۱- Introduction

۲- Literature review

۳- Problem statement and solution plan

۴- Proposed human-cloud-mobile-robot architecture

۵- Data flow mechanism

۶- Experimental result discussion and analysis

۷- Conclusion and future work

References

Abstract

Mobile robots require real-time performance, high computation power, and a shared computing environment. Although cloud computing offers computation power, it may adversely affect real-time performance owing to network lag. The main objective of this study is to allow a mobile robot vision system to reliably achieve real-time constraints using cloud computing. A human cloud mobile robot architecture is proposed as well as a data flow mechanism organized on both the mobile robot and the cloud server sides. Two algorithms are proposed: (i) A real-time image clustering algorithm, applied on the mobile robot side, and (ii) A modified growing neural gas algorithm, applied on the cloud server side. The experimental results demonstrate that there is a 25% to 45% enhancement in the total response time, depending on the communication bandwidth and image resolution. Moreover, better performance in terms of data size, path planning time, and accuracy is demonstrated over other state-of-the-art techniques.

Introduction

Mobile robots affect everyday life, as they can replace humans in several activities, such as material handling, building construction and demolition, or even planting and harvesting. Mobile robots are divided into two main categories: (i) Teleoperated mobile robots and (ii) Autonomous mobile robots. Teleoperated robots can be controlled through a wireless or wired communication system, e.g., Internet, radio connection, direct cable, or satellite. Autonomous robots can perform tasks such as house cleaning, planting and harvesting, searching and rescuing, and constructing and demolishing with minimal user intervention. Therefore, to fully use their capabilities, mobile robots should be controlled with the highest degree of autonomy. To this end, various architectures have been investigated. Robotic vision has played a major role in making mobile robot navigation safer, but its computational complexity limits autonomy. One solution is to provide the robot with high computational power; however, this requires more hardware resources, which may increase size and cost, as well as larger capacity batteries. The term “Cloud Robotics” was introduced by Kuffner [1] to support “Remotely Brained” robots with parallel computation and big data sharing over the internet.