Performance Evaluation of a Frugal Open-Source Humanoid Robotic Platform

This paper presents the design, development, and experimental validation of a frugal humanoid robotic platform named Autonomous Robot Development Open-source Platform (ARDOP). The platform is conceived as a modular, fully integrated, open-source, and cost-conscious humanoid system intended to support applied research and laboratory experimentation. In contrast to proprietary humanoid platforms, ARDOP emphasizes accessibility, reproducibility, and coherent subsystem integration within a unified Robot Operating System (ROS) framework.ARDOP is organized around three core functional subsystems: perception, manipulation, and navigation. The perception subsystem enables object detection and three-dimensional spatial localization using RGB--D sensing and convolutional neural network–based techniques. Manipulation is achieved through coordinated multi-degree-of-freedom robotic arms that support controlled pick-and-place operations within the workspace. Autonomous navigation is realized through obstacle-aware localization and map-based path planning for indoor environments.The platform is validated through simulation and hardware experiments that assess subsystem functionality and system-level coordination. Experimental results demonstrate reliable manipulation accuracy, stable navigation performance, and consistent operation across repeated trials. Rather than emphasizing algorithmic novelty, this work focuses on system-level integration, experimental validation, and practical accessibility, providing a reproducible humanoid robotic platform for applied research and educational use.