Researchers at MIT have successfully flown an autonomous, robotic plane in a confined space and around obstacles without the use of GPS. The flight by the MIT Robust Robotics Group paves the way for the development of aircraft that can fly without pilots, even in remote areas where GPS may not be available.
MITís Robust Robotics Group flew a fixed-wing vehicle around the parking garage under a university building, navigating pillars safely, with only the use of onboard sensors to direct the path of the plane, according to the university.
Researchers at MIT flew an autonomous robotic plane in an enclosed area around obstacles, demonstrating that it is possible to build a self-navigational, fixed-wing vehicle that can fly at high speeds safely without the use of GPS. (Source: MIT)
The flight was achieved through the use of custom algorithms the team developed to determine trajectory and state, or the location, physical orientation, velocity, and acceleration of the aircraft.
The plane, which MIT researchers designed and built from the ground up, also had an accurate digital map of its environment to help it safely avoid obstacles. However, it had to determine in real time its location on the map and where it had to go during the flight, researchers said, using the algorithms to interpret data from laser rangefinder and inertial sensors such as accelerometers and gyroscopes carried onboard.
MITís work is similar to research being done at the Defense Research Projects Agency to design a new miniature atomic sensor system for missiles that would eliminate GPS dependence for navigation. And new research in unmanned aerial vehicles (UAVs) is always a subject of interest for the military, which makes broad use of unmanned crafts. The Air Force and NASA, for example, have been testing a new lightweight unmanned plane to explore innovations in both supersonic and subsonic flight and wing stability.
MITís specific work was born out of plans to create control algorithms for autonomous helicopters, which researchers eventually abandoned because ďthe fixed-wing vehicle is a more complicated and interesting problem, but also that it has a much longer flight time,Ē said Nick Roy, an associate professor of aeronautics and astronautics and head of the Robust Robotics Group, in a press release. ďThe helicopter is working very hard just to keep itself in the air, and we wanted to be able to fly longer distances for longer periods of time.Ē
MIT graduate student Adam Bry worked with AeroAstro professor Mark Drela to design the body of the plane, which has shorter and broader wings than a normal aircraft. This modified design allows the plane to fly at low speeds and make tight turns, which was necessary to avoid obstacles, according to researchers. It also made the plane stable enough under a fair amount of weight so it could carry the electronic sensors running the flightís control algorithms.
The next step for MIT researchers will be to develop algorithms that will let the plane sense its environment and design a map of it on the fly, precluding the need for a preloaded outline of the flight environment, according to the university.
That's probably a big leap, Beth, just as it would be a big leap to transfer the technology to an environment that is not pre-programmed for the flying craft. Yet I could see this technology eventually getting incorporated into drones.
Wow, I have to say this is a pretty amazing video to watch, especially knowing more about how they did it. I would think this could be very useful in collision avoidance of all types. Maybe it could even apply to conventional automobiles and race cars (though I suspect the potential for a crash is part of the thrill of racing).
Robots that make maps--and that update them continuously for navigation purposes, which it's not clear that this one does--are a topic we've covered before: http://www.designnews.com/document.asp?doc_id=240288
That one is also MIT, and also from a group in its CSAIL lab.
Thanks for the link, Ann. I find it amusing that this robot uses technology from the Xbox 360. Computer games have led a number of technology developments. In the automation and control world, they're using game technology for training and simulation. The miltary is also using game technology for training.
It would be interesting to see how the algorithm would respond to a change in the environment - such as if a new structure were added (or somebody's head for that matter). Would it know enough to just avoid the obstacle or would it think it is in another location of the map?
Good question, Jack. The new object might need to be programmed in. On their next stage, this team is going to try to get the plane to map its own environment on the run. I would think that would require GPS. But maybe not.
This looks cool. I love the idea of GPS-free flighs with the saftey that GPS offers. I think it would be great as an aid for pilots too. This would be good for humanitarian drops in remote areas that are isolated or cut off in natural disasters.
I can't wait to see the evolution of this. There are lots of other moving objects in the sky. Anything that can avoid or deter birds would be fantastic for the aviation industry.
Not having to pre-program the environment into the computer seems to be a common theme in many of the comments and on first brush, it seems like not attaining it would be, in a way, a failure or limitation of the ultimate design of this system.
However, how many humans would be able to fly a plane through an unfamiliar, enclosed environment without stalling or colliding with objects? Switching to a helicopter type device would help us, because it could come to a complete stop relatively quickly without stalling. But if we were in a fixed-wing aircraft which had to maintain forward motion at a minimum airspeed, with no do-overs or a re-set button, I think few people would be able to accomplish it.
If you play video games, think of getting to a new level, or running a new race course, in a game where you HAVE to keep moving to keep from getting killed, passed or disabled. It generally takes many, many tries before a gamer can familiarize his or herself with the environment and subsequently negotiate it at full speed successfully.
While the goal may be the ideal, it is asking a lot of the onboard computer of a fixed winged aircraft to do something that our human brains and senses cannot do. Again, in a rotary type device, it is entirely feasible.
TJ--My thoughts exactly. I don't know what changes might need to be made to adapt the systems to this use but I do think that could be accomplished with some effort now that the work that has been done already. I do see the great advantage for in-flight systems where GPS is not available or has been disabled. This is great work by MIT and contributing agencies of our government.
Festo's BionicKangaroo combines pneumatic and electrical drive technology, plus very precise controls and condition monitoring. Like a real kangaroo, the BionicKangaroo robot harvests the kinetic energy of each takeoff and immediately uses it to power the next jump.
Design News and Digi-Key presents: Creating & Testing Your First RTOS Application Using MQX, a crash course that will look at defining a project, selecting a target processor, blocking code, defining tasks, completing code, and debugging.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.