Application of Multi-sensor Network and Artificial Intelligence in Health Monitoring of Medical Geriatric Care

doi:10.21203/rs.3.rs-2760872/v1

Download PDF

Research Article

Application of Multi-sensor Network and Artificial Intelligence in Health Monitoring of Medical Geriatric Care

https://doi.org/10.21203/rs.3.rs-2760872/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 05 Jun, 2023

Read the published version in Soft Computing →

You are reading this latest preprint version

The comprehensive use of data from multiple sensors for moving target tracking is a long-term problem in multi-sensor data fusion. In wireless sensor networks, state estimation and fusion are the main research issues in target tracking. Distributed is the essential feature of wireless sensor networks. Therefore, with the development of wireless sensor networks, the problem of distributed state estimation has also attracted the attention of scholars. The intelligent medical diagnosis system is composed of a traditional Chinese medicine diagnosis system, a western medicine diagnosis system, and a medical record database. The TCM diagnostic system and the Western medicine diagnostic system are the main body of the system. The TCM diagnostic system is based on a case-based reasoning model, using human body information collection equipment to simulate the process of TCM diagnostics and realize TCM diagnostic engineering, while the Western medicine diagnostic system is a neural network optimized based on genetic algorithms. Model, obtain the diagnosis and treatment method of the disease from the medical record database of the hospital information system. The basic standards for nursing homes issued by the Ministry of Health have promoted the standardized development of nursing homes. However, since the development of China's elderly care homes is still in its infancy, it makes the service content and functional positioning of elderly care homes, and orderly connection with relevant medical institutions and elderly care institutions, The main body of institutional management and the formulation of supporting policies are still in the exploratory stage. Among the chronically ill population, the proportion of elderly people ranks first. At the same time, due to the particularity of the elderly population, the recovery time is lengthened. Due to various reasons, they cannot get adequate medical services in the hospital. Due to the impact of diseases, the health of the elderly is not optimistic, and even loses their lives. Therefore, the daily immediate monitoring of the elderly is very necessary.

Multi-sensor network

Artificial intelligence

Elderly medical treatment

Health monitoring

The Kalman filter algorithm is a classic method of state estimation in linear systems. The average consistency strategy is an effective method for network-wide distributed computing tasks. In recent years, scholars have combined the Kalman filter and the consensus protocol to propose the Kalman consensus filter to improve the estimation performance by improving the consistency of the estimation of different nodes, and due to the simple distributed structure, strong scalability and robustness of the algorithm So it has attracted widespread attention [1]. According to the characteristics of Chinese medicine lacking machine inspection methods, the article adopts the method of human body information collection to simulate the operation of Chinese medicine pulse number, and realizes the inspection in the way of Chinese medicine, and the inspection results are quantitatively saved in the Chinese medicine case database [2]. At the same time, based on the characteristics of Chinese medicine, a case-based reasoning model is used to establish an expert system for Chinese medicine diagnosis. Use the growing case base as the knowledge base for reasoning and realize the diagnosis function of Chinese medicine [3]. Therefore, this paper designs a diagnosis model based on neural network, makes full use of the existing hospital diagnosis system, and uses the hospital's existing case database as network training samples to realize the diagnosis function of western medicine [4]. It has become an important issue that needs to be solved urgently, and it has also been a hotspot and focus of research in various circles in recent years [5]. This article is oriented to the health monitoring of the elderly, and has developed an intelligent monitoring terminal with no sense of violation, including a wearable vital sign monitoring terminal and a distributed plantar pressure collection terminal, which realizes the vital sign data of the elderly in free activities [6]. And gait information collection and analysis. The vital signs monitoring terminal adopts STM32 single-chip microcomputer as the microcontroller, and the ECG acquisition circuit, respiratory wave acquisition circuit, photoelectric sensor temperature acquisition circuit and wireless communication circuit, etc., construct a portable and easy-to-operate monitoring terminal [7]. The distributed plantar pressure acquisition terminal adopts a cross-switching method. Through the cooperation of software and hardware, the distributed plantar pressure data is transmitted to the PC via the wireless network for display and analysis [8].

The use of data from multiple sensors for moving target tracking is a long-term problem in multi-sensor data fusion. In military applications such as target tracking or tactical intelligence, it is mainly to estimate and predict the state of specific types of entities in the external environment [9]. The literature says that the traditional method is to gather the observation data or local estimation information of each sensor node to the fusion center, and the fusion center will perform the data fusion processing to obtain the global state estimation [10]. At the same time, remote hospitals are still short of professional medical staff [11]. This also leads to a high rate of misdiagnosis in these hospitals, especially for some intractable diseases. In addition, in some small hospitals with low medical level and lack of conditions, misdiagnosis and delay of serious diseases often occur, which may endanger the lives of patients. Therefore, the intelligent diagnosis system can avoid this situation to a certain extent [12]. At present, with the continuous development and progress of computer technology, intelligent medical diagnosis system can be applied in the application of hospital system.The literature is limited by the decline in the physical fitness of the elderly, more likely to get sick, and the rehabilitation process is lengthened, the society's attention to elderly health monitoring has increased significantly. The number of elderly patients with chronic diseases shows an increasing trend every year. Various reasons lead to the inability to receive adequate medical services in the hospital. Affected by the disease, the health of the population is not optimistic and even loses their lives. Therefore, it is particularly important to carry out immediate monitoring around the clock [13]. The literature points out that the health of the elderly embodies the characteristics of "high incidence rate, diverse disease types, long course of disease, high complications, and difficulty in treatment", which makes the elderly whose body functions degenerate have low self-care ability and high dependence [14]. There is a strong demand for nursing services; but families and elderly care institutions cannot bear the burden of elderly medical care. The literature mentions that China has unique characteristics of aging, such as getting older before getting rich, larger scale, faster speed, and heavier support burden. In addition, China's old-age security system needs to be further improved [15]. Therefore, under the severe challenge of population aging, Chinese society is facing tremendous pressure on the elderly.

3.1. Multi-sensor network algorithm

Using multi-sensor data to estimate the state of a target is one of the main tasks of data fusion. Kalman filter method is a classic method in state estimation. Standard Kalman filter and information Kalman filter are two forms of Kalman filter. This chapter will give two specific derivation and proof procedures. In addition, in multi-sensor estimation and target tracking, a central fusion processing structure is often used. A more natural approach is to extend the single-sensor KF method to multi-sensor problems. This chapter will give examples of group sensor methods and inverse covariance .

For dynamic state estimation, discrete time methods are widely used. In order to analyze and infer dynamic systems, at least

Description of the initial state: Assume that the initial state X(0) is Gaussian, with mean X(0|0) and covariance P(0|0) and

$$E\left\{\left[\begin{array}{c}w\left(k\right)\\ v\left(k\right)\\ \widehat{X}\left(0|0\right)\end{array}\right]\left[\begin{array}{ccc}w{\left(l\right)}^{T}& v{\left(l\right)}^{T}& \tilde{X}{\left(0|0\right)}^{T}\end{array}\right]\right\}=\left[\begin{array}{ccc}Q\left(k\right){\delta }_{kl}& 0& 0\\ 0& R\left(k\right){\delta }_{kl}& 0\\ 0& 0& P\left(0|0\right)\end{array}\right]$$

Where

$$\tilde{X}\left(0|0\right)=X\left(0\right)-\widehat{X}\left(0|0\right)$$

The physical process of a moving object can be divided into two parts: 1) the part that can be accurately predicted according to the equation of motion; 2) a random process with a mean value of zero and a known distribution. The Kalman filter is based on this theory to establish a motion model. According to this mathematical description, the position of a moving object at any time can be estimated.

Kalman filtering includes a prediction phase and an update phase. In the prediction phase, the optimal state estimate at the previous moment is used to predict the current state value according to the system state model, and the observation value at the current moment is predicted according to the observation model; in the update phase, Use the state value measured at the current moment to optimize the state prediction value obtained in the prediction stage, so as to obtain the final estimated value at the current moment.

In multi-sensor integrated tracking, the inverse covariance is often used to express the state estimation covariance update equation, which has a recursive formula:

$${P}^{-1}\left(k+1|k+1\right)={P}^{-1}\left(k+1|k\right)+{H}^{T}{R}^{-1}\left(k+1\right)H$$

The filter gain formula has another form of expression:

$$K\left(k+1\right)=P\left(k+1|k+1\right){H}^{T}{+H}^{T}{R}^{-1}\left(k+1\right)$$

The Kalman filter of a discrete system is a linear unbiased recursive filter. Linear means that the output of the filter is a linear function of the observed value. Unbiased means that the state estimate is equal to the mean value of the actual state. Recursive means this estimate The value can be obtained by revising the previous estimated value by using the new observation value. Therefore, the Kalman filter has properties such as linearity, unbiasedness and recurrence.

3.2. Application of artificial intelligence technology in medical systems

Modern medicine is divided into categories, resulting in extremely complicated procedures for medical treatment in hospitals. The diagnosis and treatment process of a hospital is shown in Fig. 1.

The diagnosis and treatment process of Chinese medicine is relatively simple. The top three hospitals have TCM departments, but the TCM departments are smaller in scale, and TCM uses fewer diagnosis and treatment methods. It can be said that Chinese medicine is completely different from Western medicine in terms of diagnosis and treatment, to diagnosis and treatment methods, and then to treatment process. Excellent TCM doctors can prescribe the right medicine to cure the disease by just a few simple diagnostic methods such as inquiries, tongue coating, and pulse identification. The design of a medical diagnosis system must fully consider the difference between Chinese and Western medicine, otherwise it will result in a four-different system of "painting a tiger is not an anti-dog".

The workflow of the medical diagnosis system is shown in Fig. 2.

If Western medicine is selected for diagnosis and treatment, the system will list a checklist based on the physical information selected by the patient, and the patient’s check result will be transmitted to the diagnosis system through the hospital’s existing information management system. According to the checklist, the diagnostic system calls the western medicine diagnostic subsystem based on neural networks to give a treatment plan. The treatment plan calls the western medicine medication system to synthesize the current development of western medicine and the clinical manifestations and curative effects of each medicine, and give medication recommendations, which are the same as the diagnosis of traditional Chinese medicine. The system is similar. The western medicine diagnostic subsystem also enters a cycle of inspection and treatment until the patient is discharged from the hospital and the medical record is printed.

4.1. Medical elderly care model

As a medical institution with special needs in response to the "silver wave", nursing homes for the elderly are still in the early stages of development. They are far less than other large medical institutions in terms of scale and treatment. 44.5% of medical staff are difficult to find in large hospitals, so they choose qualification experience, etc. For elderly nursing homes where the requirements are not particularly strict, 29.7% of the investigators regard elderly nursing homes as a springboard for career development, as shown in Table 1.

Table 1

Reasons why medical staff choose to work in elderly nursing homes
Reason for selection	Number of people	Percentage of surveyed population (%)
Difficult to find jobs in large hospitals	78	44.5
Elderly nursing homes are a fast-growing emerging industry	64	36.5
match of major	52	29.7
Use your own experience and waste heat	32	18.2
Use aged care homes as a springboard for career development	52	29.7
Other	6	3.4

Nursing work has the characteristics of high pressure, high labor intensity, and high risk responsibility due to its own particularities. 49.1% of the interviewees expressed a lot of mental and work pressure during work. 24.6% of the interviewees said that they are under a lot of pressure from public opinion at work, and they are unwilling to mention their work to others. In the interviews, some medical staff even said that their current work is rejected and incomprehensible by their family members. Despite working under such a heavy load, medical staff in elderly nursing homes still have problems such as unequal pay and income, limited development space, and lack of professional skills training, as shown in Table 2.

Table 2

Causes of dissatisfaction with the work of medical staff in elderly nursing homes
Reason for selection	Number of people	Percentage of surveyed population (%)
Long working hours and heavy workload	83	47.4
High job risk	73	41.7
Not respected by society	106	60.5
Poor working environment	41	23.4
Expenses are not proportional to income	89	50.9
No room for development	57	32.6
The working atmosphere is deserted	23	13.1
Other	4	2.3

Data source: According to the statistics of 175 medical staff questionnaires that have been effectively recovered.

Government departments repeat assessments and waste management resources: The health bureau, civil affairs bureau, social security bureau, and price bureau still have overlapping functions and overlapping functions in actual operation, which leads to the problem of overlapping inspection content, resulting in the human, financial and material resources of various management departments And other waste.

Relevant rules and regulations are not sound, and institutional operations are not standardized: the lack of admission standards; the current relevant government departments have not promulgated uniform discharge standards; the lack of transfer standards at this stage.

Insufficient economic compensation and slow institutional development: due to policy regulations that are too broad and general, and the lack of coordination among various departments, lack of a unified management department responsible for policy implementation and supervision and management; some elderly care home services are not included in the scope of medical insurance reimbursement; elderly groups The huge demand for medical care services has prompted social capital to actively invest in new aged care homes. However, due to the nature of public products in the medical care services provided by aged care homes, the investment in aged care homes has a high investment amount, a long payback period, and benefits. Slow characteristics.The lack of a nursing insurance system makes it difficult to protect the health rights and interests of the elderly: the differences in the economic abilities of individuals and families have caused gaps in the ability of the elderly to enjoy medical care.

Strengthen the reserve and cultivation of nursing talents, and improve professional quality: The government, as a provider and guarantor of public products and services, should continuously increase the investment and training of nursing talents, and strictly implement the "certificated work" system for elderly nursing homes. Colleges and universities should adjust, enrich and optimize the teaching curriculum according to the training level of nursing talents. Elderly nursing homes should also regularly carry out training and continuing education based on the problems exposed in their own operations and the work needs of medical staff. At the same time, in order to ensure the orderly provision of elderly care services, elderly nursing homes should strictly hold a certificated employment system. Control the quality of care from the source.

First, a nursing insurance system with mutual aid and adjustment functions. At the same time, the government also encourages the development of commercial nursing insurance through economic leverage and supporting policies; second, the source of insurance funds should be implemented by the government, society, and individuals. The principle of simultaneous payment, and can be managed by combining old-age care insurance, endowment insurance, and medical insurance. At the same time, the government grants financial subsidies to the poor and exempts them from paying insurance premiums; thirdly, the payment model is reasonably determined through the classification and grading of nursing care. Through the establishment of a scientific and standardized evaluation system, the nursing services provided are divided, and the corresponding charging system is established, so as to improve the quality of nursing services and the efficiency of supply.

4.2. Systematic design of elderly health monitoring

Power frequency interference-the space electromagnetic interference generated by the power supply equipment in the environment, the frequency is 50 Hz, the components are sinusoidal signals and harmonics, and the maximum amplitude is close to half of the ECG peak value. As shown in Fig. 3, the superposition of sinusoidal signals on the ECG graph is power frequency interference.

As shown in Fig. 4, baseline drift is usually caused by low-frequency interference, such as human respiration and electrode movement. The baseline variation amplitude at low frequencies is 15% of the ECG peak value.

As shown in Fig. 5, the EMG baseline is relatively obscure, and the voltage fluctuation is small. Features: The instantaneous energy is concentrated at 30-300Hz.

X-ray diffraction (XRD) measurement was performed on polyester fabric and graphene textile samples with an X-ray diffractometer, using CuKa radiation (λ = 0.15418nm), V = 30kV, I = 30mA. In the range of 5–60º, the scan rate is maintained at 0.1s-1/2θ, and the result is shown in Fig. 6. The diffraction curve of polyester textiles and graphene textiles are typical patterns of polyester fabrics. Due to the low concentration of graphene, it is difficult to find characteristic peaks of graphene in the patterns of modified polyester fabrics.

The process of thin-film pressure sensor: advanced screen printing. Production process: first attach conductive silver paste, nano-force sensitive semiconductor ink, etc. on the film substrate, and then make it through a special process. Working principle: As the external pressure increases, the resistance value of the corresponding pressure coating becomes smaller and smaller. Through software control, the pressure is collected and processed, and the pressure size and dynamic distribution can be obtained. The technical parameters of the membrane pressure sensor are shown in Table 3.

Table 3

Membrane pressure sensor technical parameters
Parameter category	Attribute/value	Remarks
Substrate material	PET/PI	Optional
Substrate thickness	350–400µm	Adjustable
Size	No. 42	European standard, adjustable
Film color	Black	——
Film thickness	20–25µm	Adjustable
Square resistance	5MΩ-10MΩ	Adjustable
Measuring range	0.5MPa	Equivalent to 0-50kg/cm³
Stable resistance change range	100KΩ-10KΩ	Can be adjusted according to customer needs
Hysteresis	< 5%	——
Drift	< 8%	——
Static resistance value	> 10MΩ	Generally infinity
Operating temperature	-30 ~ + 50℃	——
Operating Voltage	3-5V	Depends on the situation
Working humidity	0–90%	——
Reaction time	< 20ms	——

The determined coordinates of all points are shown in Table 4. For example, the actual coordinates of point 1 are (11, 32), and the strobe coordinates are (IO_B, OUT_1). When the point “1” is collected, IO_B is strobed. At this time, the output of OUT_1 The value is the voltage value at that point.

Table 4

Gating coordinates of plantar pressure sensor distribution
Serial number	Coordinate	Gated coordinates	Serial number	Coordinate	Gated coordinates
1	(11,32)	(IO_B,OUT_1)	15	(5.2,19)	(IO_C,OUT_7)
2	(9.5,32)	(IO_A,OUT_1)	16	(2.5,20.5)	(IO_D,OUT_6)
3	(7.8,32.5)	(IO_C,OUT_1)	17	(5.2,16.2)	(IO_C,OUT_5)
4	(10.8,28.3)	(IO_B,OUT_2)	18	(2.5,17.8)	(IO_D,OUT_6)
5	(8,28)	(IO_A,OUT_2)	19	(7.3,13.5)	(IO_B,OUT_4)
6	(5.5,27)	(IO_C,OUT_2)	20	(5.2,13)	(IO_C,OUT_4)
7	(10,25.3)	(IO_A,OUT_8)	21	(3,14.3)	(IO_D,OUT_5)
8	(8,25.3)	(IO_A,OUT_3)	22	(7.3,7.4)	(IO_B,OUT_8)
9	(5.5,24.5)	(IO_C,OUT_3)	23	(5.5,9)	(IO_C,OUT_8)
10	(3,25.3)	(IO_D,OUT_2)	24	(3,11)	(IO_D,OUT_4)
11	(22.8,9.2)	(IO_B,OUT_7)	25	(5.2,5.5)	(IO_C,OUT_9)
12	(5.7,22)	(IO_C,OUT_7)	26	(3,7.5)	(IO_D,OUT_8)
13	(2.2,23)	(IO_D,OUT_3)	27	(6.8,3.2)	(IO_B,OUT_9)
14	(8,18.5)	(IO_B,OUT_6)	28	(3.7,3.3)	(IO_C,OUT_9)

The intelligent medical diagnosis system studied in this paper is composed of a traditional Chinese medicine diagnosis system, a western medicine diagnosis system, and a medical record database. The TCM diagnostic system and the Western medicine diagnostic system are the main body of the system. The TCM diagnostic system is based on a case-based reasoning model, using human body information collection equipment to simulate the process of TCM diagnostics and realize TCM diagnostic engineering, while the Western medicine diagnostic system is a neural network optimized based on genetic algorithms. Model, obtain the diagnosis and treatment method of the disease from the medical record database of the hospital information system. A case-based reasoning model is designed, and the body information collection equipment is used to examine the patient's body, and the ever-growing case library is used as a knowledge base for reasoning to realize the diagnosis function of Chinese medicine. A diagnosis model based on neural network is designed, using the existing hospital diagnosis system, and using the hospital's existing case database as network training samples to realize the western medicine diagnosis function. The aging of the population continues to increase, and the elderly’s needs for medical care are increasing day by day, and the existing resources cannot meet the diverse needs of the elderly. The model of combining medical care and elderly care for the elderly is an innovative model and a new way of solving the current dilemma of elderly care.

Compliance with Ethical Standards

Conflict of interest

The authors declare that they have no conflict of interests

Ethical approval

This article does not contain any studies with human participants performed by any of the authors.

Data Availability

Data will be made available on request.

Sun G, Wang M, Wu L (2011) Unexpected results of Extended Fractional Kalman Filter for parameter identification in fractional order chaotic systems. Int J Innovative Comput Inform Control 7(9):5341–5352
Lo L-C, Chen Y-F, Chen W-J, Cheng T-L, Chiang JY (2012) “The study on the agreement between automatic tongue diagnosis system and traditional chinese medicine practitioners,” Evidence-Based Complementary and Alternative Medicine, vol. Article ID 505063, 9 pages, 2012
Kong L, Yang T, Fan HY et al (2020) Research progress of objective quantification of four diagnosis of traditional Chinese medicine. Chin J Ethnomed Ethnopharmacy 29(1):63–66
Huang Q, Zhang F, Li X (2018) “Machine learning in ultrasound computer-aided diagnostic systems: a survey,” Biomed Research International, vol. Article ID 5137904, 10 pages, 2018
Li X, Wang Y, Liu G (2020) Structured medical pathology data hiding information association mining algorithm based on optimized convolutional neural network. IEEE Access 8(1):1443–1452
Gope P, Hwang T (2016) BSN-care: a secure IoT-based modern healthcare system using body sensor network. IEEE Sens J 16(5):1368–1376
Ndiaye M, Hancke G, Abu-Mahfouz A (2017) Software defined networking for improved wireless sensor network management: a survey. Sensors 17(5):1031
Peng C, Qian K, Wang C (2015) Design and application of a VOC-monitoring system based on a ZigBee wireless sensor network. IEEE Sens J 15(4):2255–2268
Gao F, Wu P (2020) “A trajectory planning algorithm for medical manipulators based on adaptive particle swarm optimization and fuzzy neural network,” Journal Européen des Systèmes Automatisés, vol. 53, no. 3, pp. 429–435,
Wu H, Xu W, Yao B, Hu Y, Feng H (2020) Interacting multiple model-based adaptive trajectory prediction for anticipative human following of mobile industrial robot. Procedia Comput Sci 176:3692–3701
Jia WU, Liu CHANG, Genghua YU (2020) Effective data decision-making and transmission system based on mobile health for chronic disease management in the elderly. IEEE Syst J 17:1–12
Zhang X, Kim J, Patzer RE, Pitts SR, Patzer A, Schrager JD (2018) Prediction of emergency department hospital admission based on natural language processing and neural networks. Methods Inf Med 56(5):377–389
Cheng ZC, Jiang Y, Xu MY, Wang HY, Jiang DZ (2002) The application and development of artificial intelligence in medical diagnosis systems. J Biomed Eng 19(3):505–509
Baig MM, Gholamhosseini H, Connolly MJ (2013) A comprehensive survey of wearable and wireless ECG monitoring systems for older adults. Med Biol Eng Comput 51(5):485–495
Rahmani AM, Thanigaivelan NK, Gia TN et al (2015) “Smart e-health gateway: bringing intelligence to internet-of-things based ubiquitous healthcare systems,” in 2015 12th Annual IEEE Consumer Communications and Networking Conference (CCNC), pp. 826–834, Las Vegas, NV, USA, IEEE

Download PDF

Journal Publication

published 05 Jun, 2023

Read the published version in Soft Computing →

Reviewers invited by journal
08 Apr, 2023
Editor assigned by journal
08 Apr, 2023
First submitted to journal
31 Mar, 2023

You are reading this latest preprint version

Application of Multi-sensor Network and Artificial Intelligence in Health Monitoring of Medical Geriatric Care

Status:

Journal Publication

Version 1

Abstract

Figures

1. Introduction

2. Related Work

3. Application Of Multi-sensor Network And Artificial Intelligence Technology In Medical System

3.1. Multi-sensor network algorithm

3.2. Application of artificial intelligence technology in medical systems

4. Medical Elderly Care Model And Health Monitoring System Design

4.1. Medical elderly care model

4.2. Systematic design of elderly health monitoring

5. Conclusion

Declarations

References

Status:

Journal Publication

Version 1