Vision Based Control of Model Helicopters

The purpose of this paper is to explore control methodologies and vision algorithms to develop an autonomous unmanned aerial vehicle (UAV). UAVs include unmanned aircrafts, helicopters, blimps and other flying vehicles. An autonomous UAV brings enormous benefits and is suitable for applications like search and rescue, surveillance, remote inspection, military applications, therefore saving time, reducing costs and keeping human pilots away from dangerous flight conditions. UAVs are especially useful when (i) the working environment is inaccessible or hard to reach (planetary environments), (ii) flight is dangerous (due to war, contaminated environmental conditions), (iii) flight is monotonous, (vi) flight time is extended (atmospheric observations, data relay), (v) flight is not possible even by a skilled pilot (movie making, flight of experimental vehicles). Various unmanned vehicles are in service in military and civilian areas. Fixed-wing vehicles have long-range since they are energy efficient, but they lack the maneuverability required for many UAV tasks. Blimps are easy to control when there are fewer disturbances like wind, and lift comes from natural buoyancy, but they lack maneuverability as well. The rotary-wing aerial vehicles (also called rotorcraft) such as helicopters have distinct advantages over conventional fixed-wing aircraft and blimps on surveillance and inspection tasks, since they can take-off and land in limited space and can easily hover above any target. Moreover, helicopters have the advantage of superior maneuverability. Unfortunately, this advantage comes from the dynamically unstable nature of the vehicle and it makes helicopters very hard to control. Sophisticated sensors, fast on-board computation, and suitable control methods are required to make rotorcraft based UAV stable. Autonomy defined as “the quality or state of being self-governing”. Most of the commercial UAVs involve little or no autonomy. The goal is to increase the level of autonomy to fully autonomous operation including take-off, landing, hover (if platform capable of), way point navigation to more advanced autonomy modes such as searching, avoiding danger, combat, refueling, returning to base, etc. In addition, autonomy requires cooperation and communication with other vehicles. Currently a military UAV is supported with almost a dozen personnel to perform piloting, communications, and to control payload systems. The goal in the future is to reduce the personnel/UAV ratio (which is greater-equal to one currently) to lower than one, meaning that one personnel controlling various UAVs. Moreover, the vehicles will be autonomous to control their actions, interact with other vehicles to perform missions. This can be a commander controlling a fleet of military vehicles for combat, or a group of fire fighting UAVs commanded to extinguish a fire.