An autonomous flight control system for a model helicopter

Abstract

Flight has captured the imagination of the human race for ages: and none of these craft surpass the helicopter in agility and manoeuvrability. Despite this, they remain some of the most difficult contraptions to control, with autopilot technology being in its early stages of development especially for civilian operations, such as search and rescue. The unique capabilities of helicopters make them also the ideal choice as Unmanned Aerial Vehicles for such applications as real-time reporting/filming. This requires the use of a simple, portable, cheap but effective platform which can be controlled by an amateur without having to learn to fly one. This project focuses on developing a Fuzzy Control System for a 400-size Radio Controlled Helicopter. The project is developed on a simulator of the Esky Honey Bee King II, which should be adequate for such applications. Fuzzy Control is used to capture the subtleties of helicopter dynamics without going into the complex mathematical technicalities which would not only complicate matters considerably, but would require more powerful hardware in a physical implementation, as well as being more costly to fine-tune for different situations. The controller is developed in a modular fashion according to the controllable axes of the helicopter. The simulator used is Clearview RC Flight Simulator, with the autopilot application written in C++. Various simulations were run in order to verify correct operation of the algorithm, under both ideal and windy conditions, with more than satisfactory results. Some work was also carried out towards a physical implementation, namely interfacing with a set of sensors for orientation estimation.