Final Project For Better Sleep

December 17, 2017 by Michael Uttmark 14 Comments

It’s that time of year again, and students around the world are scrambling (or have already scrambled) to finish their final projects for the semester. And, while studying for finals prevents many from sleeping an adequate amount, [Julia] and [Nick] are seeking to maximize “what little sleep the [Electrical and Computer Engineering] major allows” them by using their final project to measure sleep quality.

To produce a metric for sleep quality, [Julia] and [Nick] set out to measure various sleep-related activities, specifically heart rate, motion and breath frequency. During the night, an Arduino Nano mounted to a glove collects data from the various sensors mounted to the user, all the while beaming the data to a stationary PIC for analysis and storage. When the user awakes, they can view their sleep report on a TFT display at the PIC base station. Ideally, users would use this data to test different habits in order to get the best nights sleep possible.

Interestingly, the group chose to implement their own heart rate sensor. With an IR transmitter, IR phototransistor and an OP amp, the group illuminates user’s fingers and measure reflection to detect heartbeats. This works because the amount of IR reflected from the user’s finger changes with blood pressure and blood oxygen level, which also happen to change when the heart is beating. There were some bumps along the road when it came to the heartbeat sensor (the need to use a finger instead of the wrist forced them to use a glove instead of a wristband), but we think it’s super cool and totally worth it. In addition to heart rate, motion is measured by an accelerometer and breath is measured by a flex sensor wrapped around the user’s chest.

With all of their data beamed back by a pair of nRF24L01s, the PIC computes the sleep “chaos” which is exactly what it sounds like: it describes just how chaotic the user slept by looking for acyclic and sudden movement. Using this metric, combined with information from breathing and heart rate, the PIC computes a percentage for good sleep where 100% is a great night and 0% means you might have been just as well off pulling an all-nighter. And, to top it all off, the PIC saves your data to an SD card for easy after-the-fact review.

The commented code that powers the project can be found here along with a parts list in their project write-up.

This device assumes that sleeping is the issue, but if waking up if your problem, we’ve already got you covered, aggressive alarm clock style. For those already on top of their sleep, you might want some help with lucid dreaming.

Video of the project explained by [Julia] and [Nick] after the break.

Continue reading “Final Project For Better Sleep” →

AI Watches You Sleep; Knows When You Dream

August 29, 2017 by James Hobson 5 Comments

If you’ve never been a patient at a sleep laboratory, monitoring a person as they sleep is an involved process of wires, sensors, and discomfort. Seeking a better method, MIT researchers — led by [Dina Katabi] and in collaboration with Massachusetts General Hospital — have developed a device that can non-invasively identify the stages of sleep in a patient.

Approximately the size of a laptop and mounted on a wall near the patient, the device measures the minuscule changes in reflected low-power RF signals. The wireless signals are analyzed by a deep neural-network AI and predicts the various sleep stages — light, deep, and REM sleep — of the patient, negating the task of manually combing through the data. Despite the sensitivity of the device, it is able to filter out irrelevant motions and interference, focusing on the breathing and pulse of the patient.

What’s novel here isn’t so much the hardware as it is the processing methodology. The researchers use both convolutional and recurrent neural networks along with what they call an adversarial training regime:

Our training regime involves 3 players: the feature encoder (CNN-RNN), the sleep stage predictor, and the source discriminator. The encoder plays a cooperative game with the predictor to predict sleep stages, and a minimax game against the source discriminator. Our source discriminator deviates from the standard domain-adversarial discriminator in that it takes as input also the predicted distribution of sleep stages in addition to the encoded features. This dependence facilitates accounting for inherent correlations between stages and individuals, which cannot be removed without degrading the performance of the predictive task.

Anyone out there want to give this one a try at home? We’d love to see a HackRF and GNU Radio used to record RF data. The researchers compare the RF to WiFi so repurposing a 2.4 GHz radio to send out repeating uniformed transmissions is a good place to start. Dump it into TensorFlow and report back.

Continue reading “AI Watches You Sleep; Knows When You Dream” →

“The Alarm Clock Ate My Duvet Cover, That’s Why I’m Late!”

August 3, 2017 by Dan Maloney 23 Comments

Some people just won’t wake up. Alarm clocks don’t cut it, flashing lights won’t work, loud music just becomes the soundtrack of an impenetrable dream. Maybe an alarm clock that rudely yanks the covers off the bed will do the trick.

Or not, but [1up Living] decided to give it a go. His mechanism is brutally simple — a large barrel under the foot of the bed around which the warm, cozy bedclothes can wind. An alarm clock is rigged with a switch on the bell to tell an Arduino to wind the drum and expose your sleeping form to the harsh, cold world. To be honest, the fact that this is powered by a 2000-lb winch that would have little trouble dismembering anyone who got caught up in the works is a bit scary. But we understand that the project is not meant to be a practical solution to oversleeping; if it were, [1up Living] might be better off using the winch to pull the bottom sheet to disgorge the sleeper from the bed entirely.

Something gentler to suit your oversleeping needs might be this Neopixel sunrise clock to coax you out of bed naturally.

Continue reading ““The Alarm Clock Ate My Duvet Cover, That’s Why I’m Late!”” →

Keep The Peace In The Bedroom With A Snore Stopping Sleep Mask

February 22, 2017 by Dan Maloney 29 Comments

Despite what my wife says, I have absolutely no evidence that I snore. After all, I’ve never actually heard me snoring. But I’ll take her word for it that I do, and that it bothers her, so perhaps I should be a sport and build this snore-detecting vibrating sleep mask so she can get a few winks more.

Part wearable tech and part life hack, [mopluschen]’s project requires a little of the threadworker’s skill. The textile part of the project is actually pretty simple, and although [mopluschen] went with a custom mask made from fabric and foam shoulder pads, it should be possible to round up a ready-made mask that could be easily modified. The electronics are equally simple – an Arduino with a sound sensor module and a couple of Lilypad Vibe boards. The mic rides just above the snore resonating chamber and the vibrators are right over the eyes. When your snore volume exceeds a preset threshold, the motors wake you up.

Whether this fixes the underlying problem or just evens the score with your sleep partner is debatable, but either way there’s some potential here. And not just for snore-correction – a similar system could detect a smoke alarm and help rouse the hearing impaired. But if the sewing part of this project puts you off, you should probably check out [Jenny List]’s persuasive argument that sewing is not just for cosplayers anymore.

ESP8266 Lullaby

November 11, 2016 by Al Williams 22 Comments

The ESP8266 is certainly a versatile device. It does, however, draw a bit of power. That isn’t really surprising, though, since you would expect beaming out WiFi signals to take a little juice. The trick is to not keep the device on all the time and spend the rest of the time in deep sleep mode. [Marco Schwartz] has a good tutorial about how to use this mode to run for “years” on a battery.

[Marco] notes that even using a 2500 mAh LiPo battery, he only gets about 30 hours of operation without sleep. By putting the chip in sleep mode, the current consumption drops from about 88 mA to just over 8 mA. That’s still high, though, because the board has a power LED! By removing a jumper or cutting a trace (depending on the board), you can drop the current draw to about 0.08 mA (80 uA) when it’s not doing anything.

Continue reading “ESP8266 Lullaby” →

10,000 Lumen Sunrise Lamp Curses The Darkness

September 2, 2016 by Richard Baguley 26 Comments

Some of us need a bit of help to get up in the mornings. This can come in the form of a sunrise lamp, which simulates the light of the sunrise to fool our poor sleep-deprived brains into waking up in the depths of winter. [Lincoln Johnson] found the ones he tried were not bright enough to wake him, so he decided to build his own: a 10,000-lumen monster that can wake him up from across the room.

It uses a lot of LEDS: 5 meters of 5630 LED strip, which pulls a circuit bending 72 watts when running at full blast. This monstrosity is powered by an Arduino Pro, which is programmed to slowly increase the brightness over a period of 30 minutes, thus simulating the sunrise. It uses PWM control to fade the LEDs, and also includes a dot matrix display to show the time. Honestly, if you are able to sleep through this thing blasting your eyes, you are probably dead.

Hackaday Prize Entry: Powering A Pi From A Battery

May 11, 2016 by Moritz Walter 28 Comments

Knocking a microcontroller into sleep mode and waking it up on demand or in intervals is common practice in many low power applications, enabling devices to stay in operation for years on a single coin cell battery. Since there are tons of applications where you might want to do similar things with a Raspberry Pi, [Patrick Van Oosterwijck] created the LiFePO4wered/Pi. The module that snaps on to eight GPIO pins of a Pi, extending it by a long life LiFePO4 battery, a charging regulator, and a proper power management. Obviously, it also makes a great UPS.

[Patrick] realized this project by expanding his already available and equally useful LiFePO4wered/USB charging regulator module by a low power MSP430G2131 microcontroller and a load switch. A daemon on the Raspberry Pi speaks to the module over I2C, allowing you to schedule a wake-up timer, let your Pi autoboot after a power outage or just read out the current battery voltage through a command line tool. Once the Pi is safely shut down, the microcontroller will also go to sleep, resulting in a standby current of 8 uA for the whole system. Together with the 500 mAh LiFePo4 cell, that’s theoretically low enough to send your Pi-ncess into a seven-year-long sleep.

LiFePO4wered/Pi is not only good for sleeping, though. [Patrick’s] runtime tests show, that the 500 mAh cell will power a Raspberry Pi Zero and a WiFi dongle for about two hours. Because the Raspberry Pi and many USB peripherals won’t complain when only 3.2 V are present on the VBUS, [Patrick] was able to squeeze out even more runtime by dismissing the boost converter from the design and driving the Pi directly from the battery voltage. If that worries you, you can either read a detailed explanation on why that works so well or just have a look at the more compliant 5 V version.

Eventually, [Patrick] used his module to create a Raspberry Pi time-lapse camera. A little script lets the Pi take a picture on boot up, set a wake-up timer and go back to sleep again. Safely enclosed in a waterproof electric box and deployed into the wild, the camera took 120 pictures on a single charge.

We’re sure the module will find it’s way into many cool projects and we’re counting the hours until we can get one in [Patrick’s] tindie store. Until then, enjoy the time-lapse video: