The Development of a ROS Driver for the Crazyflie Micro-drone and Its Use to Study Drone Proximity Detection via Air Disturbance Analysis
Unmanned aerial vehicles (drones) have been an active research area recently, the use of drones is expanding to commercial, scientific, agriculture applications, such as surveillance, product deliveries and aerial photography etc. One challenge for applications of drones is detecting obstacles and avoiding collisions. A typical solution to this issue is the use of camera sensor, ultrasonic sensor or manual control (teleoperation). However, these solutions have costs in battery lifetime, payload, operator skill. And it’s too heavy to put extra stuffs on small drones. In this research, we find a way to detect obstacle by analyzing the air disturbance around the drone from the drone's built-in flight sensors like gyroscope and accelerometer. This approach allows the drone to be lightweight and low-cost but still avoid collision automatically. My research focuses on the first phase, detecting the wind directly produced by another drone. I developed a ROS driver of crazyflie (small drones) for air disturbance detection test. A series of experiments were performed with one drone flying close to but underneath a second drone. After the drones finish their flight trips, the sensor data are collected, then analyzed by data mining to get a data pattern indicating air disturbance. Additional experiments show that the pattern can be used by a drone to reliably detect the proximity of the second, upper drone. We show that the Gyroscope data is doing better than accelerometer data.
Zhao, Qian, "The Development of a ROS Driver for the Crazyflie Micro-drone and Its Use to Study Drone Proximity Detection via Air Disturbance Analysis" (2019). ETD Collection for Fordham University. AAI22585226.