Cloud Robotics Design Implementation And Evaluation Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

This report illustrates the design, implementation and evaluation of cloud computing technology for robotics. Cloud Robotics is the application of the concept of cloud computing to the robots by delivering hardware and software computing resources as a service over the network. The key idea is to utilize the benefits of cloud technology by harnessing its processing power, memory & storage and thereby allowing the creation of smaller, cheaper, smarter and energy efficient robots. Cloud technology used in robotics can process large amounts of data, share knowledge among multiple robots or users and create new forms of user interactions. This report explains how cloud robotics can enable a smarter means of accessing, monitoring and controlling of robots.


Over the past few decades, robotics has brought significant socio-economic reforms. The field of robotics is experiencing a rapid growth with the technological advancements. The service robotics is forecasted to become a US$12 billion industry by year 2015[1]. Complex, time-consuming, dangerous tasks were possible to complete quite easily with the help of robots. The industrial use of robots has helped achieving higher production rates. However the primary cost in installation, use & calibration of robots was very high. Major portion of this cost has to be spent on onboard computation. Cloud robotics has provided a solution to this problem.

Cloud computing is the use of hardware and software computing resources that are delivered as a service over network. The use of cloud technology in industry is increasing rapidly due to many benefits it has to offer. The NIST(National Institute of Standards and Technology) defines Cloud computing as a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction[2]. According to the NIST definition, the essential characteristics of cloud computing technology include resource pooling, broad network access, on-demand self service, rapid elasticity (scalability), agility and measured service. In simple words, it helps leveraging resources, providing faster performance, unlimited storage and cost-effective.

The above mentioned benefits of cloud computing can be used to revolutionize the design of robots entirely. Instead of relying on the on-board computing for robots, we can harness that computing power from cloud. More complex robots required memory elements to store the data on-board that can be stored and extracted from cloud whenever needed through high bandwidth internet access. The cloud computing power is scalable that means for more complex tasks, more computational power can be drawn from the cloud. For the regular tasks the usual cloud services can be used. Since the cloud services are available pay-per-use, the total running cost is very less as no primary investment is required in setting up infrastructure (on-board computer, memory). The upgrade of robot could be possible real-time. This can be achieved by passing the actuator parameters for different environments around the robot. These and many other endless possibilities await as we introduce the concept of cloud computing to the robotics.

Design and Implementation

The benefits of using cloud computing technology in robotics can be illustrated by summarizing various researches in the field of cloud robotics. These researches include various paper presentations and some eminent research projects.

Roboearth Project[3]:

"RoboEarth" is an extensive network and database repository for robots. It was initiated in late 2009. The motive was to build a network where robots could share information, learn from the experience of other robots thereby allowing rapid advances in machine cognition and behavior. The ultimately goal was to create a more subtle and sophisticated human-machine interaction.RoboEarth-Diagram-small1.png

Under this project a cloud robotics infrastructure was offered. This infrastructure has a database repository to store knowledge generated by humans and robots in a machine-readable format. The data stored can include software components, maps for navigation, task knowledge and object recognition models. It allows robots to offload their heavy computation to secure and powerful cloud computing environments called the 'RoboEarth Cloud Engine' with minimal configuration. This Cloud Engine also provides high bandwidth access to the RoboEarth knowledge repository for the robots even with different hardware and software and enables them to collectively improve their knowledge.

Professor James Kuffner of Carnegie Mellon University and Engineering Manager at Google presented a paper on "Cloud-enabled humanoid robots" at the '10th IEEE-RAS International Conference on Humanoid Robots' [4]. In his paper he mentions about Physical separation of Hardware (motors & sensors) and Software (high-level processes) by introducing a concept called 'remote brain'. According to him, cloud computing technology can make the robots 'cheaper, lighter & smarter'. This is achieved by doing following three major things - sharing knowledge Base, Offloading heavy computing tasks to the cloud, setting up a skill/behavior database. Additionally, the cloud computing enables the robots to use services over the internet such as Google Maps, Google Object Recognition (Google Goggles), speech recognition and translation, also an application-specific set of instructions and device parameters for robots. This opens the doors to many possibilities for expansion of the field of robotics.

In 2011 Google announced the formation of a "Cloud Robotics team" at the Google I/O developer conference [5]. The team, in a joint talk with Willow Garage Inc., introduced the first pure java implementation of Willow Garage's Robot Operating System (ROS). This software is intended to allow integration of Google's Android operating system with ROS. The main idea of their conference was on leveraging existing Google Web Services.

In late 2011, RobotShop distribution Inc. announced its Cloud Robotics platform, '', terming it as 'the facebook for robots' [6]. The idea behind creating this platform was to remotely monitor and control the robots. It's simple to use interface and integration set up a cloud robotics base that was easier to learn and accessible for almost everyone.

Another project with a revolutionary concept was "DAvinCi: a cloud computing framework for service robots", submitted in the 'IEEE International conference on Robotics & Automation (ICRA), 2010' [7]. This is a software framework that provides the scalability and parallelism advantages of cloud computing for service robots in large environments. The primary goal was to facilitate the easy maneuver of the robots in large environments by creating intelligent maps that was achieved by developing a cloud computing environment with commodity hardware exposing a suite of robotic algorithms as a SaaS and share data co-operatively across the robotic ecosystem. It has been implemented around the Hadoop cluster[8] as parallel processing system with ROS (Robotic Operating system) as the communication framework for the robotic ecosystem. (Hadoop is an open-source software framework that supports data-intensive distributed applications). A global map of a large area was formed using a very small eight-node Hadoop cluster with the help of FastSLAM algorithm [9]. This global map was later shared with other robots introduced in the environment via a Software as a Service (SaaS) Model of cloud computing. This reduced the burden of exploration and map building for the new robot and minimized the need for additional sensors.