13.2.1 The Concept of the Smart Connected Home

There is no commonly accepted definition of the smart connected home, but in general, a smart home may denote any kind of residence (e.g., apartment, cottage, and rented living space), which involves information and communication technologies allowing for remote control, monitoring, and access (Balta-Ozkan et al., 2013). Meanwhile, some researchers further add that a smart home needs to have ambient intelligence and automatic control allowing it to recognize and possibly make decisions on its own guided by the behavior of the residents (De Silva et al., 2012; Pedrasa et al., 2010; Zhang et al., 2013).

Consequently, it can be observed that there are three types of homes: smart home, connected home, and a smart connected home. The relation between these types is illustrated in Figure 13.1.

Smart home includes systems that allow the residents to operate home appliances, typically only locally from within the house. This type of smart home tends to rely on wireline-based standards such as KNX and it is not connected to the Internet. Commonly, it tends to focus on the automation of lighting, windows, and in-house entertainment, and is associated with building automation.

Connected home allows for remote control and management of appliances, typically over IP-based networks (e.g., the Internet). Additionally, this type of home usually provides services that promote and support, for example, security, health care, and energy management. Moreover, this type of house generally includes a central hub (gateway) from which the system can be controlled together with a user interface that can be operated typically through a smartphone.

Smart connected home includes functionality from both previous types, but also adds communication and service-exchange between related areas, such as the grid and electric vehicle and on-site microgeneration (e.g., rooftop solar panels) (Balta-Ozkan et al., 2013). Smart connected homes thus merge the functionality from both the connected home and the smart home. It may also include system capabilities, such as learning, prediction of, and response to the occupants' needs and lifestyle preferences in their home environment. Typically, this type of home incorporates connected devices that exhibit some form of “intelligent” logic commonly based on machine learning algorithms for activity recognition and to perform some action automatically. In implementing this, cloud services are often used. In this case, cloud-based applications analyze the data collected and processed from the smart connected home and they are often able to take actions, sometimes autonomously. For example, a smart leak detection system may notify the home owner that there is a leak in the water heater system, and as well automatically turn off the water, gas, or electricity supply to prevent valuable resources and save money. Additionally, homes in this category may also feature advanced forms of interaction typically supporting voice detection, facial recognition, and gesture controls.

Conforming to the current home automation trends, the emphasis of this work is on the smart connected home. So, hereinafter, the term smart connected home is used to denote a residence incorporating a range of sensors, systems, and devices that can be remotely accessed, controlled, and monitored via a communication network such as the Internet. In the rest of this text, this is also associated with the concept of the do-it-yourself (DIY) home, where end users can build their own smart connected home through computing units (e.g., Arduino, Banana Pi, and Raspberry Pi) without relying on professionals.

13.2.2 Smart Connected Home Stakeholders

In IoT applications, such as smart homes or smart buildings, there are diverse stakeholders ranging from technology investors, technology developers, technology integrators, and more. Additionally, specific stakeholders may be involved depending on the actual smart home system. For instance, it is common to have a specific stakeholder such as a meter point operator in an energy-focused smart connected home, and a content provider in an entertainment-focused smart connected home.

In general, there are six main stakeholders that can be identified as described below. Figure 13.2 illustrates these stakeholders.

• Device Manufacturers. Appliance manufacturers, including smart product suppliers such as smart meter and entertainment device manufacturers. Householders may purchase devices directly from manufacturers but oftentimes through retailers or service providers. Examples of manufacturers are Samsung, Honeywell, and LG.
• Service Providers. Application service providers and utility companies that provide the end users with hardware equipment to support or enable different smart connected home services. Three examples of service providers are Verisure, AT&T, and Leak Defense.
• Network Providers. Telecommunication providers supply and manage network infrastructure like core network, radio access network, and interconnectivity network to service providers that want to offer smart home services. Effectively they are the stakeholders that connect the householders to the Internet. An example of a network provider is, for instance, Verizon.
• Regulators. Regulators are external entities overseeing business services or specific industry sectors. This can include certification and accreditation bodies related to quality, security, and safety. An example of this could be a privacy regulator that develops laws to safeguard personally identifiable information from undue exploitation. Possibly, this entity may affect the entire spectrum of industry stakeholders.
• Platform Providers. Entities supplying mechanisms, tools, and frameworks assisting in overcoming integration challenges and facilitating easier assess, customization, and automation support. Commonly, different device manufacturers work with different platform providers for third-party integration. Some organizations serving this role are Apple, Google, and Amazon.
• End Users. The end user is the stakeholder that uses services. Typically, this represents the home residents that purchase and operate the different smart connected home devices and services.

13.3.1 Energy

Energy systems are targeted to provide efficient energy consumption and management for the home. The energy domain commonly involves the use of smart meters, smart thermostats, and adaptive lighting systems. System architectures in this domain may utilize “intelligent” multiagent systems and control strategies to predict and automatically maximize energy efficiency and user comfort (Reinisch et al., 2011). Other projects put the onus of monitoring energy consumption to the householders. This can be assisted, for instance, by smart energy monitors that display in real time the energy consumption patterns, for example, the heating, hot water usage, and CO₂ emissions (Hargreaves et al., 2013; Qin et al., 2014). Two examples of commercial systems include Google Nest⁷ and Ecobee⁸ smart thermostats. Both of these systems allow for preventing a home from overheating or overcooling. They also include functionality to automatically adjust the house temperature to fit the occupants' preferences, and also allow for remote adjustment of the temperature.

The energy management domain appears to be an area where there is numerous research projects and funds especially related to reducing energy usage and optimizing its utilization in houses and commercial buildings. The driving forcing behind this tends to be associated with the rising energy costs, and reducing greenhouse gas emissions (arguably, a primary cause of global climate change). However, energy management is also important because it also contributes for peace of mind. For instance, by automatically turning on lights to help discourage potential intruders while the residents are away from home.

13.3.2 Entertainment

Smart connected home systems tend to promote entertainment typically maximizing occupants' comfort and convenience by providing personalized amusement content and social communication services. The entertainment sector commonly involves game consoles, connected TVs, and smart speaker systems. Microsoft's Easy Living Project (Brumitt et al., 2000) is an architecture that uses cameras to tailor services according to the occupants' location within the house. This system is often mentioned as a case study in well-cited surveys (Balta-Ozkan et al., 2013; De Silva et al., 2012). More recent systems adopt smart home technologies allowing the residents to get an idea of outside weather conditions without having to leave the house. This can be done, for instance, through the use of smart cameras and temperature sensors (Wang et al., 2015). In addition to that, some systems also capture and try to alter the individuals' social emotions. One way of implementing this is by utilizing smart furniture such as tables augmented with cameras and motion sensors (Yu et al., 2012). These sensors analyze the viewer's facial expressions (e.g., eyes and mouth) and then play the user's favorite music and supporting sentences to alleviate his/her mood (Yu et al., 2012). Two recent industry examples are Samsung Smart TV⁹ and Apple TV.¹⁰ Both of these systems are connected to the Internet, allowing end users to stream content from services like Netflix, and access more services, for example, a web browser, through built-in apps.

It can be observed that the entertainment domain is emerging as one of the top reasons why people purchase a smart home system. Mostly, this is for the ability to remotely control or monitor TVs and sound systems.¹¹ Additionally, it can be noted that the technologies in this domain tend to rely on microphones and cameras and may support different forms of advanced interaction, most likely in the form of speech recognition.

13.3.3 Health Care

The health care service area is focused on providing mobile health care and fitness support, and aims to provide independent healthy living. In comparison to the other domains, the health care service also involves the use of wearable sensors (e.g., wrist straps) in order to allow for possibly continuous monitoring of body signal parameters (e.g., cardiac diseases) even while not at home. Health care services can monitor the residents' personal health, generate tailored health reports, and may support remote diagnoses and chronic disease management. Generally, this includes connected devices such as wireless scales, fitness monitors, and physiological devices. Recent projects connected to the health care domain tend to utilize audio technologies, for example, speech recognition, to facilitate and ease daily living for elderly and frail persons (Amiribesheli et al., 2015; Ni et al., 2015; Portet et al., 2013). It is also common to have architectures that support continuous elderly monitoring, for example, by checking the health indicators of persons such as glucose levels. This can be implemented for instance by ontologies, web service technologies, and specialized sensors (Rivero-Espinosa et al., 2013). Notable examples from the industry include the in-home patient monitoring systems Vignet platform¹² and Philips TeleStation.¹³ These systems support connecting wireless measurement devices, such as physiological monitors, allowing individuals to be active participants in their health care.

The health care domain has been extensively researched by many authors and the amount of publications on connected health care is continuously increasing. Monitoring a person's cognitive and physical health is a precursor to ensuring a healthy society. On the flip side, it can be noted that this application involves the most privacy invasive sensors with technologies that can monitor the patient's intimate physiological conditions, such as blood glucose levels.

13.3.4 Security

Systems in the security category are often targeted to offer services, which are designed to monitor, detect, and control security and safety threats. Smart home security and safety systems typically range from remote entrance monitoring services to systems that automatically recognize physical threats, such as a fire or a burglary, and autonomously take the corresponding action. Such actions can include for instance starting the immediate fire sprinklers or automatically switching on or off appliances, such as digital locks or video cameras. Typically, this domain includes functionality that support smart door locks, cameras, and alarm systems. In the market, there are many instances of security systems, and two such examples are Verisure Securitas Direct System¹⁴ and Alarm.com.¹⁵ Both systems allow for remote monitoring and managing security systems in the home, and include 24/7 professional monitoring and emergency response solutions.

As a central observation, it can be noted that the security domain is one of the main motivations for consumers to purchase smart home systems,¹⁶ despite the various privacy concerns related to in-house video surveillance (Brezovan and Badica, 2013). Smart connected homes security systems are especially popular among residents living in towns where there are reports of high criminal activities. However, their widespread adoption is also linked to the convenience factor being offered by such systems. Furthermore, it can be noted that this domain is tightly linked to surveillance systems, involving technologies such as cameras and motion sensors that are predominantly used in intelligent buildings.

13.4.1 Sensors and Actuators

Sensors are nodes that are used to sense, measure, and detect changes in the environment including the occupants. In smart connected homes, sensors that record the temperature of an object, for example, room, or body, are used across all the previously introduced systems. Similarly, some sensor types, for example, CO/CO₂, are used in multiple domains for different purposes. For instance, these may be used for tracking CO₂ emissions in the energy domain and to indicate possible CO leakages in the security domain. Some domains, in particular the health care domain, employ specialized sensor types, for example, physiological sensors, that are exclusively used in that service area. While physiological sensors tend to be mobile and are generally worn by householders, other sensor types are often fixed, for example, wall-mounted camera, or attached to an object, for example, door contact sensor. In summary, Table 13.1 presents the different sensor types used by the different smart home systems introduced in Section 13.3. The sensor types are classified according to the type of data or parameters they measure (Amiribesheli et al., 2015).

Actuators provide the means to implement physical actions, such as turning on/off lights, raising alarms, and activating heating appliances. In smart connected homes, typical examples of these include motor controllers, switches, keys/locks, speakers, and displays. Actuators may influence sensors and can consequently cause the system or a user to activate the actuator. It can be argued that the more sophisticated and autonomous this type of interaction is, the more the connected home system is seen as smart.

In practice, it is common to assemble different sensors and actuators in a single unit. For example, Canary,¹⁷ a home security solution, has a built-in HD camera, microphone, temperature, humidity, and air quality sensors, as well as a siren.

13.4.2 Gateways

Smart connected home gateways are devices that compile, convert, and transmit information collected through sensors, sensor hubs, and commands from devices, of which some may be smart. The gateway acts as the component responsible for interfacing with the outside world allowing for routing internal network data in and out to the Internet as well as providing services to the residents. It can range from a dedicated device, smartphone, to a computing unit such as a Raspberry Pi.

This gateway tends to be connected to the home's broadband router. This is then typically connected to the Internet via a wired connection (e.g., Ethernet connection to a cable or DSL modem). Nonetheless, it is possible to have a gateway embedded with a wireless/cellular radio. In this case, the device may communicate directly to the Internet or a cellular carrier.

In setting the requirements for home gateways, the Home Gateway Initiative (HGI)¹⁸ standardization organization is actively involved. This organization is made up of multiple telecommunication companies and device manufacturers, and it is actively involved in improving the interpretability of the gateway with smart home devices. HGI has published various guidelines, standards, and requirement documents to promote an open and modular deployment for home gateways to provide compatibility among various devices. Modern gateways would typically offer the following functions:

• Notifications. Most gateways implement some form of notification system supporting the routing of messages and alerts from connected devices to the end user. For instance, in the case of an emergency, when a washing machine is flooding, the house owner might get notified of this occurrence. Typically, such alerts are in the form of SMS text messages, emails, or prompts/pop-ups displayed through a mobile application installed on an end user device such as smartphone.
• Automation. Some gateways support an automation or scheduling system that allows for the creation of different rules for the connected devices. For instance, some systems allow for the creation of rules based on time, for example, sunrise and sunset rules, and others allow for more granular rules based on events and conditions. An example of this could be a rule that automatically turns on the lights when movement is detected in the home.
• Local Control. Gateways may also feature local control. This allows for manual control and execution of rules in case the Internet or cloud services are unavailable. This interface is implemented usually in the form of a local web browser, mobile, or desktop application. However, some devices also offer a built-in display panel.
• Cloud Service. Commonly, major gateways utilize a cloud service(s). This allows for remote device management, however, oftentimes the cloud is also used for storing home data, serving the user interface, and communicating with external systems such as weather, push messages, and SMS gateways.
• Third-Party Integration. With an increasing number of smart home devices and an ensuing variety of standards, there is a need to allow for connections between dissimilar products. Modern gateways facilitate this by offering integration with third-party services, such as, Amazon Echo, Apple HomeKit, and “If This Then That” (IFTTT¹⁹), that allow so. Amazon Echo is a cloud-based, voice-activated platform that makes it possible to control connected devices. Apple HomeKit is a platform designed to make different devices communicate with each other possibly with the help of compatible bridges or hubs. IFTTT is a web-based service allowing users to pull together multiple services and to program devices, through a construct known as “recipes,” to run routines, react to triggers, or pass commands to other devices in the house. An example of an IFTTT recipe could be one that automatically powers on a smart TV after a certain amount of physical activity, for example, walking 10,000 steps, has been performed by the residents. An alternative application to IFTTT that is arguably more flexible in terms of automation support is Stringify.²⁰ This works similarly to IFTTT but instead of using “recipes” it uses a concept known as “flows.” Through flows multiple applications and services can be chained together. As an example, a flow could be designed to improve the home comfort by automatically adjusting the house lighting and room temperature through voice commands. The actual implementation of this using Stringify may consist of three “things”:²¹ Amazon Alexa, Nest thermostat, and Philips Hue Bulb. Amazon Alexa voice service (supported for instance by Amazon Echo device) can be programmed to wait for the householders to trigger it, for example, by saying “Alexa, tell Stringify Night Mode.” This trigger can then consequently adjust Hue to amber at 15% and Nest thermostat to set the room temperature to 20 °C. This setup is shown in Figure 13.4. Implementing this with IFTTT, especially if the technologies are supplied by different manufacturers, would commonly require multiple recipes.
• API/SDK. Typically, most hubs feature an Application Programming Interface (API) and/or Software Development Kit (SDK). These allow for interfacing and building customized functionality. For instance, an Alexa smart home ecosystem contains a Smart Home Skill API²² that enables the creation of capabilities (e.g., ability to play music, answer questions, provide weather forecasts) to control cloud-connected devices. Another example could be a Smart TV App Store that has a SDK allowing for screen control, app control, and download and upload via network.

13.4.3 End User Client Devices

Typically, end users access, control, and monitor the smart connected home functionality through a user interface on an end user device. Taking advantage of mobile devices, the residents can check the information of the smart home system, schedule different tasks, and send immediate commands to be performed.

For instance, end users might issue commands to open their home door or to monitor their children remotely through an application available on their smartphones, tablets, desktop computers, or specialized devices such as smart remote control systems. Furthermore, these devices may recognize speech input, hand gestures, and perform user tracking.

Commonly, a specific application is required to manage and control different smart home devices issued by different manufacturers. For instance, to control smart locks, lights, and thermostat, one would commonly require separate applications for each gadget. However, the emergence of standards in the future may allow some integration between such systems/devices, and thus, use of less number of applications.

13.4.4 Cloud Services

Smart connected home systems may involve the use of cloud services that collect and analyze data readings from home devices. Such systems tend to be implemented over private, public, or hybrid backend infrastructures (clouds) and may include service application or delivery platforms operated and managed by service providers and device manufacturers (Hassan, 2011).

Commonly, the platforms or cloud service delivery models adopt the Platform as a Service (PaaS) or the Software as a Service (SaaS) cloud resource provisioning solution. The PaaS refers to the platforms that provide cloud computing services exposed in the form of APIs, programming languages, middleware, and frameworks, allowing for data storage functionality, device management, access control, and more. SaaS solutions, for example, IFTTT, tend to be centered on data mash-up using cloud computing facilities.

Chapter 4 provides more information about the different cloud service delivery models and the utilization of cloud computing in IoT.

13.4.5 Integration Platforms

Integration platforms or frameworks are technologies that allow for better interoperability and support across the diverse smart connected home environment. Examples of current platforms that are developed and supported by companies are Apple HomeKit, Google Brillo/Weave, Allseen AllJoyn,²³ Samsung SmartThings,²⁴ and Amazon Alexa. These smart home frameworks also provide a programming framework for developers to build apps that realize smart home benefits. Moreover, there are also open source home automation platforms built on open protocols and under open source license. Three such systems are openHAB,²⁵ OpenMotics,²⁶ and Domoticz.²⁷ Commonly, these connected home platforms are popular within the DIY communities. DIY users install and configure the smart home devices themselves instead of relying on professionals.

13.4.6 Communication Protocols and Models

The Internet protocol (IP) is the basis of the IoT, and hence, it is widely used in smart connected homes. Due to its easy interoperability, pervasive nature, widespread adoption, and recent research of lightweight interfaces, the IP is considered vital to the success of the smart connected home (Kailas et al., 2012). However, non-IP-based devices can also be part of the home.

Smart connected home systems often use many different types of communication protocols. These range from wired protocols (e.g., X10, HomePlug, and LonWorks) to wireless communication protocols (e.g., Wi-Fi, Z-Wave, and ZigBee), open standards (e.g., ZigBee) to proprietary (e.g., Z-Wave), and short-range protocols (e.g., NFC, RFID, and 6LowPAN) to long-range protocols (e.g., NWave, Sigfox, and LoRaWAN).

It is also important to distinguish between low-level and high-level protocols. Low-level protocols (e.g., ZigBee) are principally used for device-to-device networking within the house and in the case of long-range protocols (e.g., 4G) to communicate with service providers. High-level protocols are used to transmit sensor data or receive remote commands through the home gateway. Examples of these are MQTT, CoAP, and XMPP. Additionally, there are high-level domain-specific protocols such as the Smart Energy Profile 2.0²⁸ that connect the home energy devices to the smart grid.

Smart connected home devices can communicate with each other through heterogeneous communication protocols and by using different models. Typically, there are three communication models for IoT environments, such as, the smart connected home: device-to-device, device-to-cloud, and device-to-gateway (Rose et al., 2015). In the device-to-device model, devices exchange direct messages with each other through low-level protocols. An example of this could be a smart lock sending a signal to turn on the smart light when a door is opened. In the device-to-cloud model, devices are instead connected to a service provider to exchange and control messages. This model is used for instance by Nest smart thermostat, for example, to support remote control through a smartphone. Most consumer devices rely on the device-to-gateway model; in this model, the gateway acts as the mediator facilitating the interaction between the devices and cloud services. As an example, a smart scale may be connected to a smart watch indirectly through the help of a local gateway and cloud services. Such a setup can be used, for instance, to help householders stay motivated and on track regarding their weight goals even when not physically being in the proximity of the scale. Figure 13.5 illustrates the three identified communication models.

13.5.1 Centralized

In the centralized model, all the data are retrieved by a single central entity (Roman et al., 2013). This entity is oftentimes a dedicated local gateway that houses the application logic, stores data, and communicates with the IoT devices that are connected to it.

Another option could be to have the cloud acting as the central connection hub (Spender, 2015). In terms of information flow, the central node collects all the information from smart devices within the home network and acts on it. Consequently, end users need to connect typically through the Internet to access the services provided by the central entity (Roman et al., 2013). This approach is, for instance, adopted by Amazon Echo that requires a constant active connection to the Internet (e.g., to support Alexa cloud-based voice service), and is broadly used by other commercial companies such as Google and Apple.

One of the challenges with the centralized architecture is that implementing effective privacy measures in this solution is arguably less flexible. For instance, here since the “intelligence” is positioned at the central entity, the other entities (the edge devices) do not have much control over the data they generate and process. Instead, this is delegated to the central object that can decide what data elements to process, whether to share a particular data stream, and so on (Roman et al., 2013).

13.5.2 Distributed

In the distributed approach, all the devices are self-sufficient capable to retrieve, process, assimilate, and provide information and services to other entities (Roman et al., 2013). In comparison to the centralized approach where information flows through a common node, in the distributed model, the architecture is similar to that of a peer-to-peer system and information is only exchanged when needed. It can be argued that connected devices in this model are considered smart on their own. Thus, in this case the majority of decisions are done locally and the Internet does not have a major role as is the case for a centralized architecture. However, it may still be needed for coordination and analysis (Spender, 2015).

An example of this could be a connected oven that detects the presence of connected light bulbs and flashes them when it finished cooking. Given the market fragmentation, this approach is difficult to have in practice. Distributed architectures may leverage a service-oriented or a mobile agent approach (Badica et al., 2013). One technology that is used to implement this type of architecture is Jini platform. This is used, for instance, by ProSyst Software company to provide products to control household devices (Toschi et al., 2016).

One of the challenges with the distributed model is that applying network security is more complex when compared to the centralized approach. In this architecture, any smart device can connect with any node at any time, not necessary knowing each other beforehand. This, especially combined with the fact that some devices might be heavily constrained, makes key management here a significant challenge (Roman et al., 2013).

13.6.1 Interoperability

Interoperability is the ability of different devices and services to work together. In the smart home environment, a use case could be when a wearable device, such as a smart watch, detects a householder waking up, and sends a signal to the connected coffee machine to start the brewing process. As discussed earlier, there are diverse integration platforms available in the market. These platforms, for example, Apple HomeKit, help in implementing such a use case, however, if the devices were manufactured by different vendors making them work together might be problematic. Albeit, they may speak IP, disparate protocols, cryptographic algorithms, and proprietary certificate formats interoperability make true interoperability a challenge (Cloud Security Alliance, 2016).

Additionally, an organization that controls different parts of a vertical market (e.g., the smart entertainment) may dominate a market, stifling competition, and creating barriers of entry for competitors. Dissimilar standards, proprietary technologies, and architectures can also lock consumers into one ecosystem of products making it hard to transfer their data from one platform to that of a competing device manufacturer or service provider. For instance, a related case occurred when the company selling the Revolv Hub home automation system²⁹ shut down its servers making the hardware and related apps useless.

A central comment is that interoperability is key to build generic smart home solutions (Ko et al., 2011), to open markets to competitive solutions in IoT (Misra et al., 2015), and as well to ease integration with the current Internet. Interoperability is a necessary component to evolve the smart connected home to a more intelligent home where devices not only connect together but are able to collaborate together toward achieving a unified goal. In achieving this, possibly devices and applications need to be decoupled. Furthermore, this may require architectures that lean toward the distributed or hybrid approaches combining the benefits of centralized and distributed styles. However, at the moment the device and the applications (apps) tend to be handled by the same stakeholder (typically the device manufacturer). Also, most of the business models adopted by vendors and most commercial applications and research works focus on cloud-based models utilizing centralized architectures.

13.6.2 Security and Privacy

Smart home devices are tightly coupled with the householders, especially in terms of their lifestyles, habits, and sensitive data such as personal photos, videos, and digital diaries. Therefore, all the information gathered, exchanged, and stored within the smart connected home, an environment meant to serve as a private space, can lead to severe security and privacy risks. While the risks vary across the different smart home systems and services, detailed personal information can be collected by devices, and used for profiling individual behavior, amongst other things.

An example, a smart thermostat may collect device usage statistics and environmental data to learn the householders' behavior. While for beneficial reasons, this data may be shared with legitimate entities such as energy providers for energy efficiency purposes, but may also be eavesdropped by malicious threat agents such as cybercriminals. This information can then be used potentially to steal the house at a time when there is no peak activity being registered on the device implying that the residents may be away. Such a scenario and related ones raise important questions regarding data governance and as to whether collected data are being kept private and processed in a secured manner by all stakeholders across the chain.

Additionally, as the smart connected home integrates more devices, it provides more entry points for malicious software and different threat sources. With heterogeneous devices in the home, the risk impact can range from low to critical. For example, attacking an Internet-connected device may disrupt a service such as preventing a device like Amazon Echo to go online and answer a query. However, attacking an actuator, such as a wireless insulin pump or a network-connected pacemaker, may have life-threatening outcomes. Moreover, the complexity of software APIs, integration middleware, and software stacks create opportunities for different threat agents to exploit. A schematic illustration of malicious threat agents attacking a smart connected home is shown in Figure 13.7.

Thus, a central concern is to put more effort in studying both malicious threat agents and their capabilities. This is an important first step to conduct a privacy and security risk assessment and eventually to develop effective risk management strategies and countermeasures. Especially in the home, which by most people is considered to be the most private and intimate place, there is also an urgent need for more knowledge on the sensitivity of the data influx, as well as, on the means to protect it from harm or misuse. However, work in this area falls behind especially in providing generic data models for smart connected homes.

13.6.3 Reliability

Smart connected homes must neither fail nor behave in unpredictable manners if the things go amiss. However, in comparison to the traditional electrical and electronic devices such as televisions, microwave ovens, and camcorders, IoT devices are of greater complexity and arguably not as reliable. Alongside with fault tolerance, reliability is a principal challenge that must be overcome for the smart home to gain a wider acceptance (Friedewald et al., 2005). Reliability is especially important for functions that control the physical environment as not doing so may result in accidents (e.g., people could fall off the stairs when lights are switched off unexpectedly) (Friedewald et al., 2005).

There are also several different facets of the reliability challenge. Specifically, when comparing connected devices to “traditional” computer systems, there are differences in the development culture, technological approaches, expectations of the markets, and regulations, as factors that affect reliability (Edwards and Grinter, 2001). These factors extend beyond the research community to stakeholders that develop, regulate, and consume these services (Edwards and Grinter, 2001).

Another aspect is that the smart connected home being integrated with external service providers transformed the house into a system of systems. This is evident when looking, for instance, at service areas, such as energy. This encompasses devices such as smart meters and smart sockets that integrate the home possibly to a smart grid infrastructure. While the benefits are many, this may also result in the home to behave in unreliable somewhat unpredictable ways. For instance, service outages, software malfunctions, and performance delays can be the direct consequence of problems in external systems. A specific question this scenario raises is how to troubleshoot such problems. This eventually makes it crucial to have resilient architectures that allow for fault-tolerance, standard protocols and interfaces, and regulatory bodies that supervise the development of such systems.

As an observation, it can also be noted that ensuring reliability in smart connected homes is key for assisted living. This is particularly crucial for the health care domain. In this area, medical devices making inaccurate measurements or inferences about the nature of the individual's behavior or health profile could result in severe or life-threatening injuries. Thus, the technology developers have to incorporate means of dealing with errors, particularly in safety critical situations. In solving this challenge, regulations and standards that guarantee a certain level of safety to the householders is key.

13.6.4 Usability

Smart home devices may be used by nonexpert users. These individuals are not familiar with inner workings of technologies. Usability defines the way an end user will interact with various smart devices and system preferences (Holroyd et al., 2010). Fundamentally, it is the quality that provides for ease of use and simplicity from the end user point of view. For instance, a generic user may have difficulty to use strong authentication mechanisms that may require a lengthy text-based password or PIN code. Usability makes the system from the end user perspective characterized as simple, transparent, and unobtrusive, ideally without compromising security and privacy.

Diverse recent advancements have been made to tackle the usability challenge. In particular, voice activation appears to be the leading natural user interface method that facilitates interaction between the householders and the smart home services (Allied Business Intelligence, 2016). Despite this, in order to have voice sensing effective, it must be able to capture voice from anywhere in the house not necessarily limited to rooms that are in close proximity to the listening devices. An alternative to voice sensing is gesture detection. This makes it easier for elderly and disabled people to interact with smart home systems through simple gestures in front of cameras (Bien et al., 2005). However, this method may not be the ideal for people with very limited or impaired body movements. Likewise, certain groups of people may prefer interacting with a smart home through a physical device (e.g. computer, smartphone, or remote control) instead of through modalities such as speech and gesture (Jeong et al., 2009). Moreover, adaptive smart home interfaces that can be tailored to fit the needs of multiple user groups possibly differing in gender, age, and physical capabilities are needed (Jeong et al., 2009).

It can be observed that more research effort is needed to find the right approach to connect end users with the smart connected home system, with the objective for a high degree of usability. Over the recent years, research into Augmented Reality and Brain Computer Interface has been getting a lot of traction to improve ways of interaction with the physical environment. Despite this, their application to commercial smart connected homes has not yet gained momentum. In solving this, standards for usability are critical, and a cross-collaboration between engineering science, behavioral science, and psychological research is needed.