Development process of a smart UAV for autonomous target detection

Abstract

The project outlines the development of an autonomous quadcopter that can be used in GPS or motion capture cameras denied environment where the quadcopter needs to localize itself. The quadcopter developed can detect targets using only onboard sensors and compute using low cost parts. Following a preset criteria, a medium sized S500 drone was modified to function using a Pixhawk 2.1 Cube flight controller for actuator control and an NVIDIA TK1 Developer board for the processing of raw sensory data, localization, and path planning.