Distance Measurement for Self-driving Vehicles Using Data Fusion and Machine Learning

Abstract

For certain mobile robots and self driving vehicles, accurate measurement of distance ahead of them is indispensable. Several sensors are utilized to achieve this. This work is on fusing the obtained data from two sensors namely Leddar M-16 and RPLidar 360. A comparison of the accuracy of the distance measured from the vehicle to obstacle using Leddar M-16 and RPLidar 360 is being done. Also these results are being compared to the improved accuracy of the resultant from both sensors after the data is fused together to produce a different set of values. Analysis on the data is done using a tool named Weka. Test bed and experiments were designed for collection of data. A machine learning technique, linear regression is used for improving accuracy of the measurement.