Project :: ECE 445 - Senior Design Laboratory

Project

#	Title	Team Members	TA	Documents	Sponsor
34	Voice Biometrics Lock	Bella Chen Lixin Guo Zaki Khan	Anand Sunderrajan	design_document1.pdf design_document2.pdf final_paper1.pdf proposal1.pdf
	Team: [Lixin Guo][lixing2], [Bella Chen][aotingc2], [Zaki Khan][zakik2] # Problem Losing your keys is not a pleasant experience. Key locks are also at the risk of being picked. Electronic locks (like the [Nest x Yale Lock](https://store.google.com/us/product/nest_x_yale_lock) )are becoming more prevalent for modern homes, but they all use passcodes for entry, which can be stolen. Physiological traits like your voice are unique and can be used as a secure way for verifying your identity. Some of the smart locks on the market may work with voice recognition like Google Home, and you can lock the door from the inside with one command. Yet when you enter from the outside, a passcode is still needed because home devices would not recognize who should not be allowed into the door. # Solution Overview We plan to design a lock that works with a real-time speech recognition app to distinguish the voice of the authorized user(s). The user could set a short command to lock or unlock the door, so the lock activates only when the exact command is spoken by the correct person. The lock would be controlled using a Bluetooth receiver that is triggered by a pulse sent from the phone. The benefit of using voice recognition is that the user can unlock the door a short distance away by speaking through his or her phone. We would use a fingerprint sensor as a backup method to verify identity. # Solution Components [Android Application] - The purpose of this component is to let the users interface with the lock. It would let the users record a command that they would like to use to unlock/lock the door. It would also record the serial number of the device they are using. Then when the person wants to unlock the lock, it would use the command and the serial number of the device being used to verify identity and send a bluetooth signal to the lock to unlock. When the user wants to lock the device, the user would push a button to lock the lock. Here it is not necessary to have voice recognition since we are just locking the device and there is no security concern with this. [Bluetooth receiver circuit] - We would design a PCB that allows a Bluetooth device to control a motor to lock or unlock the door. This circuit would be composed of an oscillator that checks the incoming signal and a flip-flop that would set voltage high momentarily to turn the motor in either direction. Once the application determines the action to lock or unlock, a pulse would be sent from the phone to trigger the Bluetooth device, which turns on the motor. [Fingerprint sensor] - The fingerprint sensor should only be used when a failed voice recognition attempt has been made through a device associated with the user profile, or be required as a second verification step if the user decides to setup 2FA. [LED indicator light] - As a visual aid, a red light would blink when the door is locked and a green light would indicate the door is unlocked. When an attempt to unlock the door fails, the red light would blink quickly. # Criterion for Success * We want the device to be able to control the lock and unlock of the door through the user’s voice based on the speech recognition function. * The app should be able to recognize an authorized device through some form of device identification. * The status of the lock should be accurately displayed in the app and by the LED indicator light. * Fingerprint matching should successfully serve as a backup verification method * Simple UI for easy navigation and user experience

Title

Team Members

Documents

Sponsor

Voice Biometrics Lock

Bella Chen
Lixin Guo
Zaki Khan

Anand Sunderrajan

design_document1.pdf
design_document2.pdf
final_paper1.pdf
proposal1.pdf

**Team:** [Lixin Guo][lixing2], [Bella Chen][aotingc2], [Zaki Khan][zakik2]

# Problem

Losing your keys is not a pleasant experience. Key locks are also at the risk of being picked. Electronic locks (like the [Nest x Yale Lock](https://store.google.com/us/product/nest_x_yale_lock) )are becoming more prevalent for modern homes, but they all use passcodes for entry, which can be stolen. Physiological traits like your voice are unique and can be used as a secure way for verifying your identity. Some of the smart locks on the market may work with voice recognition like Google Home, and you can lock the door from the inside with one command. Yet when you enter from the outside, a passcode is still needed because home devices would not recognize who should not be allowed into the door.

# Solution Overview

We plan to design a lock that works with a real-time speech recognition app to distinguish the voice of the authorized user(s). The user could set a short command to lock or unlock the door, so the lock activates only when the exact command is spoken by the correct person. The lock would be controlled using a Bluetooth receiver that is triggered by a pulse sent from the phone. The benefit of using voice recognition is that the user can unlock the door a short distance away by speaking through his or her phone. We would use a fingerprint sensor as a backup method to verify identity.

# Solution Components

**[Android Application]** - The purpose of this component is to let the users interface with the lock. It would let the users record a command that they would like to use to unlock/lock the door. It would also record the serial number of the device they are using. Then when the person wants to unlock the lock, it would use the command and the serial number of the device being used to verify identity and send a bluetooth signal to the lock to unlock. When the user wants to lock the device, the user would push a button to lock the lock. Here it is not necessary to have voice recognition since we are just locking the device and there is no security concern with this.

**[Bluetooth receiver circuit]** - We would design a PCB that allows a Bluetooth device to control a motor to lock or unlock the door. This circuit would be composed of an oscillator that checks the incoming signal and a flip-flop that would set voltage high momentarily to turn the motor in either direction. Once the application determines the action to lock or unlock, a pulse would be sent from the phone to trigger the Bluetooth device, which turns on the motor.

**[Fingerprint sensor]** - The fingerprint sensor should only be used when a failed voice recognition attempt has been made through a device associated with the user profile, or be required as a second verification step if the user decides to setup 2FA.

**[LED indicator light]** - As a visual aid, a red light would blink when the door is locked and a green light would indicate the door is unlocked. When an attempt to unlock the door fails, the red light would blink quickly.

# Criterion for Success

* We want the device to be able to control the lock and unlock of the door through the user’s voice based on the speech recognition function.

* The app should be able to recognize an authorized device through some form of device identification.

* The status of the lock should be accurately displayed in the app and by the LED indicator light.

* Fingerprint matching should successfully serve as a backup verification method

* Simple UI for easy navigation and user experience

Featured Project

I spoke with a TA that approved this idea during office hours today, and they said I should submit it as a project proposal.

# Habit-Forming Toothbrush Stand

Team Members:

- Rahul Vasanth (rvasant2)

- Quinn Andrew Palanca (qpalanc2)

- John Jung-Yoon Kim (johnjk5)

# Problem

There are few habits as impactful as good dental hygiene. Brushing teeth in the morning and night can significantly improve health outcomes. Many struggle with forming and maintaining this habit. Parents might have a difficult time getting children to brush in the morning and before sleep while homeless shelter staff, rehab facility staff, and really, anyone looking to develop and track this habit may want a non-intrusive, privacy-preserving method to develop and maintain the practice of brushing their teeth in the morning. Keeping track of this information and but not storing it permanently through a mobile application is something that does not exist on the market. A small nudge is needed to keep kids, teenagers, and adults of all ages aware and mindful about their brushing habits. Additionally, many tend to zone out while brushing their teeth because they are half asleep and have no idea how long they are brushing.

# Solution

Our solution is catered toward electric toothbrushes. Unlike specific toothbrush brands that come with mobile applications, our solution applies to all electric toothbrushes, preserves privacy, and reduces screen time. We will implement a habit-forming toothbrush stand with a microcontroller, sensors, and a simple LED display that houses the electric toothbrush. A band of sensors will be wrapped around the base of the toothbrush. Lifting the toothbrush from the stand, turning it on, and starting to brush displays a timer that counts seconds up to ten minutes. This solves the problem of brushing too quickly or losing track of time and brushing for too long. Additionally, the display will provide a scorecard for brushing, with 14 values coming from (morning, night) x (6daysago, 5daysago, . . . , today) for a "record" of one week and 14 possible instances of brushing. This will augment the user's awareness of any new trends, and potentially help parents, their children, and other use cases outlined above. We specifically store just one week of data as the goal is habit formation and not permanent storage of potentially sensitive health information in the cloud.

# Solution Components

## Subsystem 1 - Sensor Band

The sensor band will contain a Bluetooth/Wireless Accelerometer and Gyroscope, or Accelerometer, IR sensor (to determine height lifted above sink), Bluetooth/Wireless connection to the microcontroller. This will allow us to determine if the electric toothbrush has been turned on. We will experiment with the overall angle, but knowing whether the toothbrush is parallel to the ground, or is lifted at a certain height above the sink will provide additional validation. These outputs need to be communicated wirelessly to the habit-forming toothbrush stand.

Possibilities: https://www.amazon.com/Accelerometer-Acceleration-Gyroscope-Electronic-Magnetometer/dp/B07GBRTB5K/ref=sr_1_12?keywords=wireless+accelerometer&qid=1643675559&sr=8-12 and individual sensors which we are exploring on Digikey and PCB Piezotronics as well.

## Subsystem 2 - Toothbrush Base/Stand and Display

The toothbrush stand will have a pressure sensor to determine when the toothbrush is lifted from the stand (alternatively, we may also add on an IR sensor), a microcontroller with Bluetooth capability, and a control unit to process sensor outputs as well as an LED display which will be set based on the current state. Additionally, the stand will need an internal clock to distinguish between morning and evening and mark states accordingly. The majority of sensors are powered by 3.3V - 5V. If we use a battery, we may include an additional button to power on the display (or just have it turn on when the pressure sensor / IR sensor output confirms the toothbrush has been lifted, or have the device plug into an outlet.

# Criterion For Success

1. When the user lifts the toothbrush from the stan and it begins to vibrate (signaling the toothbrush is on), the brushing timer begins and is displayed.

2. After at least two minutes have passed and the toothbrush is set back on the stand, the display correctly marks the current day and period (morning or evening).

3. Track record over current and previous days and the overall weekly record is accurately maintained. At the start of a new day, the record is shifted appropriately.

Project Videos

Habit Forming Toothbrush Stand