Flight Preparation

 Required Persons During Test

 Other Attendees

Drivers & Seat Arrangement

 Success criteria

Testing Timeline

Aircraft Incident Procedures

Week-Before To-Do List

Day-Before To-Do List

Pre-Departure Checklist

Pre-flight Safety Checklist

Post-Flight Checklist

Flight Debrief

In total, multiple flights were performed. Basic functions such as flight modes, RTL, auto-land, and failsafes were tested.

Flight set #1 (basic function + flight modes)

The first set of batteries was put in for basic functionality check of the drone.

Open

Voltage of first set of batteries on the first set of initial flights

The drone was flown in three different flight modes, in the following order:

• Stabilize, Megan

• Stabilize, Anni

• AltHold, Megan

• Loiter, Megan

Flight set #2 (RC, GCS failsafe + RTL)

The first set of batteries was then used for testing RTL, RC Failsafe, and GCS Failsafe.

Open

Voltage of first set of batteries on second set of flights

The order of testing was as follows:

• Stabilized takeoff + flyout → RTL (from short range)

• Loiter takeoff + flyout → Transmitter poweroff (with RC link through RFD900’s) → Smart RTL

• Loiter takeoff + flyout → RFD900 poweroff (RC Link on drone directly). → Loiter landing.

  • “Continue if in pilot controlled modes on GCS failsafe” was checked.

• Loiter takeoff + flyout → RFD900 poweroff (RC Link on drone directly) → RTL

  • “Continue if in pilot controlled modes on GCS failsafe” now unchecked.

• Testing RTL mode actions on altitude

  • first test ensures loiter will climb to RTL altitude

  • second test ensures loiter will only descend at landing time.

Flight Set #3 (testing autoland, position hold)

A new set of batteries were used to test position hold, RTL, and autoland accuracy.

Open

Voltage of second set of batteries over flight set #3

• GPS-Only loiter accuracy was observed to be within +/-30cm when left at no stick input (very good).

• RTL accuracy evaluated to be roughly +/-50cm

Flight Test Post Mortem

Logs here: https://uofwaterloo-my.sharepoint.com/:f:/g/personal/uwarg_uwaterloo_ca/EiGR3sG4uM5Gr15i40EfLP8Bzvi_Wu2oMfD9rSTNAVL9-g?e=ehv0dz

Photos here: https://uofwaterloo-my.sharepoint.com/:f:/g/personal/uwarg_uwaterloo_ca/EtUtjlQe_kpPuI0S8IbM6YoB5-gikldpSNRogi2iGjDLoQ?e=bCaPyL

Flight Characteristics

Since this is the maiden flight of Vanny for a while, I’d like to go over some flight characteristics.

Pitch/Roll Control

Pitch/Roll control was good, however the drone was slightly sluggish to respond and the maximum roll/pitch angle could be higher (I think).

Open

AltitudeHold Roll vs Desired Roll

In the above image, you can see that there is some overshoot and oscillation on roll, which you can more clearly see on a zoomed in crop of the last few half-impulses before returning to center:

Open

Althold Roll vs DesRoll overshoot / oscillation

The same can be observed on the pitch axis:

Open

AltHold DesPitch vs Pitch graphs.

Note that during the entire impulse test, the drone was able to maintain altitude fairly well, although it’s difficult to tell since the drone is correcting for it’s own altitude during flight. This will be discussed more when I talk about altitude hold.

Yaw Control

Yaw control seems to be one of the weaker items on our drone . Not sure if this is due to poor tuning or throttle output saturation (unlikely). In general, the drone would carry a lot of momentum after letting go of the yaw input, and would tend to overshoot / continue with yawing after letting go of the stick (often leading to backstick).

Open

Pilot controlled yaw vs desired yaw

Open

(mostly) software only yaw control. Step inputs from pilot.

I’m not sure what might cause this - but I would like to compare results after running auto-tune on yaw.

Altitude Control

Altitude control was excellent, even without a rangefinder. The GPS + Barometer + IMU system on the higher end pixhawk is really doing wonders to keep the drone stable.

Open

Desired Alt vs Alt graph from flight #2

You can see pretty clearly that our altitude and desired altitude are quite close to each other, in all guided flight modes. There is some measured “oscillation”, but this is very small. Any actual difference in altitude during flight would have come from errors in measuring / calculating altitude from EK3.

Open

Pilot Controlled + Automatic Altitude Hold (desired vs measured)

Open

Throttle input vs throttle output

It is worth noting that the hover throttle was fairly consistent throughout all flights, with the individual motor outputs being around 1500 us. Angle boost was also fairly minimal, but present and effective at helping the drone maintain hover:

Open

Angle boost being applied during impulse flight. Notice hover throttle changing & learning.

Open

Individual motor outputs during flight

Red = Front Right

Green = Back Left

Blue = Front Left

Orange = Back Right

In general, it’s hard to determine if one side of the drone is significantly more imbalanced than the other, but you can clearly tell that the hover throttle for the drone is around half (i’m not sure if there’s better measurements that I can make).

Position Accuracy

Open

Target Position vs Position (N, E)

In general, position accuracy was great (considering this is one non-rtk gps). We were within +/-50cm doing multiple takeoffs and RTL’s, and the drone logs show that desired position and actual position correspond quite well. In general, this tells us that any deviation we see in position would come from measurement inaccuracies from the sensors themselves.

RCA for Red Items

Create a RCA and keep it as a child page.

Create a RCA

Action Items

Other Resources

Booking WRESTRC Field

WRESTRC