UAV SWARM Health Management Project Website
Project PI: |
Prof. Jonathan How |
|
Recent News Headlines
September 26th, 2006 : MIT's Intelligent Aircraft Fly, Cooperate Autonomously
This article is written about the indoor, multi-vehicle testbed work by Ph.D Candidate Mario Valenti (EECS), S.M./Ph.D Candidate Brett Bethke (Aero-Astro), M.Eng. Candidate Daniel Dale (EECS) and others have constructed under the supervision of Prof. Jonathan How (Aero-Astro) at MIT and in collaboration Dr. John Vian and his research team at Boeing Phantom Works in Seattle, WA. This article was written by Lauren Clark, a freelance writter for the MIT News Office.
Notes for those interested: The air vehicles are from Draganflyer Innovations, Inc. in Saskatoon, SK, Canada. In addition, the indoor positioning system is made by Vicon Peak. Check out the technical paper published on the platform using this link.
Recent Project News
May 2007: Persistent Autonomous Search and Track Using Vision
A cooperative vision-based tracking algorithm has been implemented that allows a group of two or more UAVs to cooperatively detect and estimate the position and velocity of an object of interest, which may be a ground vehicle, airborne vehicle, or other object. A number of flight experiments have been carried out in which a group of UAVs has been assigned to search for and track a number of ground vehicles in a designated search area. These experiments also incorporated a mission management component that allowed the vehicles to switch out with each other as they ran low on fuel, allowing the tracking mission to be accomplished over long periods of time. For more details, see this paper
April 2007: 6 Hour Persistent Mission With Automated Recharging
As part of ongoing research into technologies to enable UAVs to automatically refuel themselves, an automated recharging platform was developed and tested in a five-vehicle flight test. In this test, the system attempted to keep a single vehicle airborne for 6 hours. Since the flight time of each vehicle is only about 12 minutes, it was necessary for the system to automatically recharge the vehicles as they came back to base so that the would be ready to fly again in the future. For more details, see this paper
July 25th to August 8th, 2006: First MILP RHTA / Obstacle Avoidance Flight Tests
As part of our architecture upgrade, we added an interface to allow an external user to fly and test their algorithms on the platform. As a result, after writing up a tasking algorithm in Matlab, a user can connect to the platform and test their algorithm using real hardware. As part of this testing, a Mixed Integer Linear Programming (MILP) based navigation algorithm running on an external computer was connected to the platform and used to guide the air vehicle around an obstacle placed in the room. As part of this testing we made two videos. In the first video, a large penalty in the cost function is placed on trajectories which cause the vehicle to fly around the obstacle - hence, forcing the algorithm to calculate a vehicle trajectory over the obstacle. In the second video, a large penality in the cost function is placed on trajectories which cause the vehicle to fly above the obstacle - hence, forcing the algorithm to calculate a vehicle trajectory around the obstacle.
July 31st to August 2nd, 2006: First Recharge System Test Flights
After an architecture upgrade, these flight tests were used to land an air vehicle and recharge it automatically using the platform's automatic battery recharging system. During the testing, we were able to demonstrate the system's capability to land on the recharge platform's landing pad and to start recharging the vehicle's battery after landing.
July 25th to July 31st, 2006: Airplane Tracking Test Flights
As part of our architecture upgrade tests, we had a pilot fly an airplane in our testbed's flight area to see if we could track and display the vehicle's position and orientation using the testbed's 3D visual interface display. During this testing we demonstrate the system's capability to acquire the position of an airplane flown by a pilot and display the data using the platform's 3D interface.
May 3rd and 4th, 2006: Boeing Tech Expo Flight Demonstrations - Hanscom AFB, MA
At the beginning of May, we took our flight test platform on the road to the Boeing Technology Expo at Hanscom AFB near Lexington, MA. Over the two days during the event, we flew over 60 flight demonstrations (more than 30 per day) over a 16 hour period. Each of the tests involved two vehicles -- one test involved two air vehicles flying a 3D coordinated pattern, the other involved an air vehicle following a ground vehicle. These demonstrations showed that the platform is capable of flying multiple missions repeatedly, on-demand with miminal setup.
For more information on the project, use the toolbar on the left of the page to check out more links!
Boeing Program Manager |
Dr. John Vian |
|
Contact Info |
 |
|