Introduction

Cloud computing technology has succeeded from the grid concept. Cloud computing systems are generally established in bigger data centers. Cloud data centers are hosted by third party. The hosting company provides the services to the customers or the end users. Before the cloud computing technology came into existence, the traditional business applications were existed. They are very complicated. They needed different types of hardware and software to be installed, configure, test, run, secure and update periodically by a group of experts [1].

Cloud computing is a distributed computing facility. Distributed computing facility is provided to the third party. John McCarthy is the father of this model of computing. Today cloud computing is a common and major computing resource. Cloud computing tremendously changed the technology aspects worldwide [2].

Cloud computing provides a way to rent the infrastructure, run the environment and to have the services on pay and use model. Cloud style of delivering the services is defined by ‘Reese’. As per Reese, cloud services are delivered with (i) service access via web, (ii) zero capital investment, (iii) pay per use model or metered services. Cloud computing is a newly developed model with a collection of resources. This cloud computing is a web-based advancement. Cloud computing provides the platform for the services and applications. Cloud computing provides a way to configure and adjust. It would be easy to turn off the services when there is no requirement. Cloud computing is a decentralized way of computing with location independence, device independence computational process [3].

The success of this computing technology depends on Web 2.0 technologies. Web 2.0 is created for the development and usage of the contents that are created by the end user. These contents are interoperable, which means the contents can be accessed on any systems, products, and devices. The contents are dynamic and intended for the general public from all corners of the world. A word ‘ubiquitous’ is used for the cloud which means ‘being present everywhere at the same time’. These contents are configurable and sharable.

There are three technologies which work behind cloud computing. They are,

  1. (i)

    Virtualization: It shares the resources among the users.

  2. (ii)

    Service Oriented Architecture (SOA): It provides a facility to use the applications as a service irrespective of vendors or technology [4].

  3. (iii)

    Utility computing: This service model provides computing resources to the end user and the user can manage his part of the infrastructure [5] (Fig. 1).

    Fig. 1
    figure 1

    Cloud computing technology has the un-limited access to many devices

    Full size image

Cloud Technology Working Models

Cloud computing has two types of models: (i) deployment model and (ii) service model

(i) Deployment model: It refers to the way cloud computing is made available to the end user. Generally, cloud computing deployment model is a control-based infrastructure mode provisioning the services to the end user. There are four types of deployment models. They are:

(a) Public deployment model: General public could access in this model. It can be made available at a lower cost, e.g., Google, Amazon, Microsoft. These services are available with the help of Internet. Advantages of this model are: it will have less cost, and it is reliable, flexible, location independent and highly scalable. It has disadvantages also such as it has less security, less privacy, and less control [6].

(b) Private deployment: This deployment model is available for fewer people such as private use. It takes more cost, and good security can be provided. This deployment model is intended within an organization. Advantages of this model are high level privacy and more control. It has disadvantages also such as restricted area, inflexible, and limited scalability. Example: Any Bank cloud network which does not join the public cloud or a Web site which is hosted with any third-party hosting provider.

(c) Hybrid deployment model: Amalgamation of public and private. Cost for this establishment is medium, and the security it provides is moderate. Advantages of this model are scalable, flexible, cost efficient and secure. It has disadvantages also such as network issues, security compliance, and infrastructural dependency, e.g., Cloud stack, Open stack, Eucalyptus [7].

(d) Community deployment model: The services of this community model are available to a community only. Establishment cost is medium. This is a secured model. An advantage of this model is less cost-effective and more secure. It has disadvantages also such as data are accessible to others. Its management is very hard. The real-time examples of community models are government services such as healthcare schemes.

(ii) Service model: It refers to the way the cloud computing provides its service to the end user. Generally, cloud computing service model is a facility made available to the end user on demand via Internet from a cloud provider’s server. There are three types of service models. They are

(a) SaaS (Software as a Service): This would offer the software for the end user as a service. SaaS makes the software available over the Internet. It is controlled by the cloud services provider or the one who sells it. Licenses can be subscribed to software, based on the requirements. Its applications are cost-effective and available on demand. SaaS applications are automatically upgraded, e.g., Salesforce.com, Billing systems, Invoicing systems, Customer relationship applications, Human resource solutions. Example: Salesforce.com, Microsoft 365, Google Apps.

(b) PaaS (Platform as a Service): It models a platform as a service. It offers browser-based development environment, provides built in security web services interfaces, and is very simple in order to interact with other application programs within the identical cloud computing technology platform, and it would also provide web technological services to interface to connect other applications which are on other platforms, e.g., Google App Engine. Generally, there will be less administration control and these are not portable.

(c) IaaS (Infrastructure as a Service): It provides access to basic computing resources for example computers, virtual machines, load balancers, IP addresses, and software bundles, e.g., Amazon web services. The advantages of this service model are full control over the resources, flexible, efficient renting of hardware, portable, and interoperable with legacy applications.

Example: Cisco Meta pod, Microsoft Azure, Google Compute Engine (Fig. 2)

Fig. 2
figure 2

Flow of data from end user to the cloud

Full size image

Robotic Technology

The word Robot was first used by Josef Capek, derived from Slavanic rabota, meaning of a slave. Joseph F. Engelberger (1925–2015) is considered as the father of Robotics [8]. As he describes a robot as “An automated machine that does just one thing is not a robot. It is simply automation. A robot should have the capability of handling a range of jobs in a factory” (Fig. 3).

Fig. 3
figure 3

Robotic technology icon 

Full size image

A robot can be defined as a machine which would do some useful works as given by the user which would reduce his/her work and make smooth transactions. These robots can be defined into two varieties: first for industrial usage and second for personal usage. However, all industrial robots are used for commercial usage and personal robots are not used for any commercial usage [9]. Industrial robot is defined to be a “controllable device without any barrier, and they could be programmed again and again for many number of requirements, which needs to use either for a static location or for movable condition.”

0th Laws of Robotic Technology

Robotics are possible to use for variety of functional needs and could be programmed again. These robotics would have electrical elements. These elements would provide the electrical power, which would make the device to function and control. This robotics will also have some physical construction which will have some physical structure. To guide and achieve some task, it needs some software program [10]. As per Isaac Asimov, “Laws of Robotics” states that “A robot may not injure humanity, or, through inaction, allow humanity to come to harm”(0th Law).

Characteristics of Robotics

The following are the qualities of Robotics:

  1. 1.

    Body: Robots will have body parts [11].

  2. 2.

    Controlling unit: It will have a controlling unit.

  3. 3.

    Sensors: Robots will have sensors.

  4. 4.

    Actuators: Robots can move.

  5. 5.

    Program: Robots will work based on some program (or) programs

  6. 6.

    Conduct: Robots will show conduct that is coordinated by its program.

Need of Cloud Robotics

Cloud apply autonomy is a developing field of mechanical technology established in distributed computing, distributed storage, and other Internet advancements revolved around the advantages of united foundation and shared administrations [12]. It permits robots to profit by the ground-breaking computational, stockpiling, and correspondences assets of present-day server farms. Likewise, it expels overheads for upkeep and refreshes, and diminishes reliance on custom middleware. Cloud mechanical autonomy permits robots to exploit the fast increment in information move rates to offload undertakings without hard continuous prerequisites. This is quite compelling for portable robots, where on-board calculation involves extra force necessities which may lessen working term and oblige robot versatility just as increment costs. More extensive meanings of Cloud Robotics may likewise incorporate other Internet-related parts of apply autonomy, e.g., patterns toward the web-based sharing of open source equipment and programming, publicly supporting of apply autonomy subsidizing, telepresence, and human-based calculation. Different definitions stress the connections among mechanical autonomy and related rising fields, e.g., the Internet of Things, Web of Things, robot application stores, sensor systems, Big information, and others.

Initial Steps in Cloud Robotics

The DAvinCi Project has illustrated the benefits of cloud computing by parallelizing a SLAM algorithm utilizing a Hadoop cluster, shockingly not openly accessible [13].

The qualities of cloud apply autonomy are as follows:

  1. 1.

    On-request self-administration,

  2. 2.

    Expansive network system gets to,

  3. 3.

    Useful resource pooling,

  4. 4.

    Fastest flexibility,

  5. 5.

    Estimated administration (Fig. 4).

    Fig. 4
    figure 4

    Cloud robotics a pictorial representation

    Full size image

Cloud robots can be used for many applications such as:

  1. 1.

    The usage of virtualization concept on physical robots.

  2. 2.

    Web-based resources accessibility to the robotics

  3. 3.

    Noble idea to decrease the expenditure of hardware of robotics.

  4. 4.

    Creating less priced and smart robots for the usage of common public.

  5. 5.

    Usage of latest modern utilities like object recognition and voice services whenever needed.

  6. 6.

    Need for the “Master Knowledge Storage” for the benefit of easy running of cloud robots.

Cloud Robotics Technologies

A few advancements are being utilized including the Lego NXT robot, the Lego Java Operating System (leJOS), a Python mediator, the Raspberry Pi, Phidgets, and the Seattle distributed computing stage.

1. Lego NXT Robot: The Lego NXT robot contains a programmable block with a lot of sensors and servo engines. The Lego NXT robot accompanies a 32-piece ARM principle processor, 256 kilobytes of blaze memory for program stockpiling, and 64 kilobytes of RAM for information stockpiling during program execution. The Lego NXT robot contains a programmable block, servo engines, light sensor, ultrasonic sensor, sound sensor, and contact sensors [14]. The NXT block contains three yield ports that are utilized to control the servo engines. It likewise has four info ports that are utilized to associate the sensors. To get information from the sensors, an additional processor is incorporated that has 4 kilobytes of blaze memory and 512 bytes of RAM. For availability, the Lego NXT robot underpins the Bluetooth convention for remote, and USB for wired.

2. Lego Java Operating System (leJOS): It is an open source firmware swap for the Lego NXT robot. This OS gives an execution of a Java Virtual Machine appropriate for the Lego Robot processor. leJOS bolsters have an option runtime condition that permits programming the robots utilizing the mainstream programming language Java. leJOS incorporates a library of Java classes that actualize the leJOS NXJ Application Programming Interface (API). This API comprises a library of classes for work area programming that speaks with the block by means of Bluetooth or USB. It incorporates too a lot of apparatuses for troubleshooting, firmware substitution, ordering, and transferring programs into the robot’s block. A portion of the highlights upheld by leJOS include: object situated programming support, recursion, strings and synchronization, Java fundamental sorts clusters, special case taking care of, and a very much archived mechanical autonomy API.

3. Python: Python is a significant level deciphered programming language that underpins various programming ideal models and spotlights on lucidness. Right now utilized a few devices to permit execution of Python code inside the Lego NXT robot. These instruments are Jython, PyMite, and NXT-Python. Jython is a Python translator written in Java. PyMite is a lightweight Python translator particularly intended to chip away at implanted gadgets. NXT-Python is a Python library that gives orders to control the Lego NXT block utilizing Bluetooth or USB. Executing Python code is significant for utilizing different instruments including the Seattle distributed computing stage and speaking with the Raspberry Pi.

4. Raspberry Pi: The Raspberry Pi is a charge card estimated single-board PC. The most recent models are known as Model B and Model B + . They think of 512 MB of RAM, an improvement over the first Model A that has 256 MB of RAM. For capacity, the Raspberry Pi bolsters SD or MicroSD cards. The working arrangement of decision is Linux.

There are numerous dispersions of Debian and Arch Linux ARM accessible for download, being Raspian the suggested OS. The Raspberry Pi has numerous devices and programming dialects accessible, e.g., Python, C, C++, Java, Perl and Ruby. This at a little cost, somewhere in the range of $30 and $40 relying upon the model at the hour of this composition.

5. Phidgets: A phidget is made out of two words: physical and gadget. It is best portrayed as a physical usage of a GUI gadget (for example an on-screen dial gadget actualized truly as a handle). Phidgets depend on the utilization of USB ports to give minimal effort sensors and actuators. They were planned essentially for exploring different avenues regarding elective physical PC connection frameworks. Anyway on account of their adaptability and usability, they have been broadly utilized in fields like apply autonomy and savvy homes. Their name speaks to the endeavor to manufacture a physical simple to the product of the gadgets we find in GUI frameworks. Along these lines, complex physical frameworks can be built by consolidating a lot of more straightforward parts. These phidgets associate with a PC which controls them by means of the Phidgets Application Programming Interface (API). This API incredibly improves the unpredictability of taking care of these physical gadgets with the goal that designers can concentrate on the execution of uses. Phidgets applications can be produced for various working frameworks including the most famous forms of Windows, Linux, and Mac OS. There are likewise numerous programming dialects upheld including C, C++, C#, Java, and Python.

Constraints on Cloud Robotics

Resource Constraints

Even though a robot can impart its calculation remaining task at hand to different units in the system, the general adequacy of the automated system is constrained by every robot’s assets, including locally available PCs or implanted processing units, recollections, and extra room. Truly, these installed figuring gadgets are confined by the robots’ size, shape, power supply, movement mode, and workplace. When the robots are planned, assembled and sent, it is actually testing to change or overhaul their asset arrangements [15].

Communication Constraints

Normal conventions for machine-to-machine (M2M) interchanges incorporate proactive steering, which includes the intermittent trade of messages with the goal that courses to each conceivable goal in the system is kept up, and advertisement hoc routing, which frames a unique course to a goal hub just when there is a message that should be sent. Proactive steering acquires high calculation and memory assets in the course disclosure and support process. Specially appointed steering conventions experience the ill effects of high idleness as a course must be built up before a message can be sent, and are not functional if the system topology is exceptionally unique. These downsides are critical in versatile automated systems, and may prompt extreme execution corruption.

Cloud Robotics Architecture

Robots in a system can impart on the off chance that they are inside correspondence scope of one another, and with the cloud servers if the robots are near passageways of the cloud foundation. A remote M2M correspondence system can be framed by robots working agreeably with one another to course and transfer data. We call either a robot or a passageway a hub in the M2M organize. A few principles like Zigbee, Bluetooth, and WiFi Direct have been created for short-range remote interchanges between robots. For long-range correspondences, radio recurrence and microwave correspondence innovations might be utilized. A system of robots is frequently shaped powerfully and in a specially appointed way. There is no focal controller to organize the correspondence stream in the system. Mechanical units may leave and join the system, or may get inaccessible in view of flighty disappointments or impediments in nature. Moreover, the system is exceptionally powerful if robots are versatile. Every one of these contemplations make the plan of compelling steering conventions troublesome and unfeasible in certain situations. Tattle calculations are randomized techniques intended to transmit a message from a source to a goal with no unequivocal course revelation component.

In the event that two hubs are inside correspondence go, we state that they are neighbors. At the point when a robot needs to make an impression on a goal hub (either another mechanical unit or a cloud passage), it arbitrarily picks one of its neighbors and transmits the message together with a header that contains the identifier of the goal and itself, and a period esteem demonstrating the legitimacy time of the message. In another variation of the convention, the message is essentially communicated to all neighbors, however relying upon the application, (i) Peer-Based Model, (ii) (b) Proxy-Based Model, (iii) (c) Clone-Based Model

This may bring about high correspondence load in the system. We will concentrate on the convention that picks an irregular neighbor right now. At each time step, every hub haphazardly picks a neighbor to retransmit messages that are not proposed for itself are as yet legitimate. After an adequate number of time steps, all the messages will be transferred to their goals with high likelihood [16].

Addressing needs of Cloud Robotics

Cloud Robotics, the accompanying four key perspectives must be tended to:

  1. 1.

    Wide accessibility and honesty of information and information: Domain information (comprising of adages, inferred information, program code, and ontologies) and information must be straightforwardly accessible and straightforwardly obvious.

  2. 2.

    Trustability in information and calculation: Availability of data all alone isn’t adequate. Robots should have the option to trust; that calculations performed by their friends have been finished effectively. At exactly that point data can frame a solid wellspring of contribution for additional determined activities.

  3. 3.

    Re-use and interoperability across applications: Programs must be particular and semantically interoperable to encourage re-use across guidelines.

  4. 4.

    General executability of programming programs: Deterministic projects must be all around executable paying little heed to the hidden equipment, while creating repeatable outcomes. In any case, redistributing calculation and sharing the individual outcomes is outlandish [17].

Cloud Robotics Algorithms

Assessing area of Landmarks and area of portable robot all the while during map building is called SLAM. Simultaneous confinement and mapping, issue which has exceptional impact on versatile robot route and that is actually a difficult issue since it needs to assemble a guide and at a similar minute it needs to figure possess position with this guide. Answer for this issue empowers versatile robots in an obscure domain, to manufacture a reliable guide. There is a wide utilization of SLAM, similar to space portable robot application, submerged application, and open air and indoor robot application.

Khairuddin et al. [19] have described about the simultaneous localization and mapping. He explained about this technique how it is applied in the artificial intelligence. He explained how SLAM has become basic R&D filed in our latest situations presently. He says it is a solution to solve many our today’s issues related to artificial intelligence. Khairuddin explained how human intervention is not needed in terms of this AI. Simultaneously mapping and localization is possible in SLAM. SLAM algorithms are the way to proactive in real which has efficient impact on the resolving the needs of the humans.

Dai et al. [20] say that Itpsilas preferred that a hierarchical map is considered for the adoption for the needs of larger environments. The hierarchical simultaneous localization and mapping along with automaton map in the upper layer, feature-based maps also in the lower layer. These both macro- and micro-localizations are needed to be realized. The mapping progress based on supervisory control of discrete event systems is proposed by Dai. Itpsilas has embodied this particular system and also implemented as a supervisor which has permitted some sort of desired activities which would occur and it has prohibited unwanted conditions.

Wang et al. [21] related to the issues of localization, and detection procedure is general type of navigation with the non-Gauss and nonlinear model in the water below normal level, the feature map generation algorithm is based on parallel localization and mapping is proposed. The image processing and the feature extraction-based procedure based on the SLAM is demonstrated. Features of the shipwreck were matched, and it is extracted, respectively.

Norazman et al. [22] have presented the object scanning for the landmarks to be detected, and it is based on the multiple scanning of the non-stationary and discrete sensor. To make a preservation of a whole mapping and localization procedures, the massive, cost-effective and also complexed computational aspects are involved in this. The used sensors are sometimes expensive to get the services with the accurate landmarks. Norazman has proposed the mapping technique based on the Simultaneous Localization and Mapping in the unknown environment for the autonomous vehicle platform is designed. It is useful for the unmanned vehicle for finding the eligible pathways during the navigation. In this study, it has found that the infrared sensor is estimated to be more suitable for defining the shape of the detected object.

Formulation of SLAM

SLAM issue is considered as a two-stage issue: first is localization stage, and other is mapping stage

i. Localization: in which robot limit possess position on the guide which it manufacture all the while

ii. Mapping: right now map condition by utilizing laser information and possess position.

There are number of structures recommended by different analysts. In light of writings of DAvinCi, Rapyuta and C2TAM systems have wide consideration in the apply autonomy inquire about individuals. In route, mechanical mapping and odometry for computer generated reality or expanded reality, concurrent confinement and mapping (SLAM) is the computational issue of building or refreshing a guide of an obscure situation while at the same time monitoring a specialist’s area inside it. While this at first has all the earmarks of being a chicken-and-egg issue, there are a few calculations known for explaining it, in any event roughly, in tractable time for specific situations. Well-known inexact arrangement techniques incorporate the molecule channel, broadened Kalman channel, covariance convergence, and Graph SLAM. Hammer calculations are custom fitted to the accessible assets, subsequently not focused on flawlessness, yet at operational consistence. Distributed methodologies are utilized in self-driving autos, unmanned ethereal vehicles, self-governing submerged vehicles, planetary meanderers, fresher household robots and even inside the human body.

Conclusions and Future Work

Cloud is a utility-based technology with the participation of civilians for the improvement of their needs. Cloud is web-based utility figuring which offers a decision to lease and utilize the computational assets on membership premise. It has the service arranged and deployment models which would offer reasonable types of assistance to the end clients. Apply autonomy can be characterized as a machine which would do some valuable fills in as given by the client which would lessen his/her work and make smooth exchanges. Robots can be used for fulfilling the needs for the people. Cloud Robotics is an up and coming innovative territory with a mix of distributed computing and robotics. Inter-association of these two fields would benefit the humanity. A lot of progressively various assortments of uses are developing as on today. In happening to the days, cloud robots will have significant job to perform, for example, to improve the ‘open utility of innovation’.