Using Shapes on Mobiles to confuse optic flow sensors in Micro Air Vehicles appears to be feasible. Today we are studying the flight paths of Bats, Insects even birds to build and fly miniature UAVs; called Micro Air Vehicles or MAVs. There is significant research being done on this. The goal is to have mini MAVs the size of dragonflies, about 15 CM, which are autonomous and can fly through tunnels, under tree lines, under small bridges and navigate around obstacles. By using the optical flow sensors a new technology of mind boggling proportions the tiny MAV would take send out pulses and capture images, as the next pulse sees the image again it is bigger, this way the tiny MAV can sense when it is really big and getting close so it can then fly around it. There are two ideas, one is to use a single sensor unit and the other is to use multiple sensors. If you have ever studied insects, how they fly and navigate, you can see the enormity of the mathematics and algorithms it takes to make all this happen. Here is how it all works you will have to read all of these papers to understand this concept and to continue such discussion; http://www.centeye.com/pages/resources/downloads.html By the way compliments to Geoffrey Barrows for his studies and insight and recent award. Thus creating the fuzzy logic and sets of rules to determine what is an obstacle and what is another moving object. If you read one of Tom Clancys books where these little dragon flies with tiny explosives sit in the grass of the enemies runway and wait for an aircraft to take off and fly in a swarm and in front of it and allow themselves to get sucked into the jet engines where on impact with the fan blades explode, taking out the engines thus the aircraft is taken out with a swarm of micro-air-vehicles which resemble a bunch of less-than-palm-size insects. Insects have interesting characteristics. Such small units may be nearly impossible to shoot down or stop once they are set o their mission. Using sound maybe simpler than using optic network flow sensors, but the optical flow sensors can handle more things. For instance a potential mid air collision with a car, person, truck which is also moving. Also a devise, UAV, MAV created to take out another MAV, similar to a bat, which can eat 1000 bugs an hour. By lining them up and devouring them as it flies. An anti MAV, MAV, you see? The impending MAV wars will be interesting. For us to use the advantages given to us by evolution and to study the systems, hunting techniques, flows and cycles of organic devices. There are many things going on in species at all times and many different simple systems running simultaneously that are interacting. If you want to get technical you could say the wind, weather, food supply, symbiotic relationships all going on at once in the world of an insect or in the world of a human for that matter. Or a computer working on a decision matrix program whose job it is to take into account all these things and use rules to decide what to do. We must take a seat and actually think and study these things in depth. A good book to get you in the proper mind set to study these systems is Stephen Wolframs A New Kind Of Science. Start Small and Finish Big approach to the world, where simplicity and complexity are the same. Although it might not be precisely correct in all accounts and often takes liberties of explaining things from a single system program point of view it will assist you in studying the Flows of everything we see and observe. Optical Flow systems maybe somewhat more complicated and difficult to use than, sonar or sound, they may also add additional weight and therefore increase size and you could lose your advantage in that case depending on your mission. The reason I bring up these matters of how to fool a MAV or defeat one, or kill it, so to speak, since it will be behaving more similarly to an organic machine than a simple devise. Well because the academia crowd is sharing information with other countries. These MAVs after all are a combination of many sciences, many of which are life based, which are not very secret in that anyone can sit out side and watch a Bee fly in a zigzag pattern like you have to do when you taxi a tail dragger so you can see where you are going and what is coming up. Now then is an optical sensor really the best bet? Some of the best and brightest agree that it is. However Bats do not crash into things and they use sonar, or do they really use both optic flow sensors (its brain and eyes) with the benefit of sonar too? Well at least we know they always turn left when they exit a cave? Interesting. It appears that the most advanced studies are based on the Bee for the MAV, which makes a lot of sense. Now then we were talking about originally fooling the optic flow sensors with a spinning mobile like you put above the bed or crib for a small child. When you take different shapes and spin them, they will screw up the optical flow sensors, because on minute it will appear they are far away and then the silhouette turns to full size and shape which in the mind of the optic flow sensor appears to be an object coming up fast. Then it becomes small like a B-2 Stealth coming at you, that you cannot see, then it becomes big again when it turns, huge in the mind of the optic flow sensor, causing a reaction of quick change in trajectory. Then the MAV changes direction and another piece on the mobile turns side ways and again cause the unit to change directions. So then think of the problem when you put a few of these things in the path of a MAV or near a window to protect the Head of State of a neighboring ally. Your opponent cannot attack the human being inside because the MAV will turn back, meanwhile assume that other MAVs or objects are moving fast towards it every time it gets close. Thus you have fooled the MAV. If this works it maybe a good way to keep mosquitoes away from your window without a screen. Although their processing of data even with a tiny brain is much faster than that of the MAV with the multi-optic flow sensor network? Calculations per second are the issue and the speed of flight. Confusing the optic flow sensors on a MAV might be the way to go. Also the difference between sonar or sound and optic flow networked sensors on the same MAV may take into consideration these issues and the difference of the combining of data may in fact alleviate the problem. Also leaves falling from trees while flying in the canopy of trees along a road might be difficult since there are many of them and they flip around while they fall, especially in FALL, which is upon us. So this is not time to take a trip with any MAV in autumn. Certain things fool insects of certain types, while others have no problem working around those challenges. So to will different MAVs depending on how the networked sensors work, what they mimic and what their strategy is. It would be safe to say that several different ideas might be smart for MAV swarms since there is safety in numbers and different missions for each, perhaps a Mobile to confuse could be attacked by the half backs of the swarm, while the full backs and strikers stay in a holding pattern. If the decoy or optic flow network devise to confuse such as a mobile is rendered impotent then the entire swarm advances. This would be similar to a strategy of a hive, or Ants or other insects, which is somewhat based on net-centric warfare, which MAVs might play a crucial role in the next two decades? . Another issue is if swarms of organic species are so good at what they do, why try to re-invent them, just hire them as your mercenary army? Isnt that what Machiavelli said although they may not fight as hard as a citizen, there is safety in numbers as the missions will be relatively simple. As a matter of fact how do we know that the enemy may unleash a situation on us, using and organic delivery system of vector, will our MAVs be able to stay on patrol or are we better off hiring the Bats to do our dirty work and get a free dinner? And if the enemy does use organic material should we use technology of a different sort? This would keep them away and could confuse the enemies MAVs if they use sound to navigate. All options are available and all such options should be pursued since the enemy is also pursuing options as well as current allies who may someday be foes. In any case there is much work to be done to make smaller MAVs that work and can be powered for long periods for missions, which cover many miles. Designing ways to fool optic flow network sensors will be easier than designing ways to make them work, so we ought to be aware of this as we continue to figure out how we can use them to our tactical advantage. |
