Up to now, we have had a very good run of data collection and training. Last sortie, there was a bit of a scare when the Hawk landed in a clearing close to some large trees. This week our good run ended.
We launched the Hawk, carrying our newly modified vegetation stress camera, into a 15kph wind. During the flight, the wind got up a bit and the telemetry indicated a wind speed of about 30kph at 400ft AGL. However, this was quite smooth air with little turbulence.
We conducted a number of transects of our study area, hoping to collect data on the vegetation that we could compare with our spectroradiometer ground truthing. We sent the students off into the bush to collect reflectance readings, while the UAV Team flew the sortie.
When we calculated that we had almost filled the camera data card with images, we recalled the Hawk. She dutifully returned to overfly the GCS. This was the point where things began to go wrong.
The Operator entered another two waypoints and turned the camera on again to collect a short transect over the GCS. We had a blue tarpaulin laid out on the ground and we hoped to get an indication of resolution from 400ft AGL. When this transect was finished, the Operator entered L4, the command for an automated parachute landing. However, none of us realised that the Hawk would now attempt an automated approach to the last waypoint, rather than the take-off point. (We now understand that we should have entered L4 while in Return mode).
The aircraft performed a series of circles with decreasing height about 200 metres to the East of our GCS position. Expecting the parachute to deploy upwind of the GCS we were startled to see it pop out 200m to the East. In the strong wind, the Hawk drifted gently but inexorably towards the trees. The more optimistic of us thought it would land just to the West of the trees, but the parachute caught right at the end of the tallest branch about 60ft AGL.
Since we were flying from a model aircraft field, Dr. John R. knew that this was not an uncommon occurrence and telephoned a tree removal company. They arrived within about 90 minutes and recovered the Hawk. Unfortunately, the tree climber either didn't tie a good knot to the Hawk, or completely forgot, because as the branches were cut away, the aircraft crashed to the ground from about 40ft, landing on the starboard wing and breaking the tip off.
So... we are grounded until repairs can be carried out. Our trusty suppliers, Skycam UAV, analysed the flight log today and determined what we had done wrongly. They believe that the wing can be repaired and today I packed the whole aircraft into its transport Pelican case and sent it to Skycam UAV by courier.
The one saving grace is that we collected approximately 270 images with the vegetation stress camera.
Several important lessons learned - proper operation of the landing procedures, emergency recovery to normal flight, alternative landing methods, and perhaps the most important of all: full communication between the Operator, Mission Commander and observers.
We launched the Hawk, carrying our newly modified vegetation stress camera, into a 15kph wind. During the flight, the wind got up a bit and the telemetry indicated a wind speed of about 30kph at 400ft AGL. However, this was quite smooth air with little turbulence.
We conducted a number of transects of our study area, hoping to collect data on the vegetation that we could compare with our spectroradiometer ground truthing. We sent the students off into the bush to collect reflectance readings, while the UAV Team flew the sortie.
When we calculated that we had almost filled the camera data card with images, we recalled the Hawk. She dutifully returned to overfly the GCS. This was the point where things began to go wrong.
The Operator entered another two waypoints and turned the camera on again to collect a short transect over the GCS. We had a blue tarpaulin laid out on the ground and we hoped to get an indication of resolution from 400ft AGL. When this transect was finished, the Operator entered L4, the command for an automated parachute landing. However, none of us realised that the Hawk would now attempt an automated approach to the last waypoint, rather than the take-off point. (We now understand that we should have entered L4 while in Return mode).
The aircraft performed a series of circles with decreasing height about 200 metres to the East of our GCS position. Expecting the parachute to deploy upwind of the GCS we were startled to see it pop out 200m to the East. In the strong wind, the Hawk drifted gently but inexorably towards the trees. The more optimistic of us thought it would land just to the West of the trees, but the parachute caught right at the end of the tallest branch about 60ft AGL.
There's always someone with a camera to record your misfortunes!
The AUT Hawk, apparently nesting in a pine tree
So... we are grounded until repairs can be carried out. Our trusty suppliers, Skycam UAV, analysed the flight log today and determined what we had done wrongly. They believe that the wing can be repaired and today I packed the whole aircraft into its transport Pelican case and sent it to Skycam UAV by courier.
The one saving grace is that we collected approximately 270 images with the vegetation stress camera.
Several important lessons learned - proper operation of the landing procedures, emergency recovery to normal flight, alternative landing methods, and perhaps the most important of all: full communication between the Operator, Mission Commander and observers.
Labels: AUT Hawk, CofG, spectroradiometer, UAV, vegetation stress camera
Post a Comment
Comments on the technology and application of UAVs are welcome. Comments are moderated and those containing links to commercial sites will not be published.