Integrating Multi-Sensor Data With Layering Methods For Robust Autonomous Vehicle Navigation

This paper focuses on the design of a new data processing and categorization method using sensor data fusion framework for autonomous vehicle. The objective is to enhance the accuracy, efficiency and reliability in independent driving systems employing the data collected from different sensors using deep learning techniques. The input data is placed into two groups, one which includes data to be processed while the other contains data to be pre-processed. The information is processed using a new Fused Mesh Layering Method with sensor data fusion through Machine Learning, RNN and CNN. As a result, the obtained augmented layout is much more accurate and efficient for tracking the vehicles, identification of abnormalities and identification of obstacles. Also, the time period required for the execution of all methods added up is less as compared to the current frameworks.