Abstract

Self-driving cars or also known as Autonomous Vehicles have incorporate with sensors, actuators, sophisticate algorithms, machine learning systems, and high-level processors to run the software. These are vehicles that are programmed to move and be run without the need of a human being. They are of different categories depending on the low-level to high-level automation. The focus of this research is on the ability of SVM to forecast the effectiveness of self-driving cars. Especially in the context of safety and efficiency of autonomous driving considerable importance is given to the prediction of vehicle behaviour. SVM is used in this study to analyse and forecast the performance out the different sensors and environmental measures. The conclusion drawn based on the analysis of results show that the application of SVM enhances the improvement of the level of reliability of the autonomous systems. Thus, the study is demonstrated on various datasets, namely, the KITTI datasets with up to 95.64% accuracy. Finally, K-folds cross validation is performed for the robustness of the Model.