• Không có kết quả nào được tìm thấy

Short Contributions: IoT and AI Solutions for E-Health

N/A
N/A
Nguyễn Gia Hào

Academic year: 2023

Chia sẻ "Short Contributions: IoT and AI Solutions for E-Health"

Copied!
218
0
0

Loading.... (view fulltext now)

Văn bản

This is exactly what is needed in healthcare: small devices that can be wearable (detached from the body), wearable (on the body) or implantable (in the body) track the status of the patient. Authors in [20] and [10] argue that Web of Thing based middleware is suitable to ensure interoperability, as all devices can be abstracted as a web resource. The ECOHelath middleware proposed in [10] connects physicians and patients to simplify the health monitoring system and provide more accurate patient diagnosis.

Table 1. Taxonomy of middleware for IoHT
Table 1. Taxonomy of middleware for IoHT

Healthcare: Challenges, and Opportunities

1 Introduction

This list also includes smart healthcare, one of the main subjects of this article. The article addresses the phenomenon of uncertainty occurring in large, evolving systems, namely the Internet of Things through the use of the example of smart healthcare.

Fig. 1. Figure showing the Internet of Things application areas
Fig. 1. Figure showing the Internet of Things application areas

2 The IoT in Health Care

Since it has the ability to find the patient's condition and the environment the patient was in, it will greatly help the healthcare providers to understand the variations that can affect the health status of these patients. In addition, the change in the patient's physical condition may increase the percentage of his vulnerability to diseases and be a cause of his/her health deterioration [10].

3 Characteristics of the Phenomenon of Uncertainty

  • Defining Uncertainty
  • Types of Uncertainty in the Field of Healthcare
  • Sources of Uncertainty in IoT Systems
  • Causes of Uncertainty in IoT Systems

11] focused on different IoT-based healthcare systems for Wireless Body Area Network (WBAN) that can enable smart healthcare data reception and data transmission. Interoperability between different devices in different domains is a key limitation for IoT success due to a lack of universal standards.

4 Findings and Recommendations

Research Challenges

Quality of Services (QoS): The quality of service is an important parameter used in healthcare, which is a very time-sensitive system. There are several challenges to meet the quality requirements for IoT-based applications in terms of energy efficiency, data quality sensing, network resource consumption, and latency.

Major Requirements

Uncertainty is one of the key problems for most IoT systems based on RFID (Radio Frequency IDentification) technology. Ambiguity Data (acceptability, inaccuracy): sometimes radio frequencies can cause data to be reflected in reading areas, so RFID readers can read those reflections;.

Briggs, A.H., et al.: Model parameter estimation and uncertainty analysis: Report of the ISPOR-SMDM Good Research Practices Modeling Working Group-6. Our platform offers continuous health monitoring by integrating mechanisms for backing up medical data at the cloud level, thereby ensuring the notion of fault tolerance through data replication and thus the availability of information, while ensuring the integrity and confidentiality of the exchanged data (using sh1 and MD5 protocols for hashing and DES, RSA for encryption), minimal response time and low latency due to the implementation of fog computing.

2 Related Works

3 E-Health Platform

System Architecture

Web server: manages and maintains the system, including management of the (N) area administrator's accounts. All this is done by the super admin, while each admin is responsible for creating and managing the accounts of the users in his or her zone.

The Main Operating Processes

Its role is to easily process the medical data collected by sensors connected to different patients geographically distributed in areas (N), followed by the backup of these data in local databases. Cloud: a central server, which is responsible for the global storage of medical and personal data of users in all areas, with the possibility of transferring these data to an area X to which the patient has moved, if the attending physician wants to consult the file of his new patient who has just arrived in his new area.

Security of the E-Health Platform

4 Implementation and Results

Mobile Application Prototype

As can be seen in fig. 2, the sensor registers each time a new value; it sends it to the mobile application it is connected to. In the event of a heart failure as shown in the following figure, messages and alerts are sent to the practitioners treating the patient and a phone call is made from the patient's home to the toll-free civil protection number.

Simulation and Discussion of Results

While the use of threads will have an impact on the load on the communication network. While using threads, we note that the different requests will be processed simultaneously, which will therefore reduce the response time as well as the load on the network.

5 Conclusion and Perspectives

After analyzing the simulation results, we conclude that as the number of users increases, the number of requests to be processed by the server also increases, which will lead to congestion at the server level and additional traffic on the network, thus the response time extended, which is not tolerable in our system in the event of an extreme emergency. In: 2015 IEEE International Conference on Computing and Information Technology; Ubiquitous Computing and Communications; Reliable, autonomous and secure computing; Pervasive Intelligence and Computing, pp.

Sharing Architecture in Multi-Clouds Environment

In [3], ABE is used to selectively share data with doctors without allowing them to know the exact description of the patient's diseases. HSDSA gives the patient full control over the creation and management of decryption keys without relying on a trusted authority, making it more useful for public cloud environments.

2 Architecture Overview of the Proposed Scheme

Section 2 provides an overview of the overall architecture of the proposed framework and its components. Once the DO ensures that the DR is an authorized requester and that he possesses the encrypted version of the EHR (R), then the DR and DO attempt to establish a session using the Diffie-Hellman (DH) algorithm to exchange decryption keys securely .

3 Analysis of the Proposed Storage Process

The Registration Phase

Then the DO uploads its encrypted record (R) and the hash value of the original record H(R). The reconstruction phase starts when a DO or an authorized user DR wants to get the EHR, he sends a request to the framework.

The Storage Phase

Distribution is done using Shamir's secret sharing algorithm and the resulting shares are sent to cloud server providers CSP1, ..CSPn. When a DO wants to preserve an EHR, it calculates the digital signature of the original EHR(R).

4 Analysis of the Proposed Retrieval Process

The Reconstruction Phase

The Recovery Phase

Then the DO encrypts its private key (PU) with Ks and sends the result value YOU to the DR.

Table 2. Scenario of the recovery phase
Table 2. Scenario of the recovery phase

5 Conclusion

Images or other third-party material in this chapter are covered under the chapter's Creative Commons license, unless otherwise indicated in a credit line for the material. If the material is not covered by the chapter's Creative Commons license and your intended use is not permitted by legal regulations or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

Things: A Technical Review

Section 2 presents a detailed technical analysis of existing articles dealing with blockchain integration with IoT.

2 Internet of Medical Things (IoMT)-Blockchain Challenges

3 Blockchain-Based Approaches in IoMT

  • Ethereum-Based Contributions
  • Modified Consensus Protocol
  • Modified Cryptographic Technique
  • Hyperledger-Based Contributions
  • General Blockchain Concept Without Technical Specifications In [7], the authors took benefit of tamper proof feature of blockchain to

The authors proposed proof of medical stack (PoMS) as an alternative to PoS consensus model to protect smart contracts from malicious actions. In [2], the authors proposed an IoT blockchain-based architecture to enable healthcare remote monitoring.

4 Discussion and Open Issues

Programming abstractions: The integration of blockchain technology in the IoMT opens the way to many relevant applications in the health field. Our research shows that the proposed solutions lack many technical details when integrating Blockchain into the IoMT.

Table 1. Classification of researches applying blockchain in IoMT
Table 1. Classification of researches applying blockchain in IoMT

In INSect.3 we provide some medical use cases in healthcare that use this promising technology.

2 Key Concepts on Blockchain

Overview and Architecture of Blockchain

Among the important features of Blockchain, decentralization by making the ledger accessible by all participants, immutability, so the blockchain is almost impossible to tamper with and is resistant to censorship, availability by offering all peers a copy of the blockchain to take access to the entire time-stamped transaction. data and anonymity, where any user can interact with the blockchain with a generated address, which does not reveal the true identity of the user.

Taxonomy of Blockchain Systems

3 Blockchain Use Cases in Healthcare

  • Electronic Medical Records
  • Remote Patient Monitoring
  • Pharmaceutical Supply Chain
  • Health Insurance Claims

14] Hyperledger Fabric Off-chain A mobile blockchain-based health system for cognitive behavioral therapy for insomnia. 5] Hyperledger Fabric On-chain Design a blockchain-based control system for drug turnover control.

4 Research Challenges and Opportunities

IPFS

3 - User 2 sends his\her ID and ENKpubuser1(Data) to the IPFS. 4 - IPFS sends @IPFS and HIPFS to user1 b) Smart contract execution. 1 - User 1 signs his/her smart contract with his/her private key Kpriv (this smart contract requires @IPFS and HIPFS of the acceptance smart contract) and sends it to the Blockchain.

Figure 4 presents the architecture of the new solution which is composed by two phases:
Figure 4 presents the architecture of the new solution which is composed by two phases:

Monitoring System Using Internet of Things

In [9], the authors propose an ontology-based context management system that enables the monitoring of the behavior of senior citizens and the detection of risks associated with mild cognitive impairment and frailty. Thus, the combination of fuzzy logic theory with ontology is considered a solution to uncertainty.

3 Architecture of Fuzzy Ontology-Based Healthcare System

In [15], an ontology-based personalized food recommendation system is proposed to support travelers with long-term illnesses and following a strict diet. 1,13] propose recommender systems based on fuzzy ontology that efficiently monitor the diabetic patient and recommend suitable foods and drugs.

4 Proposed Fuzzy-Ontology

The state of health is the output variable that is determined based on fuzzy input variables that determine the medical measurements and the fuzzy rules. These variables are considered input variables combined with the fuzzy variables Age, BMI, and physical activity to infer whether the diet is healthy or not, which is expressed by the fuzzy variable Diet Status.

5 Health Condition and Diet Status Calculation Process

The level of physical activity is used to calculate the total calories the diabetic needs to maintain or lose weight. The results show that the time consumed depends on how many inputs the query needs to compute the output.

6 Conclusion

Example: check patient vitals today from 8 hours (start_date) to 19 hours (end_date) every 2 hours (cycle_date). A description of the various system modules is given in the following subsections.

Fig. 1. General architecture of the proposed approach
Fig. 1. General architecture of the proposed approach

2 Registration Methods

First, it allows to align subject B: T2 according to the MRI mode T1 of the same subject. Second, an afne transformation is applied to the moving image using transformation matrices (translation, rotation, scale and shift).

Fig. 1. MRI/X brain image registration: (a) mono-modal MRI/MRI atlas (from left to right: MRI image, MRI atlas and superposed images), (b) multi-modal MRI/PET (from left to right: MRI image, PET image and superposed images).
Fig. 1. MRI/X brain image registration: (a) mono-modal MRI/MRI atlas (from left to right: MRI image, MRI atlas and superposed images), (b) multi-modal MRI/PET (from left to right: MRI image, PET image and superposed images).

3 Materials

To determine the accuracy of the studied methods, we measured the results of the Normalized Cross-Correlation Coefficient (NCCC) (1) and the Normalized Mutual Information (NMI) (2). The greater the value of the normalized mutual information, the more accurate the registration process.

4 Results

In our work, we focus on the classification of scientific papers in the epidemiological domain based on the taxonomy of the epidemiological studies. In the following, we present the performance measures used to assess the performance of the different machine learning models used.

Fig. 4. Examples of MRI/X multimodal registration: (a) MRI image, (b) X image, superposed images using (c) HM, (d) SPM (e) ITK-Snap, and (f) 3D Slicer.
Fig. 4. Examples of MRI/X multimodal registration: (a) MRI image, (b) X image, superposed images using (c) HM, (d) SPM (e) ITK-Snap, and (f) 3D Slicer.

In order to meet health care policies and have good reinforcement regularity and system functionality, the proposed approach relies on an alloy analyzer to automatically verify the requirements. The results showed the effective correctness of the modeled system and the associated reinforcement mechanism.

2 Related Work

The main components of this model are Emergency Response Services, Coordination Unit, MCI, Command and Control Center and Agent Based Simulation. Catarinuccite ​​al.[1] propose a “Smart Hospital System” (SHS) for automatic monitoring and tracking of patients, staff and biomedical devices.

3 HMS Modeling

13] designs an agent-based model that predicts the response time of emergency services by taking into account the characteristics of road segments and the driving behavior of emergency service drivers. Finally, the resulting coordination unit plan will be sent to the agent-based simulation, which was used to simulate emergency response tasks in real environments, and identify the best coordination mechanism plan to achieve the best response time.

4 HMS Validation

For the effectiveness of the proposed framework, we show the correctness of the proposed model through checking and simulation. With regard to the achieved results, we ensure the correctness of the proposed model.

Fig. 2. Class diagram for HMS.
Fig. 2. Class diagram for HMS.

Inertial Sensor Based Human Activity Recognition

In recent years, several machine learning and deep learning algorithms for the recognition of human activities have been proposed. Thus, a lack of self-contained, sensor-based HAR systems with built-in machine learning and real-time response is noted.

2 Artificial Neural Networks/Feed Forward Neural Networks

3 Long Short Term Memory-RNN/Back Propagation Neural Networks

Compute the cell's intermediate state by passing our input and the previous state through the tanh activation. Perform elementwise multiplication and compute the forgotten port and multiply it by the old state of C0.

4 Experimental Results

Activity Database Collection for HAR

Compute the input gate by passing the previous state and the input through a sigmoid activation. We add these to get the new cell state which is C1 and calculate the output gate which is multiplied by the cell state going through tanh activation.

ANN Evaluation

The BlueNRG-Tile's accelerometer and gyroscope located on the right leg are used to capture data. The data has a total of 219600 data samples and is divided into 80% for training, 10% for testing and 10% for validation part.

LSTM-RNN Evaluation

5 Real-Time HAR Implementation in Raspberry PI

United Nations, World Population Prospects 2019, in Highlights, p. Chen, F., Fang, C.: The functional impairment of family support for the elderly and the outlet:. the research on the elderly support model in the underdeveloped rural areas. Nasr, N., et al.: The experience of living with stroke and using technology: opportunities to engage and co-design with end users.

Fig. 1. Hybrid WSVM-HMM system based PCA approach.
Fig. 1. Hybrid WSVM-HMM system based PCA approach.

Workflow Model for Ambient Environment

In this paper, we first describe the workflow using Ag-LOTOS [18], a formal specification model based on LOTOS. Then, the contextual workflow scheduling system (CPSw) [19] is built based on Ag-LOTOS semantics constrained by contextual information.

2 The Context-Aware Workflow Model

Ag-LOTOS for Workflows

But [8,9] allows users to model their daily activities in terms of workflow that can be adapted to context information. In A = failed, fail represents the fact that the execution of an activity A fails due to the dynamic context of the workflow.

Contextual Planning System of the Workflow

During the execution of the activity, it is possible to indicate its failure with the deactivation operator [>.A[> B means activityA can be deactivated by activityB which interrupts the main flow and uses stop instead of exit. If L = ∅, the lack of synchronization leads to the absence of interaction points between processes, this is achieved through the connection operator.

3 Case Study

If the system detects a problem in the switch at the location of the data center, it sends a request to the nearest help desk.

Fig. 1. The scenario that illustrate the case study.
Fig. 1. The scenario that illustrate the case study.

4 Conclusion

Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution, and reproduction in any medium or form. , provided you give proper credit to the original author(s) and source, provide a link to the Creative Commons license, and indicate whether changes have been made. Despite its silent effects, which are sometimes hidden from the larger audience, air pollution is becoming one of the most powerful threats to global health.

Hình ảnh

Fig. 1. Architecture overview
Table 2. Scenario of the recovery phase
Table 1. Classification of researches applying blockchain in IoMT
Table 1 contains all the security parameters of the communicating parties in e- e-health scenario.
+7

Tài liệu tham khảo

Tài liệu liên quan

Do đó, với tư cách là những người làm công tác giảng dạy lý luận chính trị trong các trường đại học, cao đẳng, nhóm tác giả nghĩ rằng, việc quán triệt những điểm