Assemblers: A modular and reconfigurable manipulation system for autonomous in-space assembly

The Assemblers project will transform in-space assembly (iSA) efforts by applying the latest technological advances in autonomous manipulation, machine learning, control theory, and state estimation to expand concepts pioneered by NASA. Assemblers will develop and demonstrate a lightweight, highly stiff assembly manipulation system consisting of stacked Stewart platforms (SPs) and a multi-agent task management system. This project builds upon LaRC’s previous efforts-7 via a demonstration of a lightweight, highly stiff assembly manipulation system and multi-agent task management system needed to robustly manage in-space and surface construction tasks. The Assemblers system will be a modular and reconfigurable manipulation architecture featuring interchangeable end-effectors for the tasks required to construct complex space systems. In addition, each SP consists of two plates connected by six linear actuators enabling six degrees of freedom (6 DOF) of motion between these plates. The multi-agent task management system will manage task scheduling and the optimization/planning needed for assembling complex space structures. Assemblers will also implement methods for error detection and correction using machine learning and autonomous decision-making within the multi-agent planning system. An example manipulator configuration consisting of two SPs is shown in figure 1. Figure 2 depicts a conceptual rendering of two manipulators each composed of four SPs collaboratively assembling a truss structure. This project will culminate in a demonstration of the complete assembly of a lunar surface solar array structure, enhancing the Artemis return to the moon mission. Three key features are highlighted in this project: (1) tolerance to component failures, (2) error detection followed by correction, and (3) diagnosis and repair in the event of failures during the operational lifecycle. Project deliverables include experimentally determined key performance parameters (KPPs), multi-agent planning and control software, an autonomous iSA design, and lessons learned for future iSA projects.