Accounting for wind
Adjusting an aircraft's flight path to ensure on-target drops
Rain and Sikorsky, a Lockheed Martin company, have jointly demonstrated autonomous fire suppression operations with a Bambi bucket suspended beneath Sikorsky’s optionally piloted Black Hawk aircraft. In May, the platform was used to validate a wind deviation model that was integrated into Rain’s suppression engine, which is used to drop water on a target after it receives approval from the incident commander.
On the flight field behind Sikorsky headquarters in Connecticut, an optionally-piloted Black Hawk helicopter equipped with Rain’s software targets a test fire in a moderate breeze while our team watches from as far away as California. We were there with Sikorsky, a Lockheed Martin company, to validate our latest wind deviation model on their optionally-piloted UH-60 Black Hawk helicopter.
Rain’s suppression engine, including the wind deviation model, was designed by Mohaimen Islam, who monitored the flight from the Sikorsky mission control center.
Rain adapts autonomous aircraft with the intelligence to perceive, understand, and suppress wildfires. Our wind deviation model uses inverse kinematics to calculate the flight path the aircraft should take to adjust for the effects of wind that the falling water will experience. Adjusting for the wind at the fire location is a critical feature of our suppression engine, ensuring that firefighters have the tools to use every droplet of water for maximum benefit.
Predicting water's behavior as it leaves a swinging bucket attached to a helicopter (that is also generating rotorwash) is a very complex problem. To predict how the water would move as it is released from the aircraft, we had already created a suppressant delivery model that accounts for factors like the mass, shape, and size of the water droplets through the entire water drop. This model must account for several evolutions of water shapes and sizes, since water behaves quite differently when immediately released from the aircraft compared to a few feet above the ground. This already-difficult problem becomes more tricky when you consider the wind at the fire location—which is often gusty and tumultuous.
Rain validated a wind deviation model with Sikorsky’s optionally-piloted Black Hawk helicopter. Wind deviation is one part of our broader wildfire mission autonomy system, built to equip fire agencies with tools to stop wildfires before they grow out of control.
Our wind deviation model builds on the suppressant delivery model, allowing us to apply real-time wind conditions to forecast the trajectory of water droplets as they are released from the aircraft—the more precise our model, the more water lands on the targeted fire.
High winds are one of the most difficult challenges encountered in aerial firefighting, so modeling the impact of wind into our existing suppressant delivery model was an important technical milestone we were eager to validate.
Having flown the wind deviation model extensively in simulation, it was now time to validate it in the real world. The chart below shows the data we gathered from a water drop with 11 knot winds. The data from the test vividly illustrated how much the suppression engine must adjust the aircraft’s flight path for an on-target water drop.
The chart shows the data we gathered from a water drop with 11 knot winds. The red dot is the fire, the blue line in the top left shows the direction of the wind, and the black triangles show both the flight path of the aircraft and the trigger and empty bucket points, respectively. The blue rectangle is the ideal drop location, and the yellow rectangle shows the actual water drop location. When comparing the flight path of the aircraft (black triangles) with the actual water drop location (yellow rectangle), this chart vividly illustrates how much the suppression engine must adjust the aircraft’s flight path for an on-target water drop.
Ahead of the flights, we defined success criteria to be a drop where the drop pattern on the ground completely envelopes the entire area of the fire, and are happy to report our wind deviation model met that criteria for each of our drops. Across water drops, we saw lateral precision between 3cm and 1.4 meters and longitudinal precision of 5.3m, all well within bounds. Because the width of a water drop is narrower than its length, we were particularly happy to validate extremely precise lateral accuracy.
Rain’s CTO, Bryan Hatton, uses a GPS device on the ground to enable accurate measurements of the test results down to the 2cm level.
In the video below, the team prepares to depart for Sikorsky headquarters in Connecticut, where the wind deviation model was flown on a UH-60 Black Hawk helicopter in collaboration with Sikorsky. Sikorsky’s MATRIX™ flight autonomy system enables rotary and fixed-wing aircraft to be flown with or without humans on board, depending on the mission, and has been a fantastic platform for integrating with Rain’s wildfire mission autonomy software.
Hitting the target in all the wind conditions we experienced validated our latest wind deviation model and the theoretical basis for adjusting for winds beyond those parameters. Wind deviation is one part of our broader wildfire mission autonomy system to equip fire agencies with tools to stop wildfires before they grow out of control.
