Optimized Drivable Path Detection System for Autonomous Vehicle in Rain Condition

Olusanya Y. Agunbiade, Selemon M. Ngwira, and Tranos Zuva


Autonomous vehicle, Vision-based system, Filtering algorithm, Environmental noise


Drivable path detection is a vital researching issue that is increasingly receiving attention from scholars because autonomous vehicle can successfully navigate, if proper drivable paths are detected. Several vision-based techniques have been proposed for drivable path detection and amazing results have been achieved by many researchers. Rain characterized as environmental noise is capable of causing autonomous vehicle accidents because of inaccurate detection of drivable path. However, after investigating the effect of rain, we address the issue of drivable path detection during rain by introducing a filtering algorithm into the drivable path detection system. The filtering algorithm used to optimize the system is capable of eliminating the effect of rain. The optimized system experimental comparison was done qualitatively and quantitatively using the following evaluation scheme: Precision (PRE), False Positive Rate (FPR), Accuracy Rate (ACC), Error Rate (ERR), Total Positive Rate (TPR), False Negative Rate (FNR), and Total Negative Rate (TNR). Results prove the system improvement in drivable path detection during rain scenario and this progress has further enhanced detection and autonomous vehicle navigation.

Important Links:

Go Back