Inventory managing system using quadruple legged robot and robot manipulator

Abstract

The robotics industry presents an attractive opportunity to improve people’s lives across many touchpoints. Despite monumental leaps in the robotics industry, the usage of robotics in everyday life is sparse. Recent studies reveal that the two most significant hurdles that robotic systems face for rapid adoption are: economic accessibility and generalizability. Existing robotic systems are too expensive and too specialized to gain traction. In this paper, we will address these concerns by developing a cost-efficient robotic system to perform the complex task of inventory management. With a minimally viable robotic system comprising a quadruple-legged robot and a robotic manipulator, we will demonstrate a wide variety of capabilities, such as localizing an object of interest with centimeter accuracy, projecting a virtual twin of an object of interest, and exhibiting complex robotic behavior such as organization of household objects across multiple platforms. These capabilities are developed through open-source frameworks such as ROS, an open-source robotics framework; Point Cloud Library, an open-source point cloud processing framework; and MoveIt, an open-source robotics motion planning framework, allowing for interoperability with existing robotic systems across a wide variety of purposes. With a no-frills robotic system that offers accurate robotic manipulation and the opportunity for contextual robotic behavior, this project will present an enticing alternative for everyday robotics.