Hello Hackers and Mobile Scrollers ^_^
[WIP] Quadcopter Design
A friend and I are building an educational platform for teaching controls and design. We have begun this project with a prototype quadcopter.
The Road to Victory: Making Embedded Processors More Accessible.
The NXP Cup was a world-wide autonomous racing competition in which, mainly undergraduate, student teams competed to program an NXP microcontroller powered model vehicle, to complete laps on unknown tracks faster than anyone else.
Most of these students were not embedded developers, and many teams failed to complete a single lap. Even in the final round, typically half of teams failed to complete the track.
I developed a Simulink Support Package that allowed teams to focus on the high level control logic, and let Simulink take care of the embedded C code and hardware drivers.
After this Support package was created, the 10th aniversary competition saw the first ever final round in which 100% of the teams completed the track.
Creating the Matrix: Simulation and Success.
Many of the teams could not afford a full scale physical track to test their vehicles on, so I built a physical simulation of the model vehicle and an App that allowed teams to design their own custom tracks to test their algorithms on.
This MATLAB App allows them to better understand how various gains in their controllers affected the behavior of the car; when the physical car did something odd they could model the behavior and figure out a fix more rapidly in simulation than they could modifying embedded code by hand.
Can you see it now? Open Source Visualizations
MATLAB is a fantastic language for creating plots of data, but many MATLAB users want to visualize a physical object, not a plot of the objects position over time.
The open source MATLAB Toolbox, PHYSVIS allows users to write the equations that define a physical system (like a simple pendulum) and automagically get an animation of that physical system.