Client Meeting October 19th - ProstheticFalconry/Capstone GitHub Wiki

Main meeting questions:

  1. What is the quality of camera needed for the Artificial intelligence identification of objects? Raspberry Pi camera resolution is the performance level needed for the quadcopter identification.

  2. How liberal is our allowed use of the Makerbot 2? Is plastic printed included in budget? Wear on machine is more important than the cost of plastic printed. Be considerate with the machine but don't be overly worried with the material. 5 cents per gram roughly.

  3. How many unique signal can be sent to/from the MIYO. By default MIYO has 5 signals running inside, theoretical signal maximum for signals is roughly 25.

Miyo training what is needed. Input to system is not same to as the MIYO, must be read and translated somehow. Is it realistic to have the MIYO control 4 attributes of flight (roll, pitch, yaw, altitude) considering. Can the quadcopter fly before this is fully designed by lab.

For shutdown we have 2 options, piggyback of MIYO or have an independent RF signal service. Independent requires a physical user switch and receiver with shutdown instructions, button to land, switch to emergency shutdown. Nic prefers a separate system to keep it independent from both failures of the of the quadcopter, and . Could have these buttons and switches on the receiver which could be mounted on the glove. This gives us options regarding the shutdown of the quadcopter that is up to the users discretion.

Beaglebone available to us? Yes but we'd have to clear code. We have a MIYO directly available to us. Did ESC's come with programming card? No. We will need shielding on the sensors to decrease noise from the rotating motors and propellers which will likely be put very close to these.

What is density of our plastics that will be 3D printed through the Makerbot 2. These densities can be tweaked at time of printing (via space made by hexagon backfilling). Any issues with 3D printing in terms of shapes? Yes but not too much info can be given, generally these problems occur only when it comes to small intricate shapes but we generally won't have these in our project. Printed objects are weakest along extrusion axis. Print perpendicular to the expected force Ex: Don't print arms width wise as motors will shear arms apart.

Beaglebone should know flight control position and motor powers to decide how much thrust is being put into lifting a payload and whether it should abandon this task. Force sensitive resistors are fine for determining actuator pressure. Padding and force sensitive resistors can still work in tandem as force carried through would not be affected by a controllability.