Audio Navigation System

for Visually Impaired Public Transit Users

In 2012, during a first year engineering design course, we worked on a two stage project. The first stage was to identify a problem within the city of Toronto, and write an RFP addressing this issue. The second stage was to choose an RFP written by another team in the class and find a solution in order to "Sustainably improve the quality of life of a community-in-need in the city". If you are interested in the first stage, please visit the ER Wait Time. Below I have described the second stage of the project. 


Objective

We were given an RFP written by another team in our class, which asked us to "Improve the way-finding process for the blind people within the TTC subway station". If you would like to see the RFP feel free to contact me.


Product

Even though a traditional solution to this problem is handrail stickers or patterned tiles, we decided to go for a radical design: an audio navigation system and personal device. The device locates the users within the subway platform, and provides them with different alternatives. Then the users select their destination by talking into the microphone. The navigation system then determines the shortest path to the destination and divides that distance by the step-size of the users. This path is sent to the users' personal device as an audio file, which guides them to their destination. 


Presentation

Our team designed a very unique banner for the demonstration day. We left the middle section of the poster blank, and we projected a slideshow on to it. This method allowed us to show slides with information and visuals relevant to the questions that we were asked on the spot; which was very convenient. 

 
ttcposter.png
 

In order to demonstrate how the device worked, we built a model of St. George station, one of the most crowded interchange stations in Toronto. I also wrote a simple algorithm to show how the device would interact with the user in order to guide them towards their destination. The video below is for the case where the user gets off of a train at one side of the station, and would like to leave the station from an exit located on the opposite end of the station. 

 
 

First, the users have to calibrate the device, by inputting their step size. Then they would place a character that we cut out of cardboard, on to the model of the station. Once the source is detected, as in their location is identified within the subway platform, they would be given different options to choose from. Once they choose their path, they would be given directions on how to get to their destination. 


My Team

In the picture below, from left to right: 

  • Sayeh Bayat 
  • Jeffrey Huang
  • Deniz Jafari
  • Parastoo Abtahi 
 
IMG_1289.JPG
 

Design Flaws 

Unfortunately, our design had many flaws: 

1. Lack of accuracy: since the users of the device were visually impaired, the directions provided to them had to be extremely accuracy. Miscalculations could risk the users life by directing them towards the track. 

2. Orientation: since we were using "left", "right", and "go straight" as key words to direct the users to their desired destination, we had to know which way they were facing. In order, to solve this issue, we had to assume that the users were orientated in such a way that their back was facing the door of the train they just exited. This limited the use of the device, because if the user was lost in a station, the directions provided would not be useful for them. 

3. Step size: we chose to use the user's step size, instead of providing them with the distance in units of meters, because we thought it would be more intuitive. There were two fundamental problems about this design decision. Firstly, the user's step size depended on how crowded the station was, and how slowly other people were moving. Secondly, for long distances, such as the one demonstrated in the video, the user was asked to take hundreds of steps, which was not intuitive at all. 

4. Voice recognition: the communication between the user and the device was based on audio signals. One issue was the accuracy of the voice recognition software used, and the other was the background noise, specially during rush hours. 

... and many other problems. Even though we were aware of these limitations, we were very proud of our prototype, we enjoyed working on this project together, and we learned a lot during the process.