The rate at which the internet is growing is unstoppable due to the large number of connected smart devices. Manufacturers often develop specific protocols for their own devices that do not usually follow any standards. This hinders the interconnection and coordination of devices from different manufacturers, limiting the number of daily activities that can be supported. Some works are proposing different techniques to reduce this barrier and avoid the vendor lock-in issue. Nevertheless, this interconnection should also depend on the context. In this chapter, the authors propose a system to dynamically identify the interconnections required each specific situation depending on the context. This proposal has been tested in case studies focused on elderly people with the aim of automating their daily tasks and improving their quality of life. Further, in a world with an accelerated population aging, there is an increasing interest on developing solutions for the elderly living assistance through IoT systems.
TopIntroduction
The Internet of Things (IoT) is becoming more and more important. This is due to the large amount of smart devices that are being developed. Recent studies estimate that by 2020 there will be 20-30 billion devices connected to the Internet (Nordrum, 2016; IronPaper, 2015). These devices can be applied for a multitude of fields such as smart cities, agriculture, automotive or healthcare (Haluza & Jungwirth, 2015). The purpose of these devices is to make people’s lives easier by automating tasks or helping users to perform them. In healthcare, the IoT paradigm allows more personalized, collaborative and preventive care, where patients are able to monitor and manage their own health. Besides, the responsibility for healthcare is shared between patients and medical staff (Metcalf, Milliard, Gomez, & Schwartz, 2016). These solutions are particularly interesting, mainly due to limited resources or difficult access to them if we consider the ageing of the population and the depopulation of rural areas.
IoT devices can perform simple tasks such as monitoring blood pressure or glucose level, performing periodic reminders, enhancing drug management or notifying of certain events. These tasks are useful but the real potential of IoT devices lies in the interconnection between them to collaborate and to perform more complex tasks. Due to the great heterogeneity of devices on the market, where there are many different manufacturers and devices, this interconnection is not easy. In order to achieve this interconnection, each manufacturer usually develops its own communication protocols. This means that there is no defined communication standard and the risk of vendor lock-in increases. This phenomenon conditions users to acquire devices from the same manufacturers in order to achieve total compatibility, also complicating the possible migration to another in the future (Roman, Zhou, & Lopez, 2013), and even to set manually each device, something that can be tedious for people who do not have a certain technical knowledge.
Different works have promoted alternative methods to make IoT devices work with each other, such as specific frameworks, such as (Shrestha, Kubler, & Främling, 2014), where a framework is developed to integrate specific domain applications into IoT, or (Kim et al., 2016), which presents interfaces and interconnection procedures based on oneM2M (Swetina, Lu, Jacobs, Ennesser, & Song, 2014). The use of ontologies and the Semantic Web are also becoming very important to solve these interconnection problems (Szilagyi & Wira, 2016). The main objective of the Semantic Web is to improve the Internet by extending interoperability between computer systems using smart agents and applications that seek information without human intervention (Barnaghi, Wang, Henson, & Taylor, 2012). These works help to solve the problem of device interconnection, but it is not an easy task, because technological diversity of smart devices must be taken into account, as well as the correct handling of the context, which is not always considered.
Besides, the development of context-sensitive software has proved successful (Perera, Zaslavsky, Christen, & Georgakopoulos, 2014). IoT devices are becoming intelligent thanks to the information gathered about the context in which they are located, from near people or other devices. In order to reduce the interaction among people and devices, this interconnection must be adapted to the context. These drawbacks can be addressed by developing software capable of adapting its behaviour to people’s needs (Perera et al., 2014; Taivalsaari & Mikkonen, 2017). In this way, several research areas can contribute to provide this adaptation, such us Context Oriented Programming (COP), Ambient Intelligent (AmI), Semantic Web, and Machine Learning (ML). Most of these paradigms allow us to define behaviours for different scenarios at design time, so the adaptation of the devices is limited to situations that developers have been able to identify, making it impossible to adapt them to other situations that may arise from the context.