries. More research and development in this area might lead to the construction of roads that are safer for all users (drivers, bikers, and pedestrians).