The Situation

Our client needed a contact-tracing mobile application & telemedicine solution to prevent the spread of the pandemic.

The Requirement

To create Europe’s 1st COVID-19 contact tracing application to be used by the health authorities and governments within the framework of fighting with the spread of the virus.

Client’s Brief

Our client is a non-profit organization from Germany currently working to stop the spread of COVID-19. Their essential requirement was to be supplied with a swift validation of the idea of using GPS satellite-location tracking as a tool for preventing the spread of the virus while simultaneously docking down the medical system overload.

How Contact-Tracing Application Can Look Nowadays
The Challenge

Given that the telemedicine industry is just emerging on the horizon of neoteric medical solutions, our tasks boiled down to conducting a completely new research. We had to pioneer the development of a contact tracing app for preventing the further spread of COVID-19. Besides, we have been tasked with developing an architecture that would store location data for more than 10 million users. While a thorough research of the GPS/BLE applicability, along with defining the tracing protocols were paramount, we have also found it exponentially crucial to limit the infrastructure expenses for our client.

Meeting the Challenge Ralabs Way

An in-depth preliminary research marks the beginning of any project at Ralabs. We have scrutinized the WHO’s position on the issue and it was clear that the organization defined the contact tracing interventions as one of the most efficient measures in controlling the COVID-19 outbreaks worldwide. Thus, while addressing the challenge, our engineers, led by our CTO, have embarked upon developing a contact tracing app that is meant to help medical authorities and governments stop the spread of the virus. establishing and conducting contact tracing mobile app development process. We released the first version in Apple's App Store and Google's Play Store within just one month.

Team

So, what have we done?We:

01

Found a solution to controlling the spread of COVID-19 and future pandemics

Rapidly developed a mobile app that will identify the virus-communicable contacts between people

02

Utilized a Pan-European Privacy-Preserving Proximity Tracing protocol (PEPP-PT/PEPP)

Used a Central Server Model to ensure anonymous identification

03

Built an infrastructure for 10+ million users

Designed the architecture and suggested ways to optimize costs and the system load.

04

Researched possible solutions for digital contact tracing

Researched contact-tracing protocols, the accuracy of GPS locations, accuracy of Bluetooth LE tracing, and Beacons

05

Decentralized data storage using Blockchain and AI Technologies

Provided full anonymization and security while complying with EU Data Protection Guidelines (GDPR)

The Process:A Stepwise Description

01

Research

The aim of our research was to find the best solution to track the spread of the virus. Our goal was to do this with the least effort required by monitoring intersections of infected people.

01

Search Queries

Up until now, Google Flu Tracker (GFT) was the only tool for monitoring flu trends. While using Google Search queries, it was meant to predict the possible flu outbreaks with a 97% accuracy. In 2013, GFT failed to predict the peak of the flu season missing on their estimates by 140%. Following that calamity, Harvard, Houston, and Northeastern researchers called GFT data unreliable and a dangerous source for any decision-making.

02

Digital Contact Tracing

Instead, we used a Contact-Tracing Approach and utilized mobile devices to identify contacts of infected persons. During the COVID-19 pandemic, several protocols were developed to allow for wide-scale digital contact tracing.

We researched both centralized protocols, such as PEPP-PT (Pan-European Privacy-Preserving Proximity Tracing) and the decentralized ones, such as DP-3T (Decentralized Privacy-Preserving Proximity Tracing), and BlueTrace. Each of the protocols scrutinized have offered a unique method for processing the contact history, as it could either be stored and processed by a central server or by individual clients in the network.When considering the limitations of mobile devices, there are several different technologies in a smartphone that can be used to track movements and intersections of the infected.

Bluetooth LE and Beacons Tracing

Bluetooth is one of the most accurate technologies when it comes to proximity identification. BLE Tracing measures the Received Signal Strength Indicator (RSSI) of a given Bluetooth connection to estimate the distance between devices. The stronger the signal, the closer the devices are to each other.

Location-based Tracking

Geolocation of devices can be measured based on the data from GNSS satellites. Nonetheless, proper quality can only be achieved by using network-based location tracking. Still, it should be acknowledged that GPS has a couple of advantages to offer:

Bluetooth LE and Beacons Tracing

Bluetooth is one of the most accurate technologies when it comes to proximity identification. BLE Tracing measures the Received Signal Strength Indicator (RSSI) of a given Bluetooth connection to estimate the distance between devices. The stronger the signal, the closer the devices are to each other.

Benefits

  • Battery-saving: Very low power usage is required to preserve device battery life
  • Keeps sensitive information private. (It doesn’t need to store location data)
  • Much better accuracy compared to GPS
  • High accuracy via proximity detection
  • iBeacon measuring of microlocation activities
  • A well-documented standardized protocol
  • High accuracy via proximity detection

Limitations

  • Testing using different devices may be needed to assure the accuracy required
  • “Calibration" is needed to measure BLE signal power
  • Bluetooth must be turned on at all times. (This could create additional risks)
  • Bluetooth requires a vast amount of time for testing
  • iBeacon results in poor quality in real cases (e.g., in museums)
  • Android/iOS can present limitations (background mode and delay receiving up-to-date values)
  • Bluetooth LE is noisy. BLE uses broadcast "advertising" to announce their presence, thus making it not ideal for outdoor environments
  • Bluetooth has technical vulnerabilities, such as BlueFrag and remote code execution

Location-based Tracking

Geolocation of devices can be measured based on the data from GNSS satellites. Nonetheless, proper quality can only be achieved by using network-based location tracking. Still, it should be acknowledged that GPS has a couple of advantages to offer:

Benefits

  • It works on both iOS/Android platforms
  • It does not require turning on specific features (e.g., Wi-Fi or cellular)

Limitations

  • Accuracy depends on the position of the GNSS satellites, characteristics of the surroundings (buildings, tree cover, valleys, etc.), weather, phone and physical chip, quality of OSS, and Wi-Fi. The app works worse in large cities (e.g., NYC, Berlin, Kyiv)
  • It is less accurate for "micro-location" proximity-based activities
  • There are privacy issues when storing sensitive location information with the potential to deanonymize. Telecommunications operators ('telcos') may hand over customers’ data
  • GPS requires a lot of power, and drains mobile device batteries quickly

Infrastructure costs planning

We have calculated the budget the customer needs for establishing a corresponding infrastructure. According to the planned load, we predicted that the optimal size for the database should maintain a minimum of 1,000,000 users, and be scalable to the amount of the users equal to the German population. Based on this prediction, we have counted the monthly cost of the database maintenance, optimized costs, and facilitated communication with AWS and MongoDB.

02

Architecture

Following our research, we chose the PEPP-PT protocol for digital contact tracing. The next step was to build Centralized network-based (PEPP-PT) architecture for keeping the privacy and confidentiality of the app users.

This diagram demonstrates:

  • How we protected the dataAll internal and inter-services communications are encrypted via HTTPS. No sensitive data is stored anywhere. The system uses a unique generated ID for any identifications of the device. The approach is based on the PEPP-PT Protocol. We used encrypted keys to preserve non-exposed users’ privacy.
  • How we cut the costs on the infrastructureWe optimized the algorithm by dividing it into two tasks. The first algorithm finds the intersections for a specific infected person within the past 24 hours, while the second finds the intersections with all infected persons within the past 24 hours. We also integrated Amazon Mongo, which resulted in a cost reduction of 80%.
  • How we connected the third-party servicesWe integrated LogDNA, Cloud Logging, Gmail, and Google Drive.

03

Development &Testing

We developed the following features:

Monitored Risk of Infection Spreading

Based on anonymized location data tracking, we implemented a Traffic Light System that identifies the risk of infection for users based on their recently visited places, people contacted, and historical health reports.

Human-in-the-Loop Documents Check

The system gathers data about the users’ location and validates reports with the infection test results pursuant to GDPR e-privacy provisions, and in total compliance with the OWASP security guidelines. After users scan the documents, the system executes human-in-the-loop checks that envisage a verification by the health care authorities. If the results turn positive, the system sends notifications to all users who have been contacted by the infected person.

Algorithm-based Intersections Matching

The checker asks users about relevant symptoms and based on the responses received, it creates a risk profile while automatically compiling the corresponding health reports. Based on this information, the system suggests custom recommendations from NHS and RKI institutions.

Other Features

We have taken proper care about some additional features for the app which were meant to make it even more convenient for the customers to use:

Algorithm-based intersections matching

Smart caching every 14 days to decrease the amount of stored data

Localization for English, German, Swiss markets

04

Release


App Store ReleaseOvercoming App Store restrictions

Google Play Release

On-the-air Update Release of updates with OTA approach

Get free-of-charge consultation from our experts

A top-quality review, analysis, and feedback on your requirements and technical procedures.Professional advice on architecture and the UX side of your current or future business.Help with the R&D plan.

Schedule a Free Consultation

Daniel

Technical Lead

Andrew

CEO

Serhii

UI/UX
Design Lead

Roman

CTO

A top-quality review, analysis, and feedback on your requirements and technical procedures.Professional advice on architecture and the UX side of your current or future business.Help with the R&D plan.

Schedule a Free Consultation

The Outcome

It was a pure joy to observe how the hard work that we put into the project has started paying us back with genuinely great results. So, this is what we have been able to achieve in less than one month:

Launched the first iOS COVID-19 tracking app on the German and Swiss markets

Gathered more than 15 million location coordinates

Reached the point of over 1,000 downloads since the app’s release

Engaged 200 beta testers into the team worldwide

Successfully validated the concept of using a contact tracing app for preventing an infection spreading

Overcame the App Store’s restrictions and released the mobile app with the OTA approach

Received 98% of excellent grades in the App Store

Let’s SeeWhat We Can Dofor You

Tell us about your project, and we will send you an estimate within 48 hours.

Olha

Engagement Manager

Tell us about your project, and we will send you an estimate within 48 hours.

I need help with