Real-Time Geometric Representation of Lane-Change Decision for Autonomous Vehicles Using Dynamic Optimization Algorithm

This paper develops a lane-change geometric representation that can be used in an on-road vehicle. The design of the proposed system uses the data collected from active a host vehicle and measures the relative speed between host vehicle and obstacle vehicles in real-time. The available distance to the target lanes measures the separated distance between the host and obstacle vehicles in real-time. These data are generated automatically using a dynamic environment and updated using time and object dynamics laws. The main algorithm uses the data to test the availability of using lane-change while checking if there is a slight break applied not to waste the testing time of the testing, then the smart searching algorithm applies to find the most suitable solution to change the lane safely as fast as possible. The desired system has been achieved with accepted real-time results that exceed the standard driver response time. Also, full accepted safety results allow the vehicle to get out of the lane with particular speed and steering angles to avoid collisions and avoid crashes in the host vehicle.