Systematic Analysis of Safety and Security Risks in Smart Homes

: The revolution in Internet of Things (IoT)-based devices and applications has provided smart applications for humans. These applications range from healthcare to traffic-flow management, to communication devices, to smart security devices, and many others. In particular, government and pri-vate organizations are showing significant interest in IoT-enabled applications for smart homes. Despite the perceived benefits and interest, human safety is also a key concern. This research is aimed at systematically analyzing the available literature on smart homes and identifying areas of concern or risk with a view to supporting the design of safe and secure smart homes. For this systematic review process, relevant work in the most highly regarded journals published in the period 2016–2020(a section of 2020 is included) was analyzed. A final set of 99 relevant articles (journal articles, book sections, conference papers, and survey papers) was analyzed in this study. This analysis is focused on three research questions and relevant keywords. The systematic analysis results and key insights will help researchers and practitioners to make more informed decisions when dealing with the safety and security risks of smart homes, especially in emergency situations.

homes have been applied to other areas, such as smart healthcare systems [3], energy systems [4], security and emergency management systems [5], and comfort and entertainment systems [6].
A smart living environment's automatic transformation mechanism is controlled by the arguing system, which is considered the brain of the smart living architectures. The smart home arguing system (SHAS) is used to make accurate decisions ensuring the security, safety, and comfort of the inhabitants and their surroundings. Keeping this in mind, researchers have intensively explored and studied multiple artificial intelligence (AI)-based systems and their uses for ambient-assisted living (AAL) systems. However, no systematic mechanism has been proposed that investigates the decision support systems (DSS) incorporated in smart living environments. Chan et al. [7] and Calvaresi et al. [8] reviewed smart living environments. Others, including Zaidan et al. [9] and Brand et al. [10], studied smart homes from single application domain perspectives, such as IoT-based communication components, and privacy concerns of IoT components in living environments. Furthermore, Wilson et al. [11] reported on smart home residents and their uses of the underlying infrastructure from socio-technical perspectives. To the best of our knowledge, there is a lack of comprehensive systematic literature review of the current context of human safety and security in smart homes.
In this paper, a systematic literature review (SLR) is presented that examines the aforementioned domain of smart homes from the following perspectives: • Determine the main goal of smart living systems; • Deliver new understanding and new knowledge about human safety in a rapidly evolving smart built environment; • Identify and characterize the key features and requirements of smart living to ensure safety and security during emergencies; • Catalogue the tools and features that provide the abilities to monitor, collect, and process real-time human sensors, and build data for efficient and effective decision making.
The rest of this paper is organized as follows. In Section 2, the proposed SLR process adapted for literature search and analysis is described. SLR results are reported and discussed in Section 3. The limitations of the proposed research work are outlined in Section 4, followed by conclusions and directions of planned future work in Section 5.

Review Process
SLR is a well-known process used to identify and evaluate available research work relevant to a particular subject or event of interest. SLR aims to present a fair evaluation of a particular research topic using a rigorous, trustworthy, and auditable methodology [12]. SLR studies are reported in many fields, e.g., networking PMIPV6 domains [13] and healthcare Big Data analytics [14]. The main reasons for performing a SLR are the following: • To explore and summarize the existing research work on a particular technology; • To find the gaps in the available technology that will ultimately lead to future investigations.
In the proposed research work, the SLR process is conducted by following the guidelines suggested by Kitchenham et al. [15,16]. The proposed review process protocol is depicted in Fig. 1 and consists of three main phases: (1) planning the review process, (2) performing the review process, and (3) reporting the results of the review process. All these steps are illustrated in detail in Fig. 1.

Review Process Planning
The proposed systematic review process is performed using the guidelines provided by Kitchenham [16,17]. Based on these guidelines, researchers must explore the existing evidence on a topic of interest and determine the need for a review process. Furthermore, the study outlines the key pre-review activities: formulation of the research questions, identification of the keywords and query formulation, selection of the peer-reviewed online digital libraries for the accumulation of the relevant primary articles for the review process, and the inclusion/exclusion criteria. Accordingly, the present systematic review was performed due to recent increasing research interest in smart homes and living environments. It is noted that there is a lack of systematic literature review study in the specific context of smart home safety and security risks, particularly during emergency situations.

Research Question Formulation
As discussed earlier, research question formulation is an essential activity in conducting a SLR. For defining the most relevant research questions, the Goal-Questions-Metrics approach of Van Solingen et al. [18] was followed, and, consequently, the three research questions given in Tab. 1 were formulated.

RQ2
What hardware components are used for early alarm mechanisms in emergency situations?
Aims to uncover the hardware devices or applications reported in the literature for early alarm purposes in case of emergencies.

RQ3
What proactive approaches are developed to reduce the loss of life or injuries in smart buildings during emergency situations?
Based on the literature, aims to accumulate knowledge on various proactive approaches that have been developed to reduce the human loss of life or injuries in emergency situations by providing guidance, communications, and monitoring facilities.

Identification of Keywords and Query Formulation
After formulating the research questions, the next and most important activity of the SLR was to identify keywords and formulate a search query to systematically select the most relevant articles from the selected online digital libraries. Keywords and the search query are shown in Tab. 2. Table 2: Keywords and query for search process ("SMART BUILDINGS" OR "SMART HOME" OR "SMART HOMES" OR "SMART LIVING ENVIROMENTS") AND ("SECURITY" OR "SAFETY" OR "RISKS" OR "THREATS") AND ("EMERGENCY SERVICES" OR "EMERGENCY EVACUATION" OR "COMMUNICATION DEVICES" OR "AUTO CONTROL DEVICES" OR "IOT DEVICES")

Online Digital Library Selection
The five well-known online digital libraries shown in Tab. 3 were selected to identify relevant research. The primary articles were selected and downloaded from these digital libraries based on the defined query. After downloading, these papers were further analyzed and explored to remove the redundant and irrelevant papers, if any. Irrelevant papers, i.e., those that did discuss smart buildings or smart homes, were removed. The finalized primary articles were then merged into a single set or directory. These are the papers that address at least one of the research questions. In summary, 99 research articles were selected as most relevant primary research articles for the purpose of the present SLR study.

Inclusion/Exclusion Criteria
Defining inclusion/exclusion criteria is the most challenging job in a SLR research process. As the main activity of the process, it ensures the selection of the most relevant primary articles for the final pool of papers for further quality assessment. Tab. 4 presents the inclusion/exclusion criteria of the primary studies relevant to the aforementioned proposed research questions. 3) Include papers published between 2016 and 2020. 4) Does the paper's title reflects enough knowledge about smart homes and smart living environments, and does the paper contain information for emergency exit and safety precautions? 5) Determine whether the abstract provides enough information about the smart buildings and safety measures during emergency situations. 6) Do the contents in the paper provide proper validation?
Exclusion criteria 1) Exclude papers written in languages other than English.
3) Exclude research papers containing less than three pages. 4) Exclude papers that fail to satisfy the inclusion criteria.
The selected primary research articles were thoroughly checked and analyzed by all the authors to ensure research validity and avoid any possible omission. Significant attention is given to the last three questions, which are defined for the inclusion of a paper in this study. A voting mechanism was considered for this step. If more than half of the authors agreed to the inclusion of the paper, then the paper was added to the final set of the most relevant articles; otherwise, the paper was excluded. This voting mechanism is based on the paper's title, abstract, and contents presented in the research article itself. A summary of the overall inclusion process is shown in Tab. 5. A final pool of 99 relevant primary articles was selected for the assessment process.

Review Process
After performing the preliminary steps of selecting the online digital libraries, identifying research questions, selecting keywords, and determining inclusion/exclusion criteria, the next phase was to perform the review using the research protocol selected as shown in Fig. 1. This activity includes defining final pool selection, data synthesis, monitoring, and quality assessment processes. All of these steps are discussed in detail in the following subsections.

Final Pool of Relevant Articles
After sorting the selected digital libraries for the relevant primary articles and performing the inclusion/exclusion process, a final set of 99 relevant articles were selected for SLR. The final pool of papers includes workshop papers, conference proceedings, book sections, journal articles, and review/survey articles. Tab. 6 presents details of the selected finalized pool of research articles. The mapping of paper sources is shown in Fig. 2, from which it is concluded that IEEE Xplore contains more papers when compared to other sources. This may reflect the interest of researchers to publish their research work in IEEE Xplore. The final paper set was also sorted based on year and publication sources. It is observed that the number of publications increases with the passage of time, which may reflect the growing interest among the research community in the proposed area of research. Fig. 3 shows the sorting results. Furthermore, the final set of relevant papers was sorted based on digital library, reference list, type of paper, and publication year, as depicted in Fig. 4. In the figure, the outer shell represents the references to the primary articles, the medium shell the paper type, and the most-inner-shell date of publication.

Quality Assessment
The quality of the selected relevant articles was assessed using the criteria defined in the SLR protocol. The set of relevant articles was reviewed and assessed against each of the research questions and corresponding criteria (QC) listed in Tab. 7.
This assessment ensured the quality of each selected paper for SLR. Furthermore, weights were assigned to all research questions based on the following criteria: • 0 if a paper has no information for the selected research question; • 0.5 if a paper has a piece of partial but satisfactory information about a research question; • 1 if a paper contains a full and complete description for the research question.
After performing the quality assessment, the most relevant articles were determined, as shown in Fig. 5.  Whether the paper provides detailed information about architecture or design followed for development of smart buildings in smart cities. QC2 Whether the paper provides in-depth knowledge and understanding about human safety in rapidly evolving smart environments. QC4 Whether the paper provides a proactive approach to reducing loss of life and injuries in emergency situations by providing proper guidance, communication, and monitoring facilities.

Results and Discussions
Each subsection below provides details of each related research question formulated for the present SLR research work.

RQ1-What are State-of-the-Art Approaches Proposed for Development of Smart Buildings?
Responses to this question were used to synthesize the literature from 2016 to 2020 in which the approaches for developing smart building architectures were discussed. Tab. 8 outlines the different approaches proposed for smart building architectures.

RQ2-What Hardware Components are Used for Early Alarm Mechanism in Emergency Situations?
Smart homes or smart living environments are developed to facilitate living with high standards and resiliency in normal and emergency situations. Tab. 9 lists multiple hardware devices or applications that are proposed for early alarm purposes during emergency situations.

RQ3-What Proactive Approaches are Developed to Reduce Loss of Life or Injuries in Smart Buildings During Emergency Situations?
Smart living environments have been developed to ensure human healthcare and security, but concerns or emergency situations exist that must be addressed, such as short circuits, other fire issues, and earthquakes, that may adversely impact smart living environments. The present SLR study uncovered a number of approaches, which are listed in Tab. 10.  [109] presents a hierarchical and combinatorial mechanism used to model and evaluate smart living system reliability. In particular, the proposed system encapsulates a multi-variant decision-diagram-based mechanism to address standby sparing, phased-mission, and functional dependence behaviors in the physical layer. The combinatorial strategies are developed using the total probability theorem.

4.
Smart home architecture These research articles [11] outline the challenges faced by researchers in the available designs and architectures of smart living environments. Based on the challenges, it presents an optimal solution for smart homes.

5.
Smart home using embedded technology These research articles [81,102] = propose a smart home architecture using embedded technology to ensure safety and basic needs of relevant objects. 6.
Ambient-assisted living (AAL) smart environments This research article [104] proposes a new mechanism named "ambient assisted living" smart environments for the South Korean people. As it is believed that South Korea is anticipated to be a super-aged society by the end of 2025, persistent steps and efforts are being made to reduce the burden of maintaining and promoting a productive and healthy lifestyle for elderly people. 7.
Conjugated smart home environment This paper [106] presents a conjugated single-chip-based controlling unit for controlling the daily tasks (morning wake-up calls, cooking timings), intrusion detection, and event control. This also ensures the security and safety of residents in smart environments. 8.
Efficient energymanagement-based smart home interface This research work [83] aims to develop an IoT-based smart home security system for real-time health monitoring technologies in a telemedicine architecture. Robot-based integrated smart home This is a robot-integrated smart home [71] developed to provide healthcare facilities for aged persons at a smart home. This is a layered architecture designed to perform guidance tasks, healthcare tasks, and other tasks for elderly people. 8.

Fire detection system
This research work [43] proposes an IoT-based fire-detection system in smart buildings. The system was designed using MQ-135 (CO2), MQ-7 (CO), MQ-2 (smog), and DHT-11 (temperature) sensors that were integrated with an Arduino board for accurately sourcing information in fire events. This book chapter [113] addresses common healthcare issues and suggests a wireless-sensor-based network architecture to monitor the air quality for carbon dioxide in complex indoor smart living environments. The research articles [29,35] proposed a wireless sensor network for elderly healthcare purposes in smart living environments.

2.
Smart bed This paper [108] discusses the test bed using real-time datasets, which is aimed at observing system efficiency through different parameters, e.g., time consumption, response times for attack detection, and storage requirements of the proposed approach.

3.
Remote health monitoring of triage and priority system This research work [83] develops an IoT-based smart living security mechanism for telemedicine systems intended to support real-time health monitoring.

Observing and restricting wireless jamming attacks
This study [84] characterizes IEEE 802.11-enabled wireless network efficiency by proposing a safe and secure model for smart homes related to jamming attacks.

Smart dust surveillance
This chapter [98] introduces multiple smart dust mesh perspectives based on the Internet of Everything and Everywhere (IoEE). It has numerous applications in the field of military and security, such as people and product monitoring, eHealth monitoring, and environment surveillance. 6.
Nexus services In this research work [89], researchers propose a nexus model for smart cities that focuses on collaboration and teamwork services.

7.
RiSH This is a robot-integrated smart home [71] developed to provide healthcare support for aged persons. This is a layered architecture aimed at supporting different guidance and healthcare functions for elderly people. 8.
Heterogeneous mechanism This study [72] proposes a heterogeneous approach for evidence identification in IoT networks. A case study is performed to validate the proposed solution. 9.
Business model in European standard Furszyfer Del Rio et al. [74] proposed a business model for European smart homes. They critically reviewed the available literature with a view to identifying and addressing common problems of security, emergency situations, and healthcare in their new business model. 10.
Acoustic surveillance system This research work [19] explores machine learning and its application to acoustic surveillance of abnormal situations. The proposed system's main objective is to help an authorized person take appropriate actions to prevent life/property loss before or during an emergency.

Limitations
The limitations of the proposed research work are the following.
• Only five digital libraries were selected for SLR. However, these libraries provide broader coverage for peer-reviewed and high-quality research articles. • A specific range of years (2016-2020 (a portion of 2020 was included in the review process) was selected for SLR. This was done to ensure the inclusion of recent trends and studies relevant to the current context of smart homes. This is also important to set the further research options based on the recent work reported in the present study. • Google Scholar is another informal source of literature and was thus omitted from the present SLR since the focus was on the systematic selection of only peer-reviewed articles. • Papers that contain the word "smart home" in their title were excluded, but their contents were not meaningful or relevant to the research questions posed. In other words, papers only containing definitions were omitted from the present study.

Conclusions and Future Work
During the past decade, IoT devices have provided state-of-the-art and smart applications for humans. These applications range from smart urban management to smart transportation management devices, smart healthcare devices, smart electrical and home devices, and many others. One of the most inspiring applications is smart homes that aim to support contemporary human living needs. Among the major concerns and challenges of researchers are the security and safety of smart homes. Embedding security in IoT-based applications has been identified as an opportunity to realize the vision of smart and energy-efficient homes and buildings. To address this problem, a SLR was performed to investigate the available literature published in the period 2016-2020 (a portion of 2020 was included in the systematic mapping). This was achieved using the SLR guidelines provided by Kitchenham [16,17].
Five different peer-reviewed online digital libraries were used for sourcing primary research articles. A total of 99 relevant articles (journal articles, book sections, conference proceedings, and survey papers) were identified for analysis and assessment purposes. This systematic synthesis and analysis of the existing research work will serve as a knowledgebase for researchers and designers interested in designing safe and secure smart homes of the future. Furthermore, the results from this SLR will inform the community about recent research trends in smart homes, which are important for formulating future research options.