Automated Contact Tracing and Surveillance
Omkar Patil
Final Year Undergraduate,
IIT Madras
E-mail: indian.op45@gmail.com
Contact: 7665103021
Idea Summary
Create an app which uses location information of the user to stifle the spread of coronavirus through
effective contact tracing and optional surveillance. The app will let people know if they have been in
close proximity to an infected person. This will be implemented by analyzing, if the user has come
within a certain distance of the affected person, based on their location data. The incentive for
people to use this functionality is to know if they have been or are in close proximity of a COVID+
person, and other updates about the spread- such as viral hotspots and nearby hospitals. People
whose location history reveal that they have been within 10m of an affected person will be notified
immediately and asked to self-quarantine themselves. The same will be let known to the health
officials. This will help in effective contact tracing to contain the secondary and even the tertiary
spread of the virus. Further, surveillance can be enforced through monitoring their location, and
other visual or sensory modalities.
Background
COVID-19 has been causing wise-spread panic and fear
among citizens all over the world, and lately in India. The
country has supposedly reached the second stage of
transmission and is moving towards the third stage-
where infected people who arrive from other countries
pass on the virus to locals. This is where contact tracing
becomes extremely crucial- if a person is identified to be
positive for COVID-19, the government has to
immediately trace all the people who have been in
contact with the infected individual and quarantine them
for a duration of 15 days. This is a tedious process, and
will soon become infeasible as the progression to the
third stage happens. Hence, utilizing technology will be
necessary in containing the spread of the virus. A similar
exercise was carried out in South Korea where travel
history of COVID-19 patients was shared to people
through messages[1]. It was one of the reasons why the
country was able to successfully flatten the curve. A
percentage of people self-quarantine themselves based on the knowledge of possible exposure.
Moreover, if information like virus hotspots are known to the people, it will further discourage them to
travel to those specific areas, mitigating the spread.
The proposal
The objective is to create an interface which will alert people, if they have been in close proximity with
a COVID+ patient. This will have multiple benefits- firstly, a section of those people will self-quarantine
themselves. Second, those people will be vigilant for any signs of symptoms and are likely to take
prompt action. Health officials can check-up on them if required through phone calls.
Today it is possible to get the fine location of the phone, and indeed many apps like Google and
Facebook leverage it. The government can use these location services for controlling the pandemic
and stifling it’s spread - without breaching the user’s privacy. Often people travel to a lot of places
before testing positive for the virus, making it necessary for health officials to perform contact tracing
of that individual. If such a situation arises with ACTS in place, the health officials can extract the
location history of that person with request or authority and upload the same on the app server. If the
app is published without government support, then the patient travel history can be obtained from
crowd-sourced data[2]. The data will be anonymized and sent to all the users, where the user's location
history will be analyzed to check if he or she has been in the vicinity of the patient. If so, the user will
immediately get a notification to self-quarantine themselves and other crucial information. The
interface will also let the health officials know the phone numbers and locations of the aforementioned
people that have come in close proximity so that they can be checked upon at a later point of time.
It is possible to get the location of a person within 10m of accuracy through Google Fused Location
Provider API[3,4]. The app will store the user’s location history for the past 15 days. The patient’s
location history will have to be manually retrieved from Google Takeout service, and then sent to the
app servers, where it will be stripped of the user information. Only the set of timestamped latitudes and
longitudes will be sent to respective phones to check for proximity with the users location history. The
task of checking for proximity can be carried out by finding the distance between the user and patient
coordinates in every 10 min interval. Many optimizations can be made which reduce the computational
power required for the same. The architecture is shown in the below figure. The computation of
proximity can be carried out on the app server also, but is purposefully done on the user’s phone to
retain his/her location data on their phone, to calm privacy fears upto some extent.
Based on the data of the location matches and positive patients, graphical models can be developed
which will help health officials and people identify virus hotspots. Accordingly, medical supplies and
personnel can be directed to the worst affected regions. This will also be an effective strategy to
disincentivize people to go into specific areas.
[1] South Korea is reporting intimate details of COVID-19 cases: has it helped?- https://www.nature.com/articles/d41586-020-00740-y
[2] API for crowd-sourced data on travel history of patients before they tested positive- https://github.com/amodm/api-covid19-in
[3] How accurate is GPS?- https://www.gps.gov/systems/gps/performance/accuracy/
[4] Google’s Fused Location API- https://developers.google.com/location-context/fused-location-provider
[5] Decimal degrees- http://wiki.gis.com/wiki/index.php/Decimal_degrees
Privacy and Feasibility Issues
Location sharing is often associated with breach of privacy; hence this feature takes care to keep the
information distributed, rather than centralized. The location history of only COVID 19 patients is
shared to the users, that too in an anonymized, cryptic format. The app only receives a set of
coordinates from the server which are associated with the travel history of the patient. The location
data collected by the app is never sent to the server, but stored in the user’s phone.
The Fused Location Provider API provided by Google, which fuses location data from GPS, WiFi and
mobile networks is accurate upto 7 digits- which implies a location accuracy in centimeters[5]. Google
as such stores the location information every 3 or 6 minutes depending upon the time of the day. The
fused location API broadcasts the location of the phone a little more infrequently, but enough to serve
our purpose.
Business Standpoint
The government of India has recently released an app in a similar vein- but a lot of countries around
the world do not have the time and workforce to create such an interface for effective contact tracing
as they are battling with this problem. So a multilingual app could be made for different countries to
implement, and could have other necessary features like enforcing quarantine of the affected persons.
