Hello everyone,

I’m currently working on developing a custom flight controller for small UAV's, and I’d like some advice on selecting the right SoC FPGA for the project. Here’s a quick overview of the requirements:

• Board Size: ~40x40 mm
• Processor: Dual ARM Cortex (preferably A9, A53, or higher)
• OS: Linux-based, capable of running Ardupilot, PX4, and Mission Planner
• Tasks: Real-time control, image processing, machine learning, and VSLAM
• Power Efficiency: Optimized for UAV applications
• Budget: The FPGA chip should cost between $150-$200 maximum

I'm looking for a price-optimized FPGA that can handle vision and ML tasks while ensuring stable real-time control. Key priorities include sufficient logic cells, high-speed I/O interfaces, and compatibility with open-source autopilot systems.

Does anyone have suggestions on specific FPGA chips from Xilinx (AMD) or Intel (Altera) that fit these criteria? I’d appreciate any recommendations and insights from those with experience in similar projects.

Thanks in advance for your help!

I would rather look at the Raspberry Pi Zero W board.

Its speed, memory, and dimensions are enough to solve your problem.

And it has a richer infrastructure to keep solving your problem. 

The solution in FPGA will be 3-5 times more costly and time-consuming.

Thanks for the feedback. This board will have to perform VSLAM, machine learning, image and video processing. I am not sure if the Raspberry Pi is powerful enough to handle all the tasks. 

When selecting an SoC FPGA for the development of a small UAV flight controller, several factors need to be considered, such as processor performance, board size, power consumption, and cost.

Xilinx (AMD)

• Zynq-7000 Series4
  • Processor: It integrates an ARM Cortex-A9 dual-core processor, which meets your requirement for a dual ARM Cortex processor.
  • Board Size: The Zynq-7000 series comes in a variety of package sizes, and you can choose a suitable one to fit the ~40x40 mm board size.
  • OS and Compatibility: It supports Linux-based operating systems and has good compatibility with open-source autopilot systems like Ardupilot and PX4.
  • Tasks: With the integration of an ARM processor and FPGA fabric, it can handle real-time control tasks. The FPGA part can be used for image processing, machine learning, and VSLAM tasks.
  • Power Efficiency: It is designed to be power-efficient, making it suitable for UAV applications where power consumption is a concern.
  • Cost: The price of the Zynq-7000 series is within your budget range of $150-$200.
  Is it help you?