Abstract:
This thesis reports a cloud robotics framework composed of mobile phone based robots and a cloud computing architecture, facilitating complex and heavily computational tasks to aid the user in the completion of difficult robotics related problems. The Cellbot based framework will complement the group of networks that the user may already be a part of, contributing as a connection bridge between real and virtual worlds. The objective of this endeavor marks the implementation of a small test/demo cloud based robotics service framework that permits distant groups of cellbots to share learned skills while simultaneously improving collaboration with non-technical human users. This cloud facility allows for robots to share new and used knowledge by downloading and uploading information as it arrives, without needing onboard storage, thus risking information freshness.
In this scenario the context aware skill arranges the system to give services such as maze traversal and path following. The small movement Cellbot used in the experiments is able to navigate its way through both the scenarios without relative difficulty after receiving commands from the Cloud server with negligible amounts of delay times. The needed input for the system to make the Cellbot move was entirely done using a single 2D camera of the Android smartphone placed on the Cellbot, whereas all storage and processing was done on the Cloud.
