Development of a Scalable Autonomy Framework for Robotic Teleoperation in Remote Greenhouses

Abstract

Human-robot collaboration is becoming increasingly important for performing complex manipulation tasks in remote or hazardous environments, where full automation remains challenging. In this context, scalable autonomy offers a promising approach by combining human decision-making and adaptability with the precision and consistency of robotic systems. This work explores this concept through the development and validation of a scalable autonomy framework for robotic teleoperation, designed to improve dexterous manipulation in constrained and remote environments, with a particular focus on remote greenhouses. These concepts are realized on a multi-modal hand-arm user interface, which is paired with an avatar robotic arm and a dexterous multi-finger end effector. Several levels of autonomy were implemented, ranging from direct teleoperation to assisted grasping and trajectory planning, allowing seamless transitions between manual and autonomous control. This framework is particularly relevant for applications such as space exploration, where direct human intervention is constrained and robotic systems capable of adapting their autonomy levels can significantly enhance operational efficiency and safety. A motion planning framework based on a Position-Position control strategy was also integrated, replacing the original Position-Velocity formulation to enhance precision in object manipulation and improve synchronization between operator and robot. Experimental results demonstrated the feasibility and the potential benefits of the proposed approach, showing an apparent tendency for improvements in task performance, stability, and reduction of operator workload. These findings suggest the potential of scalable autonomy as an enabling paradigm for future space robotics applications, particularly in tasks that require human supervision of autonomous systems operating under uncertainty and environmental variability