Target Tracking by a Quadrotor Using Proximity Sensor Fusion Based on a Sigmoid Function ?

摘要

The use of micro-class Unmanned Aerial Vehicles (UAV) for consumer applications is rapidly growing. Many such applications employ intelligent systems in order to interact with the environment around the UAV. This paper demonstrates the modelling, simulation and experimental verification of a one-dimensional object tracking quadrotor that can detect and follow a solid object in front of it by regulating its distance from the object. A combination of a noise-based filter along with a sensor fusion technique using a sigmoid function was developed for a specific combination of two proximity sensors. This system uses a Proportional-Integral-Derivative (PID) controller to generate a single high-level pitch reference based on the sensor fusion output, in order to track a target. Low-level attitude control and altitude maintenance is simultaneously performed by a commercially available autopilot system.