An IoT Reconfigurable SoC Platform for Computer Vision Applications

Abstract

The field of Internet of Things (IoT) and smart sensors has expanded rapidly in various fields of research and industrial applications. The area of IoT robotics has become a critical component in the evolution of Industry 4.0 standard. In this paper, we developed an IoT based reconfigurable System on Chip (SoC) robot that is fast and efficient for computer vision applications. It can be deployed in other IoT robotics applications and achieve its intended function. A Terasic Hexapod Spider Robot (TSR) was used with its DE0-Nano SoC board to implement our IoT robotics system. The TSR was designed to provide a competent computer vision application to recognize different shapes using a machine learning classifier. The data processing for image detection was divided into two parts, the first part involves hardware implementation on the SoC board and to provide real-time interaction of the robot with the surrounding environment. The second part of implementation is based on the cloud processing technique, where further data analysis was performed. The image detection algorithm for the computer vision component was tested and successfully implemented to recognize shapes. The TSR moves or reacts based on the detected image. The Field Programmable Gate Array (FPGA) part is programmed to handle the movement of the robot and the Hard Processor System (HPS) handles the shape recognition, Wi-Fi connectivity, and Bluetooth communication. This design is implemented, tested and can be used in real-time applications in harsh environments where movements of other robots are restricted.