ErichF
Elite Veteran
Location: Odessa, FL 33556 (Tampa Area)
| The software is a bit of a chore at first. You really need to study the manual(s) to figure it out. The hardest part was setting up the throttle and pitch curves. It's bad enough they use 8 points (no real center or 1/4 points), but you have to set the pitch curve in the AP. The AP doesn't use a pitch input from the RC receiver, only throttle. So, you have to setup your radio's throttle to be linear, 100%ATV, and 0-100%. Then, you fill in a look-up table for the autopilot to create the pitch curve values, based on the throttle input. After that, you create another look-up table for the throttle curve. The values used are fine servo values, which are numbers from -32767 to +32767. So you have to do some math to figure out what numbers to use for given throttle and pitch positions. Like I said, it's a chore.
Also, flightplans start out as "fly" files, that dictate certain actions the AP will do and when, as well as waypoint orders. They also tell the AP what to do given certain failures, such as loss of Modem link, RC failure, etc. These files have to be typed up in command line format, which for a operator/tech like me, requires a learning curve. Here's an example of such a file:
imperial
takeoff
climb 100
waitClimb 100
wait 100
flyTo (0,300)
wait 100
flyTo (-300,300)
wait 100
circuit (0, 0)
wait 100
repeat -1
definepattern=gcsFailed
wait 20
climb 300
flyto [home]
wait 30
climb 100
circuit [home]
You can create a flightplan of waypoints on the map through the GUI, and apply the plan to a fly file, which inserts the waypoint lines into the file for you. That's one good thing. Also, you can drag and edit waypoints while flying in real time without redefining the fly file and uploading it. It still requires more typing than I would prefer.
So, it's not ideal, but it does make it quite versatile.
Perry, we didn't use gel mounts on anything yet. What you see are nylon standoffs. The IMU as you called it is actually the AGL transducer (sonic altimeter board). The IMU and gyros are all mounted to the main 2128 board. We have yet to do vibration tests, and will accomplish that prior to flying with the AP in control.
The magnetometer is what's wrapped in blue tape mounted to the tailboom. It calibrated easily and is very accurate. It's a three axis magnetometer. The calibration process is two rotations upright, then two rotations with the heli on it's side. This type of magnetometer isn't prone as much to typical magnetic turning errors common with single axis units.
The initial flight test cards from the manual have you start in RCPIC mode and CIC Attitude Arcade mode. Once the low level gains are set, then you use CIC attitude arcade as a backup, and adjust position CIC arcade mode, then move through the various modes, ultimately flying in Attitude Arcade to CIC fully auto. It's not until that point do you attempt an autonomous takeoff, flight, and landing.
My gripe with these arcade modes, as they call them, is that you have to keep the machine within range of the RC transmitter for any kind of semi-autonomous manual flying. Another system we are testing allows you to turn off the RC transmitter after takeoff, send the aircraft downrange under auto nav mode, and switch to manual directed flight via PC gamepad at any time and range. You have full control with the gamepad, left right cyclic, heading with rudder, and altitude with collective. Let go of the gamepad sticks, and it holds altitude, position, and heading until you reactivate the flightplan. If the Micropilot could do that, it'd be a really nice system I think. While Micropilot has an RPV Mode, it's not tested in helicopters yet, and isn't really full control. The gamepad can only adjust heading, airspeed, and altitude. Not enough for a heli, as it totally ignores cyclic control needs.
Whew, that was a long post  Hope I answered your questions in all that.
We fly this on Wednesday and Thursday, wish us luck 
Erich
Team Kyosho Field Representative |
