An Implementation of the System on Chip Control System for a FPGA-Based Computer Vision Accelerator

In this paper, we propose a System on Chip (SoC) system to integrate a computer vision acceleration IP using Field Programmable Gate Arrays (FPGAs) and the open source autopilot system called the ArduPilot. Since our goal is to build a low cost, high efficiency and compact system, we used MicroBlaze micro controller unit (MCU) which is provided by Xilinx Also, for the flexibility and convenience of machine control, we use the commercial Flight Control Computer (FCC) Pixhawk4 (PX4). We implemented the object tracking IP and the path planner IP for controlling the motion of the machine, and designed the MicroBlaze embedded software for Mavlink packet generation and SoC system control. The entire system was tested on the drones and confirmed that real-time flight control is possible based on the tracking results of moving objects such as vehicles and human.