Minimized health risks and costs reduction by automating the real-time support service.

• ChallengeEnhance IoT healthcare application with AI
• SolutionML model to process data in the Client’s database and generate push notifications
• Technologies and toolsPython + DS, APNs, AWS Lambda, PostgreSQL, Swift (iOS)

Client

The Client is a European-based company specializing in delivering IT solutions for businesses across different industries. To improve the existing business approach by implementing modern technologies, the client needed a healthcare application integrated with the IoT and AI solutions to improve working conditions for employees and boost the efficiency of services. The Client was looking for experts in the AI software development field.

Challenge: enhance IoT healthcare application with AI

As an experienced provider of AI software development services, InData Labs was challenged to train a machine learning (ML) model to augment the Client’s existing solution. Our engineering team was tasked to build a model able to efficiently process large amounts of historical data and data gathered from wearable devices to create and deliver data-based push notifications for iOS as the output.

The IoT component was a wristband that measured a driver’s heart rate before and during each trip. This data wasn’t enough for a reliable evaluation of health status. For systemizing and crunching more data, the existing solution required an additional ML-based model.

Solution: ML model to process several types of data and generate personalized notifications

The IoT healthcare application is intended for drivers of commercial vehicles having health problems. Preventing emergency cases is a part of the business owner’s employee care and expense management program.

The InData Labs’ team started with scrutinizing the existing solution. Our AI research and development experts dived into the capabilities of the existing app to come up with the best ways to enhance it.

Apart from the data collected via wristbands, we fetched the following types of data to the model as the input:

• Driver’s heart rate
• Driver’s gender and age
• Historical heart rate data
• Weather information
• Time of day
• Route & destination

The gathered data was fed into the Client’s database for further processing. So, the following data sources were used to get the input for the ML-model:

Based on the comprehensive input data, we trained the model to perform the following tasks:

• Analyze the health status of drivers
• Detect health problems
• Generate push notifications
• Send alerts and recommendations to drivers

The algorithm splits the received output into the 3 categories or zones. Depending on one or another category, different push notifications and recommendations are sent to drivers from the server. The examples are the following:

To complete the task, the InData Labs’ team used the Apple Push Notification service (APNs) to enable the remote notifications feature. Our team also used the Client’s service to store and process data.

Result: ML-powered expense management solution, cost reduction as a result of processes automation

The InData Labs’ team successfully delivered the ML model that allowed the merging of IoT and AI to enhance the Client’s healthcare application. The provided algorithm facilitated data crunching and enabled data-driven push notifications for iOS devices.

The embedded algorithm efficiently worked with different types of data and allowed creating more personalized recommendations. Also, the solution by InData Labs allowed the owner of a commercial vehicle fleet to take better care of employees, improve working conditions for drivers, minimize health risks, and cut costs by automating the real-time support service.