Versatile educational and research robotic platform based on reconfigurable hardware

摘要

This paper presents the development of a field-programmable gate array (FPGA)-based platform. This platform is intended for the community as an educational and efficient prototyping tool, in digital signal processing, robotics, and control, but it can also be used as to develop custom robotic architectures. The proposed platform centered around an FPGA device is embedded on an off-the-shelf Phoenix hexapod robot. The platform is equipped with a camera module targeted to real time computer vision. From an academic point of view, the main use for the developed platform is to be employed as a didactic resource that facilitates the understanding of theoretical concepts commonly used in engineering courses and to speed up the development cycle in hands-on practice. A set of software tools and low-level drivers have been developed to configure the vision sensor, transfer images to/from a computer and to manage the testing of image and digital processing and control tasks in a transparent way. To validate the platform a hardware vision architecture and a locomotion control module have been created. The high level control is performed in a soft-processor that simulates the Arduino ONE micro controller. The soft-processor takes information from the vision sensor to generate controls signals for a generic module able to take a digital input and deliver a pulse-width modulation (PWM) output.