Inspiration
We wanted to create something useful and amazing so we came up with BykeSafe. BykeSafe is a wearable bike vest that has integrated turn, brake, and hazard signals.
How It Works
We used an android app to send audio signals to a passive low and high pass filter. once filters we used different sound frequencies to turn on our LED banks on the vest. We then hooked up a Myo and programmed different gestures into the android app. So when you do your normal hand signals you also turn on the signals
Challenges
Android App Newbies
None of us had created an Android application before, nor even attempted it. As such, working out the bugs with Android Studio while learning the Android SDK was a challenge in and of itself.
Arduino? No!
Our initial design had us passing a USB signal from our Android app to an Arduino, which would then send signals to the breadboard being used to control our LEDs. After several hours of work and troubleshooting on Rex's part we decided to scrap that design, as it turned out it would require a rooted phone. This was, of course, not ideal.
Aux Output for Passing Signals
Ultimately we decided on an alternative to the arduino: by passing tones from the phone to the breadboard through an aux connection, we used high-pass and low-pass filtering on the breadboard to effectively create up to 16 different combinations of signals for the LEDs.
Signal Processing and Firmware Bugs
The Myo was difficult to work with at times. The aspects of the Myo which are integral to its advertised functionality were well-tested and worked well enough. We used one of the built-in "Poses," the "Fist" Pose, to detect when the rider is choosing to brake. To detect the other motions important to us, however -- left and right turn signaling with the left arm -- we had to determine what data from the Myo was available for us to use. It turned out that the best way to distinguish between these movements was by periodically analyzing the yaw and pitch data sent to us from the Myo. This would have been easy enough to handle, except that our Myo had a firmware bug which caused the yaw to continually increase even when the wearable was perfectly still. We worked around this by calculating the change in yaw from one timeslice to the next, which gave us a useable (if unreliable) method of determining when a left turn was being signaled by the rider.
Log in or sign up for Devpost to join the conversation.