Project I-Visit is an IOT solution for the blind. It aims to help them make informed choices about everyday travel, improve self navigation experience and, while doing so, make them more confident about visiting unknown places independently and add more meaning to the blind community's social media interactions.

I-Visit is based upon the simple fact that every journey through a path makes it less unknown and that this travel need not be by the same person. It improves decision making, for travel, by leveraging the knowledge of the path gained when a blind person traverses through it. The time of travel, the number of static and dynamic obstacles encountered, the blindfriendliness and the ability to travel alone, are all recorded. So the next time, when another blind person wishes to travel though the same path, s/he receives valid feedback. Furthermore, the more a path is taken by blind users, the better and more accurate the feedback becomes. The three components that make up I-Visit are:


Since time immemorial, one of the most important items that was carried on any journey was the knowledge provided by those who had taken the same path previously. If no previous experience was available at hand, the probability of the journey being difficult increased automatically. So did the chances of being over or under prepared for the journey. No wonder then that humans have valued this resource, 'knowledge of the unknown', more than any other resource on this planet. Cut to the world of a blind person, self navigation in unknown environments remains one of the biggest, every-day life, challenges. Many modern day technologies, including smart wearables, have been successful in reducing this navigation discomfort. But what about decision making before heading out? Yes, you would be thinking in the right direction if Google Maps comes to your mind. But, would Google Maps be sufficient for a blind person to independently decide whether to visit a place alone or to take someone along, or provide an answer to whether the journey to this place is blind friendly or not. We agree that Google Maps is not meant to answer these questions, but what about Expected Travel Time? This information, provided by Google, is generally of more value to sighted users. So why not create something that can help answer these questions and probably much more?
I-Visit's solution is inspired by two common phenomenon; people love to help each other, more through social platforms and that bats personify 'being differently abled'.
People love to help         Add sign         Bat


Remember to use combination of descriptions, photos, and figures



Technical Components

There are three main technical components:

  1. Bat-Pi

    The embedded system works as a bluetooth server in this project. After receiving the ’start’ command from the smartphone, the server closes the current socket and start scanning the distance sensors placed in 3 different directions to detect the obstacles along the path. After the user arrives the destination, BAT PI will send the numbers of static and moving obstacles back to the smartphone respectively via bluetooth.To distinguish the static obstacles from the moving ones, we assume the distance passed in one second is about 1 meter. Then we set the bound value, 3 meters, to the sensor facing the front direction. Once the value returned by the sensor exceed 3 meters, the system will scan once more in a scanning period to see if the obstacle is still in the alarming area or not. For most moving obstacles, the sensor will not detect it 2 times in this process. For almost all static obstacles (e.g. walls and trees) and some kinds of moving obstacles (e.g. a passenger moving very slowly), the sensor will detect them twice in this process.
    Bat-Pi    Sensor Algo

  2. I-Visit App

    The I-Visit Android application is a blind friendly application and, therefore, does not require visual selection. It allows the user to enter source and destination via speech and uses Android's speech to text conversion feature. All instructions for the user are also audio instructions by default. The app is responsible for triggering the the sensors on when the user begins her journey and requesting the obstacle count, from the sensors, at the end of the journey. During the journey, it is responsible for relaying real time Google Map navigation directions. In the end, it accumulates all the data recorded during the journey and sends it to the cloud. Thus, the app acts like a gateway between Bat-Pi and the Cloud based data storage.

  3. The Cloud

    The cloud (an EC2 instance) hosts the IVisit database where all trip information is stored and analysis of data takes place before travel suggestions are returned.


The system prototype works in the following way:

Flow Diagram


  • Features of the Wish List accomplished


  • Potential Challenges Overcome


  • >

    Stage Milestones Crossed

  • Improved Feedback

Our Team

Sanjmeet Abrol

Sanjmeet comes from New Delhi, India and is currently pursuing a Masters in Data Science at Columbia. She was previously working as a Smart Grid Solutions Architect for Toshiba. A passionate debater and dancer, she is also an environment enthusiast who looks forward to apply her learnings in creating solutions that shall help regulate the decay of Planet Earth.
LinkedIn Profile

Yunqing Jiang

Yunqing comes from Shanghai, China, and is a graduate student of Electrical Engineering Department, Columbia University. She is enthusiastic about voluntary, has participated as volunteer in Shanghai Expo and Beijing Marathon. She is interesting in starting her career from network engineering
LinkedIn Profile

Keyi Yang

Keyi comes from Shijiazhuang, China.She got bachelor's degree in Beijing University of Posts and Telecommunications and is a current master student in Columbia University, Electrical Engineering Department. Keyi is interesting in programming and algorithm, and also have experience of taking part in the related competitions


Sanjmeet Abrol: sanjmeetabrol@gmail.com
Keyi Yang: keyi.yang@yahoo.com
Yunqing Jiang: annejiangyq@outlook.com

Columbia University Department of Electrical Engineering
Class Website: Columbia University EECS E4764 Fall '16 IoT
Instructor: Professsor Xiaofan (Fred) Jiang