Adaptive Neuro-Fuzzy Technique for Autonomous Ground Vehicle Navigation

Auday Al-Mayyahi,Phil Birch,William Wang

doi:10.3390/robotics3040349

Auday Al-Mayyahi, Phil Birch + Show 1 more

Open Access

https://doi.org/10.3390/robotics3040349

Copy DOI

Journal: Robotics	Publication Date: Nov 19, 2014
Citations: 60	License type: CC BY 4.0

Affiliation: University of Sussex

Abstract

This article proposes an adaptive neuro-fuzzy inference system (ANFIS) for solving navigation problems of an autonomous ground vehicle (AGV). The system consists of four ANFIS controllers; two of which are used for regulating both the left and right angular velocities of the AGV in order to reach the target position; and other two ANFIS controllers are used for optimal heading adjustment in order to avoid obstacles. The two velocity controllers receive three sensor inputs: front distance (FD); right distance (RD) and left distance (LD) for the low-level motion control. Two heading controllers deploy the angle difference (AD) between the heading of AGV and the angle to the target to choose the optimal direction. The simulation experiments have been carried out under two different scenarios to investigate the feasibility of the proposed ANFIS technique. The simulation results have been presented using MATLAB software package; showing that ANFIS is capable of performing the navigation and path planning task safely and efficiently in a workspace populated with static obstacles.

Highlights

Autonomous ground vehicles are an important key of industrial automation
An artificial workspace has been created with seven identical static obstacles that were placed in different positions in the workspace
The workspace dimensions are fixed by four corner points having coordinates (–2, –2), (18, –2), (18, 18), (–2, 18) to combine the workspace grid

Summary

Introduction

Autonomous ground vehicles are an important key of industrial automation. Such vehicles could be utilized in different applications such as monitoring, transportation and many other potential applications. The path planning and navigation problems are one of area of current research. A considerable attention has been paid in recent years to deal with such problems. The autonomous vehicle must be able to gather and extract information from its surrounding using sensors. These sensors are fed to a controller to plan and execute its mission within its environment without human intervention [1]

Methods

Results

Conclusion