In this paper enhancements have been done to the previously proposed Support Vector Machines (SVM) based path planning algorithm [1]. SVM, a data classification technique has been applied in conjunction with k-Nearest Neighbors (k-NN) algorithm for autonomous navigation in unknown road-like environments. The features of the road such as lane markers and the obstacles in the robot’s visibility are divided into two classes using k-nearest neighbors (k-NN) algorithm and then a maximum margin hyperplane is obtained using SVM that optimally separates both the classes. This hyperplane represents a collision-free path. The proposed algorithm has been tested in a variety of environments and the scenarios where it was unsuccessful have been identified and addressed for improvement. The simulation results of the enhanced algorithm have been presented and its performance has been compared with vfh+, a purely local obstacle avoidance algorithm based on artificial potential field method.

This content is only available via PDF.
You do not currently have access to this content.