Computing Optic Flow with ArduEye Vision Sensor

摘要

Abstract : Flight control of micro aerial vehicles (MAVs) is challenging due to the size of small-scale robotics platforms, which places size, weight, and power limitations on onboard sensors, as well as restricts the data processing algorithms that can be implemented on the platforms. We focus on computing optic flow in order to control a MAV using small vision chips as opposed to a complex camera system in order to cut down on the size and weight of the sensors as well as the processing power needed. We propose using an externally produced vision sensor by Centeye, ArduEye, to develop an algorithm for computing optic flow, from existing algorithms tailored to the sensor, to be used to control a quadrotor. Such efforts will benefit the various MAV research thrusts by providing a successful optic flow sensor and data processing algorithm that can be applied to the flight control of other robotic platforms.