Page 94 - ITU Journal Future and evolving technologies Volume 2 (2021), Issue 5 – Internet of Everything

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 5

This work will focus on the interfaces between users and In industry and in particular in the IoT ield, chatbots
IoT‑enhanced environments. Indeed, the chatbot will are entering in work lows in a capillary way. This is
provide users with a user‑friendly interface to assist them because thanks to their characteristics they allow to
in creating their pro ile and managing their services and stem the dif iculties of con iguring and troubleshooting
objects. The virtualization of the user is also introduced to devices encountered by operators and users.
suggest relevant services that are expected to be the most The irst problem that arises when new solutions have to
suitable and interesting for each user based on their pro‑ be introduced into existing systems is to understand the
ile and thanks to context‑awareness mechanisms. The impact in terms of complexity. To understand the com‑
use of chatbots in the IoT already has a history, but the plexity in [5] the authors analyzed the possibility of cre‑
main limitation comes from the fact that these are vertical ating a general architecture that would allow the integra‑
and domain‑speci ic solutions. The solution we propose, tion between chatbot and IoT systems in a simple way.
thanks also to the intermediation of user/device virtual‑ The study found that what chatbots and IoT have in com‑
izations, allows the reuse of the same chatbot interface on mon is that they adopt their services through relatively
different IoT applications. simple, often RESTful, web APIs. In this scenario, adopt‑
Accordingly, the contributions of the paper are as follows: ing a service‑oriented development approach to devel‑
opment, integration is feasible thanks to RESTful HTTP
• We analyze the key requirements for the develop‑
standards and protocols. In this case the ISO/OSI appli‑
ment of a chatbot for the IoT scenario and provide
cation level is the only level concerned, without having to
a description of the architectural components of the
go down to the underlying levels. It is therefore clear that
chatbot‑enabled user virtualization; with design precautions, the integration between chatbot
and IoT platforms is extremely simple.
• We discuss the integration of the chatbot system in
a fully distributed virtualization‑based IoT architec‑ In the literature there are several examples of systems
ture; that use chatbots to interact with IoT devices.
In [6] the authors implement a chatbot integrated with an
• We provide the details of the implementation that agricultural plant monitoring system. In their implemen‑
have been carried out to develop a prototype; tation they use fuzzy logic and Natural Language Process‑
ing (NLP) to interpret user inputs. The user asks the plant
• We present some experimental results for the eval‑ a question and it answers. An orchid was used for the ex‑
uation of the capability of the proposed solution to periments. The success rate of the interaction between
identify correctly the user intention interacting in question and answer was 71%.
the IoT environment. An interesting proposal is presented in [7], an IoT system
with AI chatbots for plant monitoring capable of monitor‑
The paper is structured as follows. Section 2 discusses
ing various parameters useful for knowing the health of
the major related works in this area. Section 3 presents
houseplants. Alongside the IoT system, we implement a
the key requirements in designing a chatbot system in a
chatbot to inform the owner about the current conditions
virtualization‑based IoT platform. The system archite‑
of the plant and its current needs. The data is also stored
cure is presented in Section 4, while in Section 5 we illus‑
and through the bot the user is able to analyze the graph
trate a use case to better understand the reference sce‑
and determine the level of wellbeing of the plant and any
nario. The implementation and the experimental results
problems.
are shown in Section 6 and Section 7 respectively. Finally,
Section 8 concludes the paper. In [8] an integrated Chatbot‑IoT system is implemented
to make the monitoring and improvement of water qual‑
2. ST ATE OF ART ity quick and ef icient. For monitoring, a network of IoT
sensors was created, supported by a cloud platform. In‑
The irst entertainment chatbot was developed in 1966, side, a chatbot has been integrated that uses text mining
it was called ELIZA [3] and was a parody of a psychother‑ techniques to interpret user inputs. The result showed
apist who answered the patient’s questions with other excellent performance with high precision and recall for
questions, obtained by rephrasing the patient’s questions. each class.
In 1995, Richard S. Wallace built A.L.I.C.E. [4] a chatbot In [9] and in [10] two IoT platforms for home monitor‑
made entirely with open source software that uses the ing and remote control are presented. They have a built‑
AIML language, child of the XML language from which it in chatbot that can understand text or voice commands
inherits extensibility, which thus allows the chatbot to using NLP. Using different APIs and protocols, the au‑
hold a conversation. With the growing interest in arti i‑ thors have obtained user‑friendly systems for controlling
cial intelligence and with the idea of simplifying the inter‑ home devices. They also demonstrated how an architec‑
action between man and machine, more and more com- ture structured on multiple services is effective and easy
panies, have developed or directed part of their research to implement.
on chatbots.

82 © International Telecommunication Union, 2021