الفهرس | Only 14 pages are availabe for public view |
Abstract Cloud computing gives a dynamically scalable computing power, storage, and other re- sources depending on the user’s demands. Accordingly, it is feasible to exploit cloud resources to the offload tasks that are too heavy for mobile robots hardware capabilities to reach in real time. However, existing cloud computing cannot efficiently deal with the soft real-time applications such as online gaming, video streaming, and cloud-based vision system for mobile robots. Such applications require real-time performance and shared computing environments. The main objective of this thesis is to propose a human cloud mobile robot architecture which enables a single user to control multiple mobile robots. In addition, this architecture allows mobile robot vision system to reliably achieve real-time constraints using cloud computing through 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 mod- ified growing neural gas algorithm, applied on the cloud server side. The experimental results demonstrate that there is enhancement in the total response time, depending on the communication bandwidth and image resolution over other state-of-the-art techniques. Moreover, better performance in terms of data size, path planning time, battery life, and accuracy is demonstrated. |