Engineering for Access 2020 shortlist: Krisian Chhatralia

My entry is Smart Glasses for the blind. These glasses are equipped with a camera and computer vision (Optical Character Recognition – OCR and real time object detection) to detect the world around the user. It will also contain GPS, a microphone and speaker so users can ask “Where am I” if lost and set a destination before leaving their house and provide real time directions when travelling towards their destination. The glasses will relay the information back to the user when required. This is useful in situations where braille is not present, it will also prevent an overload of information of the world around them.

This product will empower users as it promotes independence and freedom by allowing the blind to spend more time outside without the need of a chaperone or to learn a route to their destination before travelling by themselves.

For example, users can set a destination on their app and it will give real time navigation via the speakers near the temples of the glasses. They will no longer need to feel worried about taking out their phone in public as the glasses will provide navigation and they can use a white cane to understand where to turn as normal. Users can go shopping by themselves and understand which product they are holding, the price and which aisle they are in without the need of a chaperone or braille. If lost, users can ask the glasses “Where am I?” and set destinations without the need of taking out their phone. Furthermore, they will no longer have to spend time learning a route with a chaperone before feeling comfortable leaving by themselves to reach a certain destination. They will also be able to pick up and read any product or book in front of them as the speakers by the temples will repeat information back to the users. This can be on command so as not to overload the user with information about the world.

The glasses could be connected to a smartphone, allowing apps could be created for the glasses specifically. These apps could include face recognition so that users can identify a friend in a crowd. This may also prevent the user from becoming scared of others approaching them as they will be notified of this and they will expect it. Travelling apps could be created for users who want to take tours around cities. The glasses will recognise landmarks and other sightseeing attractions and could give out information about the history of the city, landmarks and also provide event information if users wish to visit concerts etc, essentially working like a VR tour. Furthermore, updates could be made to the glasses to understand if a user is under attack or if a theft is taking place. Since it is connected to an app, an emergency signal can be created to notify the emergency services. (Also mentioned further below).

The glasses could also contain a barcode scanner so they can read the price of certain products where pricing is harder to find or if the text is harder to read, such as clothing or handwritten prices. Since the glasses could be connected to a smartphone and contain certain sensors, the potential is limitless. As long as apps can be developed to make use of the product, users can benefit from a multitude of apps, provided they make use of the sensors within the glasses. The problem surrounding this product is the weight and size of the glasses.

However, since this product contains certain sensors, the improvements are all software based.

This product is designed to help the blind at any age. However, it would be most beneficial to those who travel alone and are at an age where they can live independently. This will depend on the personal development of the user, however, this product may be suitable for teenagers or even younger if parents wish to provide the child a sense of the world around them and how to navigate it. It will also allow younger children to feel more independent as they could pick up a book or object and understand what is in front of them as the camera will read for them and speak the information back to the child.

I made acquaintances with a blind man at university at the start of my third year. I told him I am a black belt in karate and he asked me if I could teach him how to defend himself as it might be a useful skill in a time of need. Further into the conversation I asked if he would like to meet me at another location. He agreed, however he told me he needed to learn the route in order to meet me there in the future. I then began to think about safety for the blind and how they must feel taking out their phone and other possessions in public as well how to improve navigation. I believe that the blind may fear theft whilst out in public as well as other potential crimes. Below will explain the development of my idea to help improve safety and navigation as well as object recognition.

After speaking to my blind friend my first idea was to develop a “smart” white cane. This would allow users to set a destination before leaving to prevent them from taking out their phone or feeling lost. Users can either feel vibrations from haptic feedback in the handle or hear directions from a speaker in the cane to turn left or right. If haptic feedback was chosen, the cane would have had to understand whether the user was left or right handed and also where the fingers and palm was placed on the cane. In the case of a right handed user, vibrations in the palm would indicate turning right and vibrations in the palm would indicate turning left. Vibrations would increase as the user would come closer to the turning point – similar to a satnav, but with vibrations. However, to my dismay, this product has already been created under the name of WeWalk. I did not feel comfortable submitting this idea as I could not think of an improvement to their design.

I then turned to other ideas. Currently, I am taking a computer vision module at university. This birthed the idea of using a camera on a pair of glasses to recognise words. The glasses would be connected to a speaker or earplugs that would be used to read to the user. However, again to my dismay, this product has also been created under the name of Orcam.

I have thought of many other ideas, however eventually I decided to submit the smart glasses as I believe it is an improvement to the above products. Whilst the smart white cane can

provide navigation and GPS, it cannot provide object recognition or character recognition. Similarly, Orcam is only used for reading purposes. My idea improves upon the two as it allows the user to use both features under one product and I believe it has great potential. The glasses could be connected to a smartphone, allowing apps could be created for the glasses specifically. These apps could include face recognition so that users can identify a friend in a crowd. This may also prevent the user from becoming scared of others approaching them as they will be notified of this and they will expect it. Travelling apps could be created for users who want to take tours around cities. The glasses will recognise landmarks and other sightseeing attractions and could give out information about the history of the city, landmarks and also provide event information if users wish to visit concerts etc.
Furthermore, updates could be made to the glasses to understand if a user is under attack or if a theft is taking place. Since it is connected to an app, an emergency signal can be created to notify the emergency services. However, this may require an accelerometer or gyroscope in order to understand motion.

The function of the glasses has been explained above, however, I will describe the sensors and parts required in order to enable the function.

The camera on the glasses will be used for object and character recognition. This could be placed in a similar position to google glass along the sidearms of the glasses to balance the weight and prevent the glasses from slipping down the nose. Other sensors that must be included are GPS for navigation, a microphone for the user input, a speaker for navigation and object recognition and a bluetooth chip and memory. Since the glasses are connected to a smartphone, the software on the phone can process the computer vision and navigation. Potential sensors include: A gyroscope or accelerometer for motion detection in the case of an attack or crime.

However, a camera could also be used to detect a potential crime. Furthermore, the microphone will need a filter to remove the majority of noise from the outside world. The speaker must be loud enough to transmit noise to the ear but not so loud as to cause pain. A volume controller may be necessary. Lastly, the design of the glasses must balance the weight, heat and size of the glasses as well as understand whether the user is wearing the glasses. This will need a pressure sensor under the arms of the glasses, and could have a clip to prevent them falling off. The speaker could also signal to the user where the glasses are in the case of the glasses falling off or to locate the glasses.

• Camera – ISOCELL S5K2L1 £13.95 [Iparts4u]
• Electret Microphone (2 for noise cancellation) – £1.25 (1), £2.50 (2) [Farnell]
• Speaker (2 for each ear) – up to £26.56 for 2 [mouser]
• GPS – £11 – £500 [mouser]
• Microcontroller/microprocessor (audiovisual) – £22.82 [texas instruments]
• Bluetooth – Bluetooth Modules (802.15.1) EMB1061 BLE Module £2 [mouser]
• Accelerometer – Kionix £5.48 [mouser]
• 3D printed housing (filament)- up to £50

Total = £134.31

This price may not reflect the full price, as the quality of the product depends on the components. The price increases with quality, however, this may seem like a reasonable estimate – perhaps on the lower end.

Note* Orcam sells for $2500/$3500, taking a 15% – 30% cost, orcam could cost between $375 to $750 for the lower priced model and $525 to $1050 for the higher priced model. However, in any case this similar product falls below the £5000 prototype budget.