Machine Learning-Based Angle-Of-Arrival Estimation for Enhanced V2V Communication in Autonomous Navigation and Smart Cities

There are many applications for angle of arrival (AoA) estimations in car-to-car communications, and in this paper, machine learning methods are used to improve the accuracy of 21 AoA estimations. Several machine learning algorithms, namely Random Forest, Gradient Boosting, XGBoost, LightGBM, and Support Vector Regression, are applied to AoA estimations. The best result was obtained by Gradient Boosting, with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.9496 on freeways. The most important features were distance, frequency, and interference, as these features have a significant impact on the accuracy of AoA estimations; therefore, they are used to improve the accuracy of AoA estimations. The proposed machine learning method reduces the root mean square error (RMSE) to less than 5° on freeways and less than 10° in other scenarios, thus allowing it to be used for real-time applications. The complexity of the machine learning model is O(n log n), which enables its application in real-time scenarios for autonomous vehicles.