This is the page description.

—  AI & Code—

AI and Software Projects


Bluetooth Low Energy Lock

This device uses BLE (Bluetooth Low Energy) to enable keyless access. Schedule times for entry, revoke access on the fly, or email your keys to your friend waiting outside. With Sha-256 encryption, you can be sure of security. Here's a video of the prototype in action. 

Developed by Jivitesh Dhaliwal


How It Works

Device is on low power mode before being activated. On being activated, it starts broadcasting that it is available to pair.

Pairing can occur without user's intervention. The app recognizes locks it has paired previously with.

The BLE Lock verifies that only an authentic app can pair then sends a request signal asking for user's data.

Encrypted data received from the user's phone is compared with data generated onboard the device, and the door is unlocked if the data matches.


Why This is Useful


If you've ever waited outside a friend's apartment, or have worried about unrestricted access to your home in your absence, this device is the perfect solution. You can email keys to a friend, limit access times, and never have to worry about forgetting your keys in the car or back home. 




Indoor Navigation For Patients

The Post Graduate Institute of Medical Education and Research, Chandigarh (more commonly known as PGI) is one of India's premier medical institutes. It is also one of the largest in terms of patient influx. For 2015-16, the OPD attendance stood at over 2.4 million patients. An overwhelming majority of patients belong to the economically weaker section of the society and are uneducated. As a result, they face numerous obstacles and discomforts in obtaining healthcare. One of these obstacles is navigating through the different departments at the institute. Doctors write down the name of the department the patients need to go to next, but not having prior experience and being uneducated, they are unable to. A lot of patients have a tough time finding their destination especially since they are ill. 

I was going through Professor Patrick Winston's lecture on A* search on MIT OCW to build my game programming skills and realized that it could be extremely useful for the current problem. Given a set of locations and distances between them, the algorithm could be used to plan the most efficient path. Since the architectural plan of PGI was easily available, the major part of the work was easily accomplished.

The next challenge consisted of figuring out a way to input the patient's locations. Since most patients are too poor to own smartphones, it was imperative that the solution also work on low cost phones (< $20). For this, a set of numbers were assigned to different locations and the patients could send a free SMS to the service with their destination and their current location (as numbers). Instructions to the locations come as both written and symbolic instructions. Patients with smartphones can use interactive maps and bar code scans to find the way.

Still under iteration, this project will grow in scope to include the main problem of helping patients know the time of their appointment. The process of registration at PGI involves the patients getting medical cards made. These cards are then submitted to the designated OPDs. Registrations end at 10AM, after which the cards are collected. However, since the cards are collected early in the morning, the patients form a mental map that their appointments will be scheduled soon. As a result, those that got their cards made later sometimes have to wait over 7 hours. To prevent this, I intend to add a 'time estimate' feature, where registered patients get a text message with the approximate time of their appointments.