Technical Proposal for Smart Cane

 

Benjamin Nguyen, Jie Yu, Mariam Khan,

Rezwana Kabita, Sanjidah Abdullah

English 21007 Writing for Engineering

Professor Maryam Alikhani

City College of New York

November 21, 2018

---

Table of Contents

Executive Summary ……………………………………………………………………. 2

Introduction.………………………………………………………………………………. 2

Statement of the Problem ….…………………………………………………………… 3

Objectives …………………………………………………….…….……………..………. 4

Plan of Action ……………………………………….…………….………..…………….. 4

Management Plan …………………………………….…………….………..………….. 6

Conclusion …………………………………….…………….………..……………………. 7

References ……………………………………………….…….…………….………….. 8

---

Executive Summary

The following is a proposal for the Smart Cane, a technical and responsive smart walking cane that allows for the blind to better navigate themselves in public settings such as grocery stores, parks, and other locations that can be unfamiliar. Our goal is to be able to design a technological walking cane for the blind to navigate through unfamiliar areas. The walking cane will have a sensor on the rear end which will respond to touch or once the cane comes into contact with an object. Once the sensor is activated, it will activate the camera attached to the walking cane and the camera will use Image Processing tools to identify the object. After the object is identified, a speaker or bluetooth headset will verbalize what that object is. The project will require a knowledge of both the hardware components of the camera, speaker, bluetooth headset, and how they will correspond to Image Processing software.

 

Introduction

Blindness is the inability to see due to an injury, disease, or a congenital condition. This eye condition affects about 36 million people worldwide, while there are a total of 1.3 billion people who suffer from various forms of visual impairment altogether (World Health Organization, 2018). There are many types of visual impairments such as cataracts and diabetic retinopathy that can be treated through surgery. Although, now, technology plays a crucial role in the development of products to help treat blindness and even restore vision. For instance, Second Sight created a bionic eye brain implant called Orion to restore some vision to people who are blind due to causes other than preventable or treatable conditions. This solution costs about $150,000 and is available in 15 centers in the United States (Mullin, 2017). However, for those who prefer an alternative to a biomedically engineered product, an external piece of technology such as WeWALK’s Smart Cane is available. The Smart Cane is a technological cane that alerts visually-impaired people if there are any obstacles in front of them through voice alert and vibration, which could allow them to walk with less accident (Wahab-Mutalib et al., 2011). Thus, through the advancement of modern technology, we are able to use viable solutions to help treat and even cure blindness.

 

Statement of the Problem

The problem we are addressing is the need for more accurate sensing for the blind to live a more comfortable and inclusive life. Blind people have been forced to use basic canes that have a purpose of detecting a physical object near them by having the cane come into physical contact with the nearby object. However, the blind have to rely on their sense of touch to figure out what that object actually is. With time, many have figured out how to live with their canes and sense of touch, but what about the people that have yet to master the skill of understanding the physical world simply by touch? There have been attempts to solve this issue using products such as the bionic eye and current existing but lacking smart canes. However, these solutions are drastically expensive and their accessibility is limited. To note, the current existing smart cane, for example, does not alert users of specificity of the object and the bionic eye is not universal for all types of blindness. The only other alternative is learning the aforementioned skill. Learning this skill takes years of experience and even then, performing basic daily tasks such as grocery shopping can be difficult without the aid of another person. In order to achieve a better way of navigating life with blindness, an alternative solution is needed.

 

Objectives

Our objective is to enhance the lives of those affected by blindness by providing a smart cane that helps detect objects using a camera, eliminating the need of another person. With visual detection, the specificity of a product will no longer be a mystery for the blind. This will help visually-impaired people with basic daily tasks such as shopping and navigating through unfamiliar areas. Additionally, we want to make this solution accessible to all, regardless of the form of visual impairment they have and their financial status. Rather than creating a limitedly accessible product, we would research affordable alternatives for the equipment while still retaining the high level of quality. Our research would be implemented into our product so that it will improve the quality of life for the blind.

 

Plan of Action

The main course of action we want to take is in doing research on features that are yet to exist so that we can incorporate it to our version of the smart cane. We plan on building the software component of the smart cane first, and then moving onto building the hardware component. Once both components are developed, we will integrate them and run multiple tests to determine the most efficient version of the smart cane through trial and error.

For the software component of our smart cane, we want to achieve our objective by researching object detection using cameras and develop effective image processing programs to build the most efficient form of our smart cane.  In order for our research to be used correctly, we plan on integrating it with the sensors we will be using. The image detection software will take into account the colors, edges, logos, and many other visually distinct features on objects and send this information to a computer that will process it. The “first step in image classification is to simplify the image by extracting important information and leaving out the rest” (Maruti Tech Labs, 2017). After that, the software must begin “segmenting the image and analyzing each and every segment for pixel differences” (TestingWhiz, 2014). While the information is sifted through by the image processing software, it will compare the information to online image databases with an algorithm that helps find visual similarities. After comparing the information given with the databases, the computer will choose the keyword that best matches which will ultimately be sent back to the cane. This keyword is then verbalized to the user of cane through the attached speakers.

In regards to the hardware component, it must coalesce well with the Image Processing programming software and be able to relay information to each respective component so that the cane can come together as a cohesive unit. The sensor must be able to register touch and once it activates, it must be relay that information to the camera to activate. Once the camera activates, it is imperative that the Image Processing software activates and register the image that is taken by the camera to process it via an online database. In order to power the entire unit, there must also be a convenient power supply that is light, efficient, and portable as to not hinder the walking experience of the user. The power supply must be able to power all of the electronic components swiftly and effectively. And lastly, the cane itself will also have to be extremely durable and retractable in order to suit the daily needs and convenience of a visually impaired person. To do this, the cane will be composed of a light and durable carbon fiber, thus adding to the integrity, and overall composition of the cane.

 

Management Plan

Our plan is to upgrade the existing cane with a more modern and affordable smart

cane for people from all walks of life. It can be made within a very short period of time. Specifically, the software component could be created within three to four months while the hardware component could be engineered within two to three days. The following table lists all expected costs of the supplies necessary for building our updated smart cane.

These prices are based on consumer prices. Prices may vary and can be less expensive if the manufacturing company buys them in bulk (Huang, 2014).

Item	Cost
Cane (Walking Stick Skeleton)	$15.00
Ultrasonic sensor	$15.00
Camera (360 rotation)	$30.00
Bluetooth	$10.00
headphone and speaker	$15.00
Power supplies and battery pack	$10.00
Insulated wiring	$15.00
Total	$110.00

 

The software consists of two parts which include object detection and image processing. Three Artificial Intelligence experts can complete this research within three to four months. The salary for three AI experts would approximately be $80,000 (Salary: Artificial Intelligence, n.d.). Additionally, raw materials for research would cost around $10,000.

 

Conclusion

In conclusion, our updated SmartCane has numerous advantages over the original SmartCane and other existing devices for the blind. For instance, it can be easily popularized since the traditional SmartCane can only detect obstacles in the way, and the bionic eyes can only be applied to certain blind conditions. However, our smart cane not only has the directing function but can also help all visually-impaired people discover the world through object recognition. Furthermore, our updated smart cane is very practical in that all the components needed for the smart cane are relatively cheap and accessible. Additionally, the most important part of our smart cane, the object recognition system, can be conducted using the technology of computer vision algorithms which can be optimized and integrates into our software. This will save much money and time in developing our product.

Nevertheless, the perfect product does not exist. There are a few limitations for the smart cane we designed. First, the recognition may fail due to the fact that the computer vision algorithm only recognizes an object by analyzing its shape with thousands of similar images in the database and providing the most possible output. For example, it may confuse a sculpture with a real object. We believe that by enhancing the database of our software, the smart cane will be able to identify most daily-life objects correctly. Additionally, the current computer vision algorithm is only able to recognize objects in an image, however, it cannot differentiate the correct relationships between objects in an image. If computer vision can be developed to analyze pictures as our brains do, we will be one step closer in creating inclusive lives for the blind.

 

 

 

References

 

Huang, W., Mcnamara, H., Pasarkar, A., Rizzo, R., & Molodan, D. (2014). Smart cane. Retrieved from https://soe.rutgers.edu/sites/default/files/imce/pdfs/gset-2014/SmartCaneFinal

Maruti Tech Labs. (2017, September 5). What is the Working of Image Recognition and How it is Used? Retrieved from https://www.marutitech.com/working-image-recognition/

Mullin, E. (2017, September 18). A company is reviving efforts to make a bionic eye brain implant for the blind. Retrieved from https://www.technologyreview.com/s/608844/blind-patients-to-test-bionic-eye-brain-implants/

Salary: Artificial Intelligence. (n.d.). Retrieved from https://www.glassdoor.com/Salaries/artificial-intelligence-salary-SRCH_KO0,23.html

TestingWhiz. (2014, September 7). How Image Comparison Works in Website Testing. Retrieved from https://www.testing-whiz.com/blog/how-image-comparison-works-in-website-testing

Vision impairment and blindness. (2018, October 11). Retrieved from https://www.who.int/news-room/fact-sheets/detail/blindness-and-visual-impairment

Wahab, M. H., Talib, A. A., Kadir, H. A., Johari, A., Noraziah, A., Sidek, R. M., & Mutalib, A. A. (2011). Smart Cane: Assistive Cane for Visually-impaired People. International Journal of Computer Science Issues,8(4), 2nd ser., 21-27. Retrieved from https://arxiv.org/ftp/arxiv/papers/1110/1110.5156.pdf.