×

Monitoring and predictive analytics of employee health status. Wearable sensors and devices for Smart PPE/Smart clothing

July 2020

Analytical Report (full version)

Request cost of the full version of the study: news@json.tv

Analytical Report (full version)

Monitoring and predictive analytics of employee health status. Wearable sensors and devices for Smart PPE/Smart clothing
Monitoring and predictive analytics of employee health status. Wearable sensors and devices for Smart PPE/Smart clothing
July 2020

Monitoring and predictive analytics of employee health status. Wearable sensors and devices for Smart PPE/Smart clothing

July 2020

To register or Log in, to download the file

Download file

+7 926 561 09 80; news@json.tv

Write, call, if you have questions

The main goal of new research by J'son & Partners Consulting was to analyze the global market for algorithms (models) that can be used to build predictive analytics on the health status of employees working in hazardous industries like mining, petrochemical, etc. The analysis of such algorithms allowed to identify the key parameters of the human condition (Vital Signs) for monitoring. In parallel, the authors also carried out an analysis of the market for wearable sensors & devices used for collecting data on the human condition and the environment. As a result, there have been developed recommendations for creating the so-called Smart Clothing or advanced version of different elements of Personal Protective Equipments (PPE), available on the market today.

 

In today’s world, we can see a large number of algorithms for evaluating the state of personnel in various industries, taking into account their specifics, which can be put into predictive analytical systems. In addition, there are various machine learning models created for solving specific medical problems and predicting the human condition. Usually, in this case, we are talking about post-hospital rehabilitation or observation of people with chronic diseases. Various research groups carry out studies in the world and in Russia using machine learning models to predict the state of certain groups of the population, and such studies have not yet reached the stage of commercial implementation.

 

In general, evaluation of the condition of such a complex biological system as a human body requires the use of a set of indicators that are significant in terms of diagnostics. These indicators characterize the functioning of various systems of the body — first of all, cardiovascular, as well as respiratory, nervous, musculoskeletal, etc. If any of the human body systems are affected in any way, the organism maintains homeostasis (dynamic constancy of the internal environment and basic physiological functions) making changes in the functioning of other systems.

 

Scientific studies show that in order to monitor all vital signs of a person (heart rate/pulse, pressure, respiratory rate, temperature, blood oxygen level), four types of sensors are enough. They can receive and compare signals using known algorithms and then produce all the necessary data. Basing on these indicators, you can create multi-sensory "smart clothing" which is able to monitor most of the necessary parameters, like ECG, PPG (photoplethysmogram), GSR (galvanic skin response) and temperature.

 

PPG and temperature sensors have become the de facto standard for most commercial fitness devices, and GSR (stress/emotion) sensors are less common. Relatively simple optical PPG sensors using various processing algorithms allow obtaining heart rate, blood pressure, respiratory rate, blood oxygen saturation (SpO2), as well as heart rate variability. Along with motion sensors (accelerometer, gyroscope, GPS), this allows to use simple mass devices with acceptable accuracy to carry out a very wide range of tasks.

 

Fig. 1. Ability to extract key vital parameters of a person by four types of sensors (ECG, PPG, GSR, temperature)

 

As for taking ECG parameters remotely, it should be noted that this is the most detailed, and most importantly, a more accurate indicator which is as close as possible to the reference stationary outpatient solutions, on the basis of which you can make far-reaching conclusions (heart rate, pressure, breathing, heart rate variability or blood oxygen level, as well as the potential diagnosis of almost all diseases). This area though remains the domain of narrow-industry solutions. As a rule, sensor patches are used for taking ECG parameters. A patch is attached to the chest area for 1-7 days. Or such solutions are represented by similarly highly specialized T-shirts made of conductive textiles. The advantage of such a solution is that it allows to place a larger number of sensors (in addition to ECG it can be temperature, breathing, etc.), as well as a wearable device that combines the function of a battery and an aggregator of received data or a communicator. It is also the preferred option for creating complex PPE, all physiological parameters, in this case, are aggregated by sensors in the tee, and a protective coverall (and/or hard hat) with environmental analysis sensors are put on top of it.

 

According to a number of open medical scientific articles and various estimates mentioned in them, there are already more than 100 different algorithms for evaluating, interpreting, comparing various physiological indicators, and building predictive models based on them. Most of them to some extent take into account cardiovascular system indicators as an integral cross-section of the human condition.

 

Speaking about the choice of the type of algorithm for assessing the overall physiological state of a person, we can mark the two based on direct ECG analysis and indirect reading through PPG. Both options allow to capture and determine basic indicators, but ECG is more accurate (and more expensive), approaching professional outpatient analogs, while PPG is simpler and cheaper (and its measurement error is higher). Both methods have strengths and weaknesses, but algorithms based on them allow to show heart rate variability (HRV) which is a necessary indicator for measuring stress, general physiological load and fatigue and helps prevent overstrain and possible accidents at work. Regarding monotonous professions (drivers, automated control system operators, etc.), analysis of GSR (electro-thermal resistance) sensor data requires special attention.

 

As for the analysis of environmental parameters, a number of portable and wearable devices are already available on the market, but they are usually intended for the mass market with appropriate tasks (air quality, fine particle content, black carbon, radiation level). Noise and light levels, as well as pressure, are becoming more common and standard functionality for many wearable commercial devices from leading manufacturers. However, when it comes to solving specific tasks (monitoring of flammable gases, warning of high voltage), only specialized monofunctional devices are currently available on the market. Integration of their capabilities, including monitoring of human physiological parameters, is one of the tasks of the near future.

 

As for the capabilities and reproducibility of such technologies in Russia, integration of sensors for reading human physiological parameters and environmental analysis sensors (both common and specialized) is at the same stage as in the rest world – there are no ready-made complex and commercially available solutions yet, research and development are ongoing, and some technologies are being piloted.

 

In addition, there are at least two solutions that can be used to create smart clothing/PPE in the future. These solutions are supposed to be used in the mining and oil and gas industries. They are IQ-Beat developments that supplement T-shirts with an ECG external jumpsuit and environmental analysis sensors (gas and temperature analyzers) attached in it. The other one is "Goodwin-Neva", a solution based on a radio communication system with the function of production conditions monitoring, which can be supplemented with external devices, such as gas analyzers/dosimeters/high-voltage sensors, and ECG sensors/fitness bracelets. Another possible option for creating a full-fledged PPE is to expand the functionality of "Biotelemetry", an industry solution by the ITPS integrator for the oil and gas sector based on the Russian-developed HealBe GoBe2 bracelet (an addition to ECG and environmental analysis sensors). A possible alternative option for IQ-Beat is the diagnostic system "Screenfax" by Medscreen which is able to diagnose more than 40 diseases using ECG (4 leads) – only the stationary version is currently in commercial operation, but the company is working on a more compact solution, suitable for wearable devices (2020).

 

The task of integrating human physiology and environment sensors with a comprehensive commercially available PPE solution has not yet been solved anywhere in the world. Both in Russia and in the rest world we can see that individual industrial-grade elements (gas analyzers, ECG shirts, voltage sensors, smart bracelets, etc.) have already been worked out and are available, the first experiments and pilot projects are underway to integrate them.

 

Fig. 2. Perspective Smart clothing elements available on the market today

 

Devices reading ECG directly or those equipped with simpler PPG sensors can determine the general physiological state of an employee according to the state of their cardiovascular system with high accuracy. Comparison of these data with temperature, respiratory rate, and sweating, as well as motion analysis only increases the correlation, often without requiring comparison with environmental data to understand that an emergency has already occurred. In this case, it remains only to promptly inform the other employees and provide assistance to the victims. In this regard, the use of additional specialized wearable devices and sensors, for example, voltage indicators or various gas analyzers, is aimed, first of all, at the timely notification of a person before entering the "red" zone, preventing emergency and avoiding production injuries.

 

In order to monitor the basic physiological parameters of a person, it is advisable to identify four key types of sensors, basing on the readings of which everything else can be calculated. This is an ECG sensor, PPG, skin electrodermal activity sensor (GSR), and temperature sensor. Also, Smart Clothing/PPE requires motion sensors, gyroscopes, an accelerometer, and additional sensors for analyzing the state of the environment, due to the specifics of a particular industry (for example, temperature and gas analyzers for the mining or oil and gas industries).

 

____________________

 

This information note was prepared by the J'son & Partners Consulting. We work hard to provide factual and prognostic data that fully reflect the situation and available at the time of release. J'son & Partners Consulting reserves the right to revise the data after publication of new official information by individual players. 

 

Copyright © 2020, J'son & Partners Consulting. The media can use the text, graphics, and data contained in this market review only using a link to the source of information - J'son & Partners Consulting or with an active link to the JSON.TV portal

 

™ J'son & Partners [registered trademark] 

 

Detailed research results are presented in the full version of the report:

 

Monitoring and predictive analytics of employee health status. Wearable sensors and devices for Smart PPE/Smart clothing

 

Contents

1. ALGORITHMS, PREDICTIVE MODELS AND CRITICAL PARAMETERS FOR MONITORING

1.1. STUDY OF THE MARKET FOR ALGORITHMS (MODELS) THAT CAN BE USED FOR PREDICTIVE ANALYTICS BASED ON THE EMPLOYEE'S HEALTH STATUS.

1.1.1. Study of the market for available algorithms (models) used for monitoring and preventive control of physiological parameters

1.1.2. Algorithm for processing/analyzing biomedical signals in remote monitoring

1.1.3. Choosing the optimal task for wearable systems when solving human health predictive analytics problems

1.1.4. Specific models for evaluating the employee’s health status and their parameters which can be embedded in predictive models

1.1.5. Commercially available machine learning models created for specialized medical goals (monitoring and forecasting)

1.1.6. Scientific studies and examples of other models and algorithms used for predictive medical solutions

1.2. BASIC PARAMETERS FOR MONITORING HUMAN HEALTH WHEN USING WEARABLE DEVICES (VITAL SIGN MONITORING).

2. MARKET FOR WEARABLE HUMAN CONDITION SENSORS

2.1. OVERVIEW OF THE MARKET FOR WEARABLE HUMAN CONDITION SENSORS/DEVICES 

2.1.1. ECG

2.1.2. PPG (SpO2)

2.1.3. GSR

2.1.4. Temperature

2.2. RANKING OF SENSORS ACCORDING TO THE FOLLOWING CRITERIA

2.2.1. Measurement accuracy (in comparison with stationary measurement)

2.2.2. Applied technology (reproducibility in the Russian Federation) and its cost

2.2.3. Integral applicability indicator (accuracy of predictive analytics based on data collected from wearable sensors)

3. WEARABLE ENVIRONMENT ANALYSIS SENSORS

3.1. REVIEW OF WEARABLE ENVIRONMENT ANALYSIS SENSORS

3.2. COMPARABLE THREATS, POTENTIALLY DETECTED WHEN USING SENSORS IN THE OIL AND GAS INDUSTRY

3.2.1. Detection of explosive gas concentration

3.2.2. Detection of hydrogen sulfide (H2S) presence

3.2.3. Detection of electric voltage near a person's location

4.        CONCLUSIONS

 

List of figures

Fig. 1. Matrix of the main processes for data mining from wearable devices

Fig. 2. Schema of predictive modeling process organization

Fig. 3. GSR bracelet 

Fig. 4. Device for tracking the position of a driver's head and eyes

Fig. 5. How the SleepAlert system works

Fig. 6. The key formula for light absorption peaks from the Brain Bit patent for non-invasive spectral geometry and a general view of the device

Fig. 8. Patents for algorithms evaluating the human body condition based on the HealBe GoBe2 bracelet

Fig. 9. The architecture of the SafeOperator solution, which includes SafeLife

Fig. 10. The key feature of IQ-Beat is a convenient reading of ECG and other human parameters by sensors sewn into the fabric

Fig. 11. Goodwin-Neva, a radio communication system with personnel monitoring, labor protection and eco-monitoring functions 

Fig. 12. List of diseases diagnosed by the Medscreen ECG

Fig. 13. Plans to develop the Screenfax diagnostic system from a stationary version to a wearable device 

Fig. 14. Parameters analyzed by Welltory basing on VHR monitoring

Fig. 15. Example of reports in the Welltory app based on VHR analysis results

Fig. 16. Screenshot of the IBM Watson Analytics system user account used by the APP911 project (defining scales)

Fig. 17. Wearable devices by partners: Garmin (tracker), Guardhat (smart helmet), Mitsufuji (T-shirt with sensors), as part of the integration solution based on IBM Maximo Worker Insights

Fig. 18. Samsung s-Patch Cardio, general operation principle 

Fig. 19. Astroskin, smart clothing able to change sensor indicators

Fig. 20. Screenshot of one of VivoSense medical data analysis platform modules

Fig. 21. Zio XT and a patient status report based on wear results

Fig. 22. Ultra-precise neural network correctly detects atrial fibrillation (AF) from other rhythms (not AF) basing on PPG taken from a wrist

Fig. 23. The generalized architecture of wearable systems for human health monitoring

Fig. 24. Capabilities to extract key vital parameters of a person by four types of sensors (ECG, PPG, GSR, temperature sensors)

Fig. 25. Capabilities of the Valencell PPG module to synthesize and monitor various physiological parameters

Fig. 26. Compliance of some ECG reading T-shirts with international standards (medical device certification)

Fig. 27. Composition of the safety++ solution, a joint project by MIT & ENI (gas analyzers, sensors for reading basic medical indicators, special carbine boots and sensors for installation work).

Fig. 28. Prevalence of sensor types for detecting various gases

 

List of tables

Table. 1. Health assessment and predictive analytics systems available on the market

Table. 2. Integrated table for the most versatile wearable devices with sensors allowing to take other vital physiological parameters (PPG, GSR, temperature) in addition to ECG. Certificates

Table. 3. Parameters of gas analyzers, modules and sensors

Table. 4. Parameters of gas analyzers, modules and sensors for hydrogen sulfide (H2S)

Table. 5. Parameters of devices and sensors for measuring electrical voltage

Table. 6. Wearable devices and sensors reviewed in the report and comparable threats in the oil and gas industry that they can prevent