INTRODUCTION

Smart Human Dynamics develops two products:

• Light shift (Sensor, Mobile application and Dashboard)

• Pregnancy coach (The same Sensor, another application and Dashboard)

BUSINESS OBJECTIVES

The objectives of the project set for IOMICO team by the Client were the following: 

• to review and improve the existing sensor firmware; 

• to review, improve and maintain the existing iOS and Android application launching releases; 

• to integrate a dongle into sensors network for their simultaneous work;

• to test the solution, prepare the products for clinical research.

BUSINESS PROBLEM/OPPORTUNITY

The Body position Sensor is the clinically-proven solution designed to minimize back sleeping occurrences by up to five times without causing disruptions to sleep. In the second or third trimester, when a pregnant woman assumes a back (supine) position, there is a risk of compression of the inferior vena cava (IVC), potentially leading to reduced oxygen supply to the fetus. Studies have consistently demonstrated a correlation between back sleeping and elevated risks of stillbirth, fetal growth restriction, and neonatal ICU admissions, with pregnant women spending an average of 1.5 hours in the back sleeping position. The same reasons could cause the risk of choking during COVID illness.

APPROACH

The iomico team reviewed and improved the existing sensor firmware, made code review of mobile applications.

1. We flashed the nrf 52833 dongle for installing a stable Bluetooth connection. It was a challenge to adapt the Bluetooth library to existing firmware, because there were no approved libraries before. The dongle enabling allowed to extend the solution architecture and use it as a broker to get data from several sensors and transmit them to the nurse's station. It’s useful when patients are cared for in an inpatient department.

2. The sensor has the accelerometer and the pedometer. All the data are measured, filtered and analyzed with both firmware and mobile application. The result depends on the user's personal data like weight, height and age. For pregnant women the phase of pregnancy is also important for data analytics. Finally the mobile application and dashboard visualize the human activity getting charts and showing figures for further analysis. The application can warn users about long back sleeping or other risky body positions. The sensor uses vibrator and LEDs to warn the device users.

3. The process of data analysis is very complicated. We had to learn medical studies and read clinical research results for better data interpretation, in order to improve the process of caring and make a better doctor's report. The combination of data from different sensors aligned with an internal timer makes the user’s sleep more safe. The application can distinguish different phases of sleeping and reduce the risk of missing an important medical event when the user is activeless.

TECHNICAL DETAILS

The product is supposed to have two architectures

1. Remote treatment

In this case a patient or pregnant woman connects the sensor (accelerometer and pedometer) with an installed mobile application. The app shows the live data and advises the patient when some issues are obvious. It stores all statistics into the internal DB, and is able to send it to cloud DB. The doctor can have access to it and see the analytics to ensure remote monitoring and care. 

This way has several benefits but weaknesses as well. The benefit is the additional support where presence in the inpatient department is not obligatory but meanwhile being at home without any control is risky.

The same things are going with COVID patients. Another point is the need to use the App and device remotely, f.e. in cases when inpatient departments are full by the reason of pandemic or epidemic period.

Version 1. Remote care

2. Hospital treatment

In this case the sensor connects to the gateway and transmits data to the stationary post via gateway. This architecture allows to enable several gateways and a big number of sensors at the same time. In this case all the devices can send urgent messages to the stationary post and the medical personnel provide immediate actions. The benefit of such architecture is emergency and quick technical support, sensors are easy to reuse by different patients.

Both approaches have a good prospect for the healthcare domain.It’s an example of enabling IoT with different approaches in dependency of the epidemic level during the difficult period.

LESSONS LEARNED

The development of hybrid mobile applications requires extended knowledge in Android and iOS systems.

• We have got experience in Codepush server for application update (UI updating).

• Gateway based architecture is a challenge if it was not supposed to be implemented before the start of engineering works.

• There is a lack of libraries for Node.Js for gateway support in Bluetooth communication.

• Cordova framework is not supported by many OS and there are no good libraries for implementing many modern features. For new projects it’s better to choose other frameworks.

CONCLUSION

The project was not supposed to scale, so a big part of the codebase was refactored.

• Each project should have been planned and structured, and have to have a clear vivid architecture for further development and maintenance.

• Projects of the healthcare domain requires a comprehensive testing, double check and deep code review.