A Smart System for sleep monitoring by integrating IoT with big data analytics
ABSTRACT
Obtrusive sleep apnea (OSA) is one of the most important sleep disorders because of it has direct adverse impact on the quality of life. Intellectual deterioration, decreased psychomotor performance, behaviour and personality disorders are some of the consequences of OSA. Therefore, a real-time monitoring of this disorder is a critical need in healthcare solutions. There are several systems for OSA detection. Nevertheless, despite their promising results, these systems not guiding their treatment. For these reasons, this research presents an innovative system for both to detect and support of treatment of OSA of elderly people by monitoring of multiple factors such as sleep environment, sleep status, physical activities, and physiological parameters as well as the use of open data available in smart cities. Our system architecture performs two types of processing. On the one hand, a pre-processing based on rules that enables the sending of real-time notifications to responsible for the care of elderly, in the event of an emergency situation. This pre-processing is essentially based on a Fog Computing approach implemented in a smart device operating at the edge of the network that additionally offers advanced interoperability services: technical, syntactic, and semantic. On the other hand, a batch data processing that enables a descriptive analysis that statistically details the behaviour of the data and a predictive analysis for the development of services, such as predicting the least polluted place to perform outdoor activities. This processing uses Big Data tools on Cloud Computing. The performed experiments show a 93.3 % of effectivity in the AQI prediction to guide the OSA treatment. The system’s performance has been evaluated in terms of latency. The achieved results clearly demonstrate that the pre-processing of data at the edge of the network improves the efficiency of the system.
EXISTING SYSTEMS:
Human beings experience changes in our bodies and in our lives. One of these changes is the alteration of sleep that occurs with age. In particular, obstructive sleep apnea syndrome (OSA) is one of the mostcommon and dangerous respiratory disorders that occur during sleep. OSA consists of the obstruction or partial blockage of the upper respiratory tract for at least 10 seconds and that prevents proper oxygenation of the blood even over 20-30 times an hour of sleep. According to number of interruptions per hour and by using the apneahypopnea index (AHI), OSA can be classified into 3 categories from higher to lower severity; if these interruptions occur between 5 and 15 times per hour as “mild”, if these interruptions occur between 15 and 30 times per hour as “moderate”, and if these respiratory interruptions occur more than 30 times per hour as “severe” .
The sleep fragmentation produced by these respiratory interruptions have clinical consequences associated including depressive disorders, irritability, intellectual deterioration, decreased psychomotor performance, behaviour and personality disorders [3]. As a result, the quality of life (QoL) can be significantly decreased and thereby increase the associated health problems and medical costs. Sleep apnea affects 21% of the US population (~70 million people), and account for an estimated $16 billion in health care expenses each year. OSA at any age is a major concern due to the health problems it can cause, but it is even more problematic for older people who are more likely to have respiratory problems at night, but are less likely to be diagnosed; (more than ~ 80–90%) are not diagnosed or simply diagnosed as snoring. Research reveals that between 13 and 32% of people over 65 suffer from OSA and it is a growing problem in developed countries, which are average life expectancy has increased. The difficulty or trouble falling asleep, combined with the lack of deep sleep, results in a poor QoL and a greater health risk for elderly people. In addition, the long-term implications of chronic sleep disorders include an association with an increased risk of death[6]. Nowadays, the main reference tool for the diagnosis of OSA is a conventional polysomnography (PSG) study. PSG is an examination that lasts all night in a specialized clinic or in hospitals under constant medical surveillance, which means that the patients must go to a medical facility frequently, which will inevitably increase the burden on hospitals. Additionally, this method incorporates many sensors on a person’s body, which is considered intrusive
and, in turn, can disturb sleep. In addition, the high cost of PSG makes it a very impractical monitoring method to be implemented in the long term. On the other hand, there are many approaches available for OSA treatment, including weight loss, sleep hygiene techniques, positional and continuous open airway therapy (COAT), continuous positive airway pressure (CPAP) and surgical interventions. CPAP is an effective treatment for OSA [8], nevertheless adherence to treatment is suboptimal because to low perceived disease risk by the patients. This in turn may bring discomfort to the patients and lead them to interrupt therapy. Therefore, there is a significant need to provide an unobtrusive real-time- systems that not only allows for the detection of the OSA but also supports its treatment at home. Several researches have proposed a variety of systems to detect OSA episodes, generally using wearable sensors incorporated in smart devices such as bands, bracelets, watches, and telephones. These systems achieve near realtime detecting of the OSA based on especially monitoring of physiological parameters such as the respiratory rate, the heart rate, and the oxygen saturation through wireless technologies such as Bluetooth, Wi-Fi, and ZigBee with promising results. Nevertheless, as far as we know, these systems do not support the long-term treatment of this syndrome. In addition, almost all existing systems cannot work without a smartphone, which is used as a receiver and a processor of the data.
Therefore, such proposals are unsuitable to monitor OSA patients athome, since the complex tasks of data processing can have a great impact on the daily use of the smartphone and thus in operating the system. The effects of sleep apnea and its complications have heightened the urgency for the patient to have not only a rapid diagnosis but also treatment.
PRPOSED SYSTEM:
In this work, an architecture of an OSA monitoring system based on the Internet of Things (IoT) and Big Data is proposed. The three-layered architecture integrates Fog and Cloud Computing capabilities to support both diagnosis and treatment of sleep apnea by creating of various services including remote monitoring, real-time alert notifications, data analysis and information visualization. The proposed system envisions assisting health professionals in medical decision-making.
The system performs the monitoring of the OSA harnessing advantage of the combined use of different technologies, components, and complementary open standards such as 6LoWPAN, ZigBee, BLE, Smart IoT Gateway, FIWARE and lightweight and secure IoT protocols such as MQTT and CoAP. The data related to the physical activities of the elderly, the sleep environment, the sleep status, the physiological parameters, and the context information collected are transmitted directly to a Smart IoT Gateway operating at the Fog Computing level using different Low Power Wireless Networks such as Bluetooth 4.0, ZigBee, and 6LowPAN. The different types of interoperability provided (technical, syntactic, and semantic) by the Smart IoT Gateway allows seamless interoperability between the different networks and communication protocols used. In order to guarantee an immediate response from the system in emergency situations with low latency, the preprocessing of the data independently is also implemented in the Fog Computing level. The pre-processed data is stored and made available to users at the Cloud Computing level through a generic enabler which can greatly improve the administration and availability of the data.
Additionally, at this level, an analyzer based on Big Data is implemented to support the processing of data by extracting and analyzing the data coming from both the Fog Computing level and the open data catalog available in smart cities, in particular, from the city of Valencia. To do so, the big data architecture is based on Lambda architecture because it provides a fault-tolerance, scalable and reliable system. Using the batch layer described on the lambda architecture, the historical data is stored in a Hadoop Distributed File System (HDFS) cluster and exploited by using the Apache Spark platform. Despite lambda architecture proposes a speed layer to real-time processing, this is not considered in the Big Data architecture proposed because the real-time processing is implemented in the fog layer.
In the same level, a web-based graphical user interface (GUI) is also implemented that enables health professionals, caregivers, and emergency medical centers to remotely access the data on elders’ physical activity, sleep stage, sleep
environment, and medical condition in order to assess whether treatment needs to be changed, as well as to contact them when in need. The proposed system has been successfully implemented and its real feasibility in the monitoring and treatment of OSA has been fully tested.
CONCLUSION :
QoL has become a need in society that will continue to be even more important if we consider that in the future the number of older adults will represent more than 14% of the world’s population. OSA is one of the diseases that most compromises the QoL of the adults who suffer from it and causes important complications that can affect their health.
The performance evaluation of the proposed system is evaluated in terms of latency. The results demonstrate that the detection of apnea episodes making at the Fog layer reduces the latency on the communication infrastructure compared to at the Cloud layer. In the future, we will focus on integrating the system with other solutions applied to the healthcare domain derived from Inter-IoT project with the aim of facilitate the delivery of elderly smart healthcare services. In addition, a future evaluation with more patients will provide both useful lessons learned and results to be used to further enhance the proposed system.