Self-driving cars is a trending topic of the modern world. The ability to control a vehicle without human interaction can significantly increase safety on the road. In this research project, the main focus is on making the behavior of self-driving cars more like it was driven by a human driver in situations when pedestrians or animals are close by. While self-driving cars are equipped with cameras they can record and process the situations on the road and respond accordingly. It can be accomplished by image processing techniques, differentiation of objects, if they are alive or not, if they are moving or not, whether they are human or not and even from the determination of pedestrian’s age. Analysis of such data shall provide a prediction of an object behavior. The algorithm that is being developed shall give a car increased safety for pedestrians and animals that may potentially appear in the vehicle’s path. Computer vision library OpenCV, image processing, machine learning techniques and neural networks will be combined and used to build this algorithm. Presented is the current progress of this research project to design machine learning algorithms for the behavioral prediction of objects for self-driving cars.
Byeloborodov, Yevgeniy and Rashad, Sherif, "Design Of Machine Learning Algorithms For Behavioral Prediction Of Objects For Self Driving Cars" (2021). 2021 Celebration of Student Scholarship - Oral Presentations. 64.