Automation and robotics have played important roles in space research, most notably in planetary exploration. While an increased need for automation and robotics in space research is anticipated, some of the major challenges and opportunities for automation and robotics will be provided by the Space Station. Examples of these challenges are briefly reviewed.
NASA is currently considering the establishment of a Space Missions for Automation and Robotics Technologies (SMART) Program to define, develop, integrate, test, and operate a spaceborne national research facility for the validation of advanced automation and robotics technologies. Initially, the concept is envisioned to be implemented through a series of Shuttle-based flight experiments which will utilize telepresence technologies and real-time operational concepts. However, eventually the facility will be capable of a more autonomous role and will be supported by either the Shuttle or the Space Station. To ensure incorporation of leading-edge technology in the facility, performance capability will periodically and systematically be upgraded by the solicitation of recommendations from a user advisory group. The facility will be managed by NASA, but will be available to all potential investigators. Experiments for each flight will be selected by a peer review group. Detailed definition and design is proposed to take place during FY 86, with the first SMART flight projected for FY 89. This paper discusses the objectives and rationale for the proposed SMART Program, potential implementation scenarios, and the management approach. The main purpose of the paper is to make the reader aware of this upcoming program, and to encourage participation beginning with the concep-tual definition phase.
The National Aeronautics and Space Administration has initiated a major effort to establish a permanent manned presence in space to expand the exploration and use of space for activities that enhance the security and welfare of mankind. In particular, Space Station would:
- Establish a means for the permanent presence of people in space.o Provide a capability for routine, continuous utilization of space for science, applications, technology development, commercial utilization, national security and general operations. - Lead to the development and exploitation of the synergistic effects of the man/machine combination in space. - Provide for essential system elements and operational practices for an integrated national space capability. - Reduce the cost and complexity of living in and using space.
The progress of technology is marked by fragmentation -- dividing research and development into ever narrower fields of specialization. Ultimately, specialists know everything about nothing. And hope for integrating those slender slivers of specialty into a whole fades. Without an integrated, all-encompassing perspective, technology becomes applied in a lopsided and often inefficient manner. A decisionary model, developed and applied for NASA's Chief Engineer toward establishment of commercial space operations, can be adapted to the identification, evaluation, and selection of optimum application of artificial intelligence for space station automation -- restoring wholeness to a situation that is otherwise chaotic due to increasing subdivision of effort. Issues such as functional assignments for space station task, domain, and symptom modules can be resolved in a manner understood by all parties rather than just the person with assigned responsibility -- and ranked by overall significance to mission accomplishment. Ranking is based on the three basic parameters of cost, performance, and schedule. This approach has successfully integrated many diverse specialties in situations like worldwide terrorism control, coal mining safety, medical malpractice risk, grain elevator explosion prevention, offshore drilling hazards, and criminal justice resource allocation -- all of which would have otherwise been subject to "squeaky wheel" emphasis and support of decision-makers.
This paper describes results of a project to build a prototype expert system for automated fault isolation and correction of a regenerative CO2 removal device that is typical of functions of the air revitalization group in the Space Station environmental control and life support system (ECLSS). The software was developed using one of the powerful commercial knowledge engineering environments. The goal of the project was to evaluate the feasibility of using a software development environment to rapidly design, construct, test, and change expert system software. This paper discusses the use of expert systems to enhance automatic controllers, and the use of information on device design and on device troubleshooting and repair procedures in developing expert systems. This paper also describes the development of the prototype expert system and presents results of the evaluation.
Space Station will require a tremendous increase in autonomous power management capabilities over previous spacecraft. America's first space station, Skylab, was operational from July 1973 until March 1974. The eight kilowatt electrical power bus required fifteen ground support personnel for monitoring, analysis, and control as well as extensive periods of onboard crew involvement. In contrast, the Initial Operational Configuration (IOC) Space Station has a requirement for 75 kilowatts of primary power distribution while the growth Space Station will require 300 kilowatts. It is anticipated that the Common Modules will each have the capability of managing up to 50 kilowatts of power; 25 kilowatts to be routed to adjoining Common Modules or attached payloads and 25 kilowatts for consumption within the Common Module. Minimization of crew involvement and ground support is a critical requirement for the complex Common Module electrical power system. The goal is to make this system as autonomous as is practical. Expert systems are envisioned to play a critical role in the electrical power system in both the IOC and growth versions of Space Station.
This paper describes the development of an expert system for defining and dynamically updating procedures for an orbital rendezvous maneuver. The product of the expert system is a procedure represented by a Moore automaton. The construction is recursive and driven by a simulation of the rendezvousing bodies.
This paper addresses the basic operating considerations resulting from the unique environment of space along with those resulting from the current and projected state of Automation and Robotics (A&R) which will influence the initial layout and maintenance of the Space Station. In this paper, we introduce a concept called "robot-factors" which deals with the telerobot working environment and its organizational relationships with other robots. Robot factors are discussed in this paper from the point of view of the overall system architecture of the Space Station. That is, we present basic design considerations concerning the physical nature of the Space Station complex as well as those concerning the data management system. In many ways, robot factors is quite analogous to conventional human factors. The emphasis of the study is on making the robot's tasks safe and easy to perform. It is also on the concern of the telerobot's welfare in terms of that of other cooperating telerobots in the performance of a common task.
Automation of the space station is necessary to make more effective use of the crew, to carry out repairs that are impractical or dangerous, and to monitor and control the many space station subsystems. Intelligent robotics and expert systems play a strong role in automation, and both disciplines are highly dependent on a common artificial intelligence (Al) technology base. The AI technology base provides the reasoning and planning capabilities needed in robotic tasks, such as perception of the environment and planning a path to a goal, and in expert systems tasks, such as control of subsystems and maintenance of equipment. This paper describes automation concepts for the space station, the specific robotic and expert systems required to attain this automation, and the research and development required. It also presents an evolutionary development plan that leads to fully automatic mobile robots for servicing satellites. Finally, we indicate the sequence of demonstrations and the research and development needed to confirm the automation capabilities. We emphasize that advanced robotics requires AI, and that to advance, AI needs the "real-world" problems provided by robotics.
The results of a National Aeronautics and Space Administration (NASA) sponsored study, performed in order to establish the feasibility of remotely manipulated or unmanned welding fabrication systems for space construction, are first presented in this paper. Possible space welding fabrication tasks and operational modes are classified and the capabilities and limitations of human operators and machines are outlined. The human performance in remote welding tasks is experimentally tested under the sensing and actuation constraints imposed by remote manipulation in outer space environments. Proposals for the development of space welding technology are made and necessary future research and development (R&D) efforts are identified. The development of improved visual sensing strategies and computer encoding of the human welding engineering expertise are identified as essential, both for human operator assistance and for autonomous operation in all phases of welding fabrication. Results of a related follow-up study are then briefly presented. Novel uses of machine vision for the determination of the weld joint and bead geometry are proposed and implemented, and a first prototype of a rule-based expert system is developed for the interpretation of the visually detected weld features and defects.
A general conceptualization of Khatib's potential-field approach to manipulator collision avoidance is presented. The potential-field approach is shown to consist of four algorithms: a. repulsion away from obstacles, an attraction towards a goal, a method of combining these, and a resulting method of incrementing the arm. Alternatives for these algorithms are discussed. A multiple-robot system demonstrating the concepts is presented. The system uses a detailed rigid model of the entire arms and surrounding objects to avoid collisions. The system operates in close to real time, and is demonstrated with two PUMA robots moving concurrently. Results are applicable to any type of anthropomorphic arm, including the Remote Manipulator System.
The goal of this research project is the development of a computer vision system to monitor and control life sciences experimentation on hoard space stations. The vision system is organized as a multi-processor system with distributed processes selectively analyzing hierarchical imagery in order to monitor and control the appropriate instrumentation.
Video data is used in a wide variety of computer vision tasks. Applications range from mail sorting to medical diagnostics to industrial inspection. For Space Station applications, however, video imagery has certain limitations. Outside a spacecraft the ambient illumination and viewing background can cause problems for a video system. Identifying a satellite at an unknown attitude and distance may be very difficult to do with 2D imagery. Consequently, investigators are looking at other sources of data to supplement or replace video data for vision tasks on the Space Station. Laser systems can provide range information, and laser scanners can provide reflectance and depth information in image format. Yet other approaches are being considered. This paper will discuss some of the advantages of the different approaches in the context of anticipated Space Station applications. The issues associated with the problem of integrating data from various sources to most effectively and efficiently accomplish a given vision task will also be addressed.