Prospect for Designing Wearable Devices for Children’s Asthma Attack Detection Based on Internet of Things and Cloud Computing

Indonesia’s geographical location as an archipelago makes this country has a humid tropical climate. This climate will affect asthmatics. Asthma attacks a patient suddenly, this is very dangerous especially if the patient is a child. Before attacking usually, asthma will react to the body with the characteristics of changes in breathing frequency, pulse frequency, mild, severe, even to the point of stopping breathing. Asthma attacks will be very dangerous for children who are not under the supervision of their parents (being in school for example). It can be imagined if asthma with severe clusters aims. The purpose of this study is to propose the design of wearable-type devices used in children at home to facilitate monitoring of asthma attacks through the internet network.


Introduction
The population of Indonesia, which reaches more than 260,000,000 people, has a diverse ethnic, religious, ethnic, cultural and geographical location that is characteristic [1]. That much population does not have national data related to asthma in childhood with review of emergency visits, hospitalization, and death in a hospital. Then the International Study of Asthma and Allergies in Childhood (ISAAC) conducted a study on determining the prevalence of asthma in children in several provincial capitals, see Table 1 [2].
Disruption of the respiratory system, asthma sufferers will feel shortness of breath and coughing this is due to lack of air. Asthma attacks can occur at night or early morning [3 Globally, Indonesia is ranked 20th for asthma-related deaths. About one in 22 people suffer from asthma. However, only 54% were diagnosed with 30% of cases well controlled [4]. The results of the Basic Health Research (Riskesdas) in 2018 showed an average result of the prevalence of asthma in Indonesia of 2.4%. The number of patients with asthma in outpatients aged <4th year as much as 15.42% [5] .
The National Child Asthma Guidelines issued by the Indonesian Pediatrician Association in 2015 classified asthma attacks into three: Moderate mild asthma attacks, severe asthma attacks, and Asthma attacks with the threat of stopping breathing, see Table 2 [6]. Based on Table 2 Asthma attack symptoms can be detected using sensors known as body sensor networks (BSN). In BSN there are several sensors such as heart rate sensor, finger sensor, oximeter sensor, and peak expiratory flow meter [7] [8]. The value of Peak Expiratory Flow can indicate a person's health. Using the MPX 5100 sensor mounted on a pipe orifice plate can measure the air pressure of a person's breath so that asthma attacks can be detected [9]. However, this research does not include wearable-type devices. Photoplethymography (PPG) sensor that can measure a person's pulse and oxygen saturation level. Determination of the location of PPG sensors greatly affects the accuracy of the placement of sensors with wearable-type must be precise [10] . There are also many similar studies that measure the pulse using a heart rate sensor and oxygen saturation level using an oximeter whose data is sent to the monitor system via a 2.4 GHz frequency. This system is ideally installed in a hospital or clinic because it cannot be categorized as a wearable-type device. Therefore in this paper we propose the design of wearable-type devices used in children at home to facilitate monitoring of asthma attacks through the internet network.

Related Works
Internet of Things (IoT) is a new paradigm in communication networks, the data generated is processed in various ways including using certain algorithms [11] [12] . Cisco Systems proposes a model called the Internet of Things Reference Model with seven layers: Physical Devices & Controllers, Connectivity, Edge (Fog) Computing, Data Accumulation, Data Abstraction, Application, and Collaboration & Processes, see Figure 1 [12] .
Other research has also discussed the IoT e-Health Framework for monitoring asthma attacks consisting of three layers: patient, fog computing, and cloud computing, see Figure Figure 4. At the patient layer and around attached several sensors related to health, fog computing processes data from sensors in patients to help diagnose signs of disease, and cloud computing is a layer for storing data and collaborative analysis with various parties including the family [13] .

Body Sensor Network (BSN) or Body Area Network (BAN) is a new generation of Wireless
Sensor Network (WSN) installed on the human body [7] . BSN consists of biomedical sensor nodes that are small in size which aim to continuously monitor vital signs of the human body. BSN applications are very broad, ranging from applications in the military, entertainment, sports to health. The IEEE 802.15.6 standard categorizes BSN applications into two broad categories namely medical and non-medical applications (consumer electronics) [8]. The main characteristic of all BSN applications is that they are built to improve the quality of human life. Physically, the form of BSN implementation in the medical / health world is divided into three types of sensor nodes as follows [14] : 1) Implant node: Is a sensor node that is installed in the human body, where the installation is usually under the surface of the skin. 2) Body surface node: It is a sensor node that is mounted on the surface of the skin. This type is the type of node that will be used in this proposed research. 3) External node: It is a sensor node that does not come into direct contact with the skin / human body. Photoplethysmography is a method for detecting changes in blood volume in capillaries using the property of reflection and light absorption and produces biomedical information that can be used such as heart rate and oxygen saturation [10] . There are many types of Photoplethysmography sensors that are easy to use such as clock and ring sensors. While the Piezoelectric Sensor has been used widely in the world as a system for monitoring the vital signs of the body [15] . One of the functions of Piezoelectric sensors is to be able to monitor the respiratory pattern in particular respiration rate [16] . Prediction and probability are approaches that are very often used in the KDD (Knowledge Discovery Databse) method which aims to identify something that has not happened before. The approach taken to monitor historical data on asthma patients in children can increase the vigilance of patients' parents towards their children. A general example of an asthma monitoring approach is discussed comprehensively in an article conducted by Sajeeda, et al about the extraction of detection and prevention of asthma attacks [17]. Another study conducted by Raji et al. Presented a monitoring system of respiratory rate with IoT so that it was able to carry out periodic monitoring of patients [18]. The role of IoT is also used to detect asthma attacks in real-time applications by notifying them when behavior and pattern changes occur outside normal conditions [19].

Methods and Approaches
In this paper we adopt a trust classification consisting of several metrics: trust metrics, trust sources, trust algorithms, trust architecture, and trust propagation, see Figure Error! Reference source not found. [11] . However, in this study the focus is on trust source and trust architecture for handling detection of asthma attacks in children at school or home. 1) Trust Source, this research is still open regarding the model of trust both direct, indirect or hybrid sources. A node from IoT to detect asthma attacks can use any trust source model. 2) Trust Architecture, the detection of asthma attacks the trust architecture model adjusts to the regulations of each region related to the distribution of health information. However, research recommends using a cloud-based model with the aim of several interested parties being able to read the results of asthma attack detection. Of course this kind of model needs multilevel authority. For example, interested parties related to asthma attacks in children are parents, doctors, hospitals / clinics, and the government. In Indonesia is the Indonesian Ministry of Health or Kementerian Kesehatan Republik Indonesia.

Result and Discussion
Based on the above theoretical considerations at this initial stage, we map papers that have been published until the latest developments. The basic approaches we use include the IoT Reference Model, Architecture of IoT e-Health, and the Model of Trust Classification in IoT. This is stated in TABLE III. about research maps and publications related to IoT in asthma attacks.  Table 3 then we can summarize that detection devices for asthma attacks in children require a special tool to monitor asthma attacks that can provide immediate action. The criteria for the delivery of these devices must be in the form of wearables that are good in terms of design and comfortable to use for children. The device consists of a sensor that measures the pulse frequency, oxygen saturation level, and PEF Index that is connected to a direct gateway connected to the internet. This device should be in a special room or a house so that the quality of the internet network is maintained. In addition to using sensors attached to the body also need other sensors installed or using Social-IoT such as weather, air quality, temperature and air pressure. Data processing can use cloud computing technology [31] with limited authority. Users of this information include parents and medical teams that have been appointed.
So we propose the architecture of detection of asthma attacks in children that are connected with cloud computing technology that makes it easy for parents to monitor the development of a child's illness and conditions in the home and the medical team that has been assigned can provide fast and appropriate action when an asthma attack occurs in a child. Devices should be designed with the Wearable type for children's convenience that is connected directly to the IoT e-Health cloud computing system, see Figure 4.

Conclusion
IoT e-health care to anticipate sudden asthma attacks in children is designed with the type of wearable body surface node must be designed with an emphasis on the comfort of fiber does not interfere with children's activities. The IoT e-health asthma attack detection device that we designed senses the pulse frequency, oxygen, and PEF, and biases the sensor in children's play environment, especially in the weather, temperature changes, and dust density. The device and Social IoT are directly connected to IoT e-Health Cloud-based via an internet gateway. So parents and the medical team can monitor easily.