Hardware software co-design for quadcopter control

Abstract

Modern quadcopters development is becoming more computationally intensive and has brought more complexity than before, resulting in the need for faster and more reliable processing architectures. With emerging reconfigurable platforms tightly coupled with fast processor and high performance reconfigurable fabric, it promises the shift in focus from microcontroller to reconfigurable-based applications exploiting the limited resources available on an embedded device. These reconfigurable platforms offer significant advantages like speedup, better reliability and higher levels of determinism in timing critical applications. Furthermore, much emphasis is placed on computer vision in quadcopters pushing it to the limits to become completely autonomous. This approach is highly suited to implement on reconfigurable architectures and can also benefit in communication from the tight coupling between the hardware and software area. In this paper, we present the functional execution of a quadcopter running on a commercial available reconfigurable platform (the Xilinx Zynq), which serve as a basis for further research and development in this area.