Real-Time Performance Evaluation of Household Object Detection, Recognition and Grasping Using Firebird V Robot

Abstract

Real-time object detection, recognition, and grasping are essential capabilities for autonomous robotic systems operating in dynamic environments. In this paper, we present a comprehensive performance evaluation of these functionalities using the Firebird V Robot, a versatile platform renowned for its agility and robustness. Leveraging state-of-the-art computer vision algorithms, we assess the system's ability to detect objects in real-time, recognize their identities, and grasp them reliably under varying environmental conditions. Our experimental framework encompasses a series of tests conducted in both simulated scenarios, enabling a thorough analysis of the Firebird V Robot's performance. Through meticulous experimentation and analysis, we provide valuable insights into the practical viability of deploying these functionalities in real-world applications. Our findings such as detection and recognition accuracy (92.2%), grasp success rate (0.8), total execution time (30sec to 60sec) and efficiency shed light on the strengths and limitations of the methods being evaluated and Firebird V Robot as a platform for real-time object manipulation. By evaluating the performance of these critical functionalities, we contribute to the advancement of intelligent robotic systems capable of seamlessly integrating into various domains, from industrial automation to service robotics and beyond.