The Concept

Snorlax is a smart mattress pad that tracks a user's sleeping patterns and positions

Sleep Analysis

Snorlax is perpetually monitoring the on-mattress accelerometers and pressure sensors for small movements. Snorlax constantly logs sleep information to the cloud, so users can view their sleeping patterns as soon as they wake up.

Position Tracking

Users can calibrate the mattress to recognize up to 4 sleeping positions. The system uses the wide pressure sensor network to sense small changes in weight distribution of the user when sleeping in different positions. The system uses an SVM model to predict current sleeping position.

Position Alert

For patients suffering from sleep disorders or certain medical conditions, sleeping in the wrong position can be harmful. The Position Alert feature allows the user to configure one or more incorrect sleeping positions. The system will make a soft sound to alert them should they end up in the wrong position. The sound is turned off automatically when they correct their position.

Persistent Alarm

Say goodbye to oversleeping and ‘accidentally’ shutting off your alarm. The Snorlax Persistent Alarm system will not stop blasting you until you literally get off the bed.

Our Motivation

Currently, sleeping patterns can be monitored by the average user through various phone apps and wearable devices such as Fitbit, which are all intrusive and use primarily the user's movement to calculate sleeping patterns.

The incentive for this project was to have an entirely non-intrusive system (including wearables), that augments the functionality of current mobile phone applications. By spanning the entire mattress with sensor coverage and having multiple accelerometers at different parts of the mattress, our goal is to broaden the sensor data that phones can collect and have more accurate sleep information.

Competitive Analysis

Eight Sleep

This is our primary competitor, and is a smart mattress pad that does all of the sleep tracking without the need of a wearable device. It has features to heat the mattress pad to desired temperature, communicate with Nest thermostat, control lighting and monitor heart rate and noise levels. This product is still in development and it is unclear if it will have all the features suggested.

Sleep as Android

A sleep tracking Android app that determines the user’s optimal waking point through the use of accelerometers. The main differentiation of our product from this app is that the app requires the phone to be put on the bed in order to get any readings, while our product does not require any conscious effort to do anything different to your normal sleep routine.

RestBed

A smart, actuated bed that tracks sleep data and responds to your sleep. While RestBed provides more comprehensive data on the way someone sleeps, it is very expensive, costing over $200. Our product tracks users' movements, sleeping positions and any noise to determine the sleep data at a smaller cost.

The Requirements

Functional Requirements

Requirement Description
Construction Collect movement and position data using accelerometers and pressure sensors
Communication The mattress pad will connect to the specified WiFi network and communicate / log data to the web server over the internet.
Alarm The smart mattress pad shall have an alarm feature, whereby users can set an alarm time through the web application.
User interface Users will have access to a web application that they can log into and view reports of their sleep activity over the night.
Calibration Users will be able to calibrate their smart bed to work for their body type.
Buzzer Users will be able to set a buzzer to sound when they are sleeping in an incorrect position.

Non-Functional Requirements

Requirement Description
Confortable A poll/survey on the users of the product should yield that at least 95% users do not feel anything out of the ordinary when sleeping on the mattress pad.
Non-invasive The Snorlax mattress pad shall be entirely non-invasive. The user will have no wearable component.
Real-Time Features such as the sleeping position buzzer must operate in real-time. The buzzer must sound within a minute of the correct time.

Components Needed

Hardware

Microphone

We used a microphone to listen to bed movements to determine a user's sleep cycle.

Accelerometer

We also used a three-axis accelerometer to listen to bed movements to determine a user's sleep cycle.

Velostats

We hand-fabricated pressure sensors called velostats. Velostats are conductive fabric that changes in resistance when pressure is applied.

Raspberry Pi

The core of Snorlax is the Raspberry Pi. We used the raspberry pi to read from the various sensors and send the data to a webserver via Wifi.

Analog to Digital Converter

Analog to Digital Converters (ADC)s were essential to Snorlax.

Software & Protocols

Django Web Framework

Developed the web application using the Django Web Framework built in Python.

SQL Database

We stored all the sleep data gathered from the bed in a SQL database.

WIFI and HTTP

Our primary method of communication was through Wifi using HTTP requests.

Design

Sensor Architecture

Communication Pipeline

Snorlax Core

Snorlax Peripheral

Our Team

Pratik Prakash

Did things no one else wanted to do

Ameya Kamat

Lead Calibrator

Harsha Chivukula

PCB Designer

Benson Qiu

Test Subject