While it’s clear from the Epilogue that more work is needed on the drift PIDs, the concept is proven, and the project moves on in order of feasability, importance and cost.
- I’m still obsessed about noise suppression for the accelerometer along with weight reduction, so I’ve reworked the breadboard to use a smaller one, and intend to seat it on a carbon fibre platform that has a 10mm silicone foam layer between it and the platform seated between her legs. That platform itself will also be replaced with a carbon fibre equivalent- I’m just waiting for the CF sheet to complete that.
- No drift in windy conditions uses the same PID logic as deliberate movement in non-windy conditions: a horizontal speed PID target of zero meters per second in windy conditions keeps the PID stable by dipping Phoebe’s nose down into the wind; a horizontal speed target of non-zero means Phoebe’s nose will drop in the desired direction of movement. That opens up the option of user control via WiFi radio / remote control to set a desired target for the PIDs; I’m already 50% there on the physical build / components – 2 joysticks control the desired speed along the X, Y and Z axes, allowing horizontal movement and ascent / descent.
- Turning left / right (as opposed to moving left or right) will need a yaw PID which takes a target from the remote control as the desired angle of turn. Then as she moves forwards with her nose down, Phoebe’s yaw will change the direction her power is applied and she’ll turn as expected. That outer PID to rotate by a fixed angle doesn’t exist yet, only the inner one to stabilize the yaw.
- Takeoff from a non-horizontal platform requires the ability to measure the takeoff angle using a compass; the MPU9150 is pin compatible with the MPU6050 I use for accel / gyro but also contains the compass – breakout boards are available but not in the style matching my MPU6050 breakout, so I’m investigating building my own. The compass will also be useful for the new yaw PID for turning corners by a fixed angle.
- Phoebe needs protection against power and RC connectivity failure to trigger a stable landing
- Adding a GPS would allow her to track her path, either just for display, or to allow her to return to base under problem conditions like above. She could also follow apre-programmed path autonomously.
So plenty to keep me busy despite the winter weather. But first I need fine tuning on the drift on longer flight times so I can confirm Phoebe’s suitability for a move into the garage.