Experimental investigation on automated assembly of space structure from cooperative modular components

Highlights

• A series of simplified cooperative modular components are developed.

• An automated assembly strategy to fulfill the assembly task is proposed.

• An experimental demonstration of the automated assembly strategy is achieved.

Abstract

This paper presents a proof-of-concept investigation for demonstrating the feasibility of robotically assembling a space structure from cooperative modular components. The work is motivated by the requirement of in-orbit assembly of large space structures which play an increasingly important role in space engineering and science. The main goal of this paper is three-fold. The first is to design and develop a series of cooperative modular components. The second is to develop an automated assembly strategy to fulfill the assembly task based on visual measurement using an economic camera. The third is to experimentally demonstrate the automated assembly of the modular components into a much larger structure using the proposed assembly strategy and the developed hardware platform. In the demonstration, the robot grasps modular components, manipulates them to target positions with Iterative Learning Control and then assembles them to a central modular component fixed on a rotating platform to form a full structure. The whole assembly process takes about 4 mins and the video showcasing the experiment can be seen at https://youtu.be/OI7aaPwY_qQ.

Introduction

With the development of space science and technology, the sizes of space structures are becoming larger and larger, such as giant antennas and space telescopes. Due to the delivery capacity limitations of space launch vehicles, extremely large space structures are difficult to be deployed in orbit as a whole, but can only be constructed by In-Space Assembly (ISA) technology from separate parts. Roughly speaking, in-space assembly can be divided into two types: human participation and automated assembly. Manned ISA is mainly completed by remote control on the ground or by astronauts in space. There are two kinds of automated assembly in space. The first one uses specific space robots to automatically complete target capture [1,2], transportation and assembly [3,4]. Another kind realizes ISA by autonomous rendezvous and docking between modules [5,6]. Our work is focused on the problem of automated ISA of the first kind.

In the past few decades, NASA Langley Research Center has explored the feasibility of applying automated assembly to space missions which require structures that are larger than available launch vehicles [7]. In recent years, researchers have made remarkable advances in developing Space Assembly of Large Spacecraft Structural System Architectures (SALSSA) [8,9], which has the potential to drastically increase the capabilities and performance of future space missions and spacecraft while significantly reducing their cost. ISA technology plays an important role in decreasing deployment risk, assembling systems too large to fit into a single launch vehicle, and enabling repair and upgrade. There are many possible applications of ISA technology, such as asteroid redirect vehicle, artificial gravity vehicles, and high-resolution space telescope [10]. This indicates a pressing need for investigating robotic ISA and related technologies. So far, different robotic architectures for assembling large space structures have been proposed, such as fixed-to-base self-assembly robot, tracked self-assembly robot, end-over-end walking robot, free-flying assembly robot and docking-to-mobile-track free-flyer robot [11], etc. For example, fixed-to-base self-assembly robot, as shown in Fig. 1(a), is the most fundamental and simplest type for ISA. In this architecture, a robot arm is rigidly fixed on a base satellite to pick and place modular space structures from their stored position to their assembly position. However, for a large structure, it is less versatile due to the limit in its reachability. Compared with fixed-to-base self-assembly robot, free-flying assembly robot has an advantage in sizes of space structures. Fig. 1(b) illustrates the concept of this architecture, which involves two free-flying agents: a free-flying assembler with robotic arm and a free-flying satellite platform where robotic assembly will be performed. The modular components may be stowed in the assembler agent or the satellite platform separate from it, and are usually expected to be cooperative in the assembly process. Prompted by the concepts of these robotic assembly architectures, we aim to present an investigation for demonstrating the feasibility of robotically assembling a large space structure from cooperative modular components with ground experimental validation.

As a relatively economical way to conceptually verify ISA technology, much ground-based experimental work concerning robotic ISA has been reported. For instance, Doggett [12] used a fixed-base robotic arm on a planar-motion base to assemble and disassemble an 8-m structure in which the truss structure was assembled in rings on a rotary motion base. Senda et al. [13] developed an experimental system emulating a free-flying space robot for automatic truss structure assembly where the robot simulator floats on a two-dimensional planar table by air bearings. Jenett et al. [14] presented a robotic truss assembler designed for Space Robot Universal Truss System (SpRoUTS). The overall layout of the system mainly consists of a storage cartridge for flat pack elements and a built structure rising out of the robot. In Ref. [15], Jenett and Cheung presented a relative robotic platform for manipulating a modular 3D lattice structure, in which the robots can operate without external metrology. Wong et al. [16] constructed an autonomous, distributed robotic system to assemble a backbone truss for solar arrays for a solar-electric propulsion tug. In the robotic system, there are a long reach manipulator, a strut and joint manipulator, and a six degree-of-freedom (DOF) precision manipulator which collaborate with each other to finish in-space assembly task. Karumanchi et al. [17] reported on a payload-centric autonomy paradigm and presented results from laboratory demonstrations of automated assembly of structures using a two-armed robot called RoboSimian. The deployable payloads are 1-m trusses that are robotically assembled to form a 3-m diameter kinematically closed loop structure to subcentimeter accuracy. In Ref. [18], Levedahl et al. gave an overview of the current status of Trusselator development, which can address the in-situ fabrication of solar array support structures and achieve the balance of in-situ fabrication and robotic assembly.

The paper aims at conceptually verifying the robot-based ISA technology of large space structures, especially in the aspects of structure design, planning algorithm, and experimental demonstration. More specifically, the main goal of this paper is three-fold. The first is to design a series of cooperative modular components with special connectors, which can be scaled to a larger size and a more complex structure. To facilitate rapid prototyping of the proposed design, all the components are made by low-price, consumer-grade 3D Printing. The second is to develop an automated assembly strategy to fulfill the assembly task based on visual measurement using an economic camera. The strategy covers the technical details of the assembly strategy such as module localization, assembly sequencing and motion planning, etc. The third is to experimentally evaluate the proposed assembly strategy using the developed hardware platform, and to verify whether the accuracy of robotic manipulation satisfies the precision requirement of the assembly task.

This paper is organized as follows. Section 2 presents the design of the components and the final structure constructed by assembly, and introduces the developed hardware system for performing assembly task. In section 3, the assembly strategy is given in detail, including module localization, assembly sequencing and motion planning based on Iterative Learning Control. Section 4 presents the experimental verification of the proposed assembly strategy using the developed hardware platform. Section 5 provides conclusions and future study directions of the work.