Security of Smart Things REU 2017

This repository contains all the code used to complete the Security of Smart Things REU project of privacy preserving indoor localization.

• REUServer directory contains all the Java code running on the Fingerprint Server
  
• REU2017 contains the Android Client Side Code

This program requires the MySQL Driver to communicate with FingerPrint server, found here

The JAR file containing all the homomorphic encryption schemes can be located at the following repository and with its documentation

Installation

Server Installation

You can run the server using Gradle as followsing

bash ./gradlew run -PchooseRole=Localization.server

Android Installation

Please note the Android installation has only been tested on Samsung Galaxy 3. But it should work with all other Android devices.

Upon downloading the repository, open the REU2017 folder using Android Studio. Follow the instructions here on how to load an application onto your phone.

Usage

Server set-up

To run the server, you need to populate the environment variables MYSQL_USER and MYSQL_PASSWORD to access the Fingerprint database.

Note: Upon initialization, it will expect MySQL to be running! If it isn't the server will not turn on.

Upon initializing the server will
- If there is preprocessed data, determine the number of APs used for each column - Assuming a port number was provided, check if it is a valid port number and open the server socket. Default is port 9254 - Create the database called "FIU" and a Table called "trainingpoints" to store the raw Wi-Fi scans.

It will also create a shell to do the following operations:
- clr Clear the shell - print Assuming the fingerprint database preprocessed the training data, print the lookup tables to a CSV file. - frequency Print a frequency map of all detected Access Points. Use this to tune your FSF parameter. - k Disabled for now. - test-FSF [0, 1] For a value between 0 and 1, determine how many access points would get filtered. - FSF [0, 1] Sets new FSF parameter. However, if you want to see this change, you will need to build a new lookup table. - process Build the lookup table for all floor maps - reset delete all lookup tables - exit close the shell and server

Upon each iteration of the shell, it will update you on: - If the fingerprint server has a lookup table - Number of columns in the lookup table - Number of Access Points detected in training data - Number of fingerprints of each floor map - Current value of K, for MCA/DMA (default set to 2). Currently, this is hard set as I fail to see the benefit in changing this value.

Phone setup

This is the

• Reset: Delete the Lookup table on the server
• Undo: Delete the last scan of Access Points/Wi-Fi signals
• Train Database: Open new menu, which opens a floor map to train with Wi-Fi APs and scans
• Process DB: Create lookup tables based on the training data
• Scan: TO BE UPDATED. Update server of new maps
• Localize: Open up new menu, and select which floor map you want to find your location in.
• Localization Scheme: On the bottom right, you can select a combination of server/client side, homomorphic encryption scheme, and localization algorithm.

Training Workflow

Localization Workflow

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Authors and acknowledgment

Code Author: Andrew Quijano

A. Quijano and K. Akkaya, "Server-Side Fingerprint-Based Indoor Localization Using Encrypted Sorting," 2019 IEEE 16th International Conference on Mobile Ad Hoc and Sensor Systems Workshops (MASSW), Monterey, CA, USA, 2019, pp. 53-57, doi: 10.1109/MASSW.2019.00017

P. Armengol, R. Tobkes, K. Akkaya, B. S. Ciftler and I. Guvenc, "Efficient Privacy-Preserving Fingerprint-Based Indoor Localization Using Crowdsourcing," 2015 IEEE 12th International Conference on Mobile Ad Hoc and Sensor Systems, Dallas, TX, 2015, pp. 549-554, doi: 10.1109/MASS.2015.76.

The work to create this repository was initially funded by the US NSF REU Site at FIU under the grant number REU CNS-1461119.

License

MIT

Project status

The project is functional at its current state. However, some optimizations do need to made for ease of use. This would only work best if I can re-obtain access to Broadway to test on multiple floors/add new floors as needed.

Also, there are a few things I want to upgrade: - Streamline the process of the client uploading/receiving floor maps. - See how to allow the client to set the value k for localizations. - I will provide screenshots on the Localization/Training Phase later this week.