Rain classification for autonomous vehicle navigation : A support vector machine approach

• Main Authors: Abdul Haleem, Habeeb Mohamed, Muhammad Aizzat, Zakaria, Mohd Azraai, Mohd Razman, Anwar P. P., Abdul Majeed, Mohamad Heerwan, Peeie
• Format: Article
• Language: English
• Published: Penerbit UMP 2020
• Subjects: TJ Mechanical engineering and machinery; TS Manufactures
• Online Access: http://umpir.ump.edu.my/id/eprint/33644/1/Rain%20classification%20for%20autonomous%20vehicle%20navigation.pdf; http://umpir.ump.edu.my/id/eprint/33644/; https://doi.org/10.15282/mekatronika.v2i2.7022; https://doi.org/10.15282/mekatronika.v2i2.7022

The advancement of LIDAR technology used in the autonomous vehicle (AV) system has made it increasingly popular. Despite that, the ability of the sensor to adjust to human behaviour in sensing and perceiving different environments is still unsolved as it significantly impacting the performance of LIDAR, causing the effect of missing points and false positives detection. The immerging of machine learning algorithms that have greatly impacted solving uncertainties and LIDAR's reliability in making judgments has proven a great success. This paper aims to classify different rain rates conditions in a controlled environment with real rain using a LIDAR. Then, the feature extraction using the time-domain method was employed to generate more features with a variation of SVM models in developing classification models. The preliminary observation shows that the Poly-SVM model can achieve a test classification accuracy of 97%. Noting that, the proposed method has the potential to evaluate weather classification.