Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
9 Robotics, Automation, and Artificial Intelligence BACKGROUND AND STATUS In 1987, the National Research Council report Space Technology to Meet Future Needs stated that "the time has come to add a new technology, automation and robotics, to the other major technologies propulsion and power, materials, and information management that are considered essential to U.S. capability to operate effectively in space. There are three reasons: affordability, achievability, and need" (NRC, 19871. Since that report, the technology is no longer "new." Much has been done, but it is still young relative to other space technologies. However, the reasons for its application are becoming even stronger. The application of automation and robotics to NASA's mission is in a state of transition. NASA is increasingly moving toward the use of small spacecraft and associated systems and technology that may minimize the need for robot-assisted servicing. At the same time, the servicing and maintenance of major projects such as the Hubble Space Telescope and the Space Station are being planned using teleoperated robotic arms. This section is directed toward the following applications for robotics, automation, ant} artificial intelligence as they relate to small spacecraft applications that the pane! believes offer great potential payoffs: small microrovers capable of rapid deployment and numerous operations, for robotic exploration of Mars; and small robotic spacecraft and intravehicular robotics, which can contribute to crew productivity and the health and maintenance of major space assets such as the Space Station and the Hubble Space Telescope. In addition, the pane! believes that increased applications of computer-based systems will benefit all aspects of mission operations, including training and simulations. These systems are discussed in Chapter 2 of this report. 70
Robotics, Automation, and Artificial Intelligence SMALL ROBOTIC PLANETARY EXPLORERS In the area of planetary exploration, small spacecraft systems are already being developed. For example, the NASA Mars Pathfinder project scheduled for launch in 1996 will employ an innovative microrover (termed "Rocky") operating in the vicinity of a Mars landing craft. This craft serves as a science base and as the communications center with Earth. With the ability to operate somewhat autonomously in the vicinity of the lander craft, Rocky can significantly increase the total returned knowledge. The Mars Pathfinder baseline is for remote task control, where an Earth-based operator can observe the three-dimensional environment of Rocky and program in the next sequence of maneuvers. Rocky's perceptive ability is based on a laser line-striping scanner. However, its task-control ability is limited by an 8-bit processor and the associates} limited memory. Rocky has sensors for operation as well as limited spectral-analysis capability for rock samples. The potential for limiter! autonomous operation after the completion of the principal mission objectives is being considered for Mars Pathfinder. However, there is concern that the microrovers will prove to be limited in overall performance due to technological deficiencies identified in this report. The NASA Mars Pathfinder project is the focal point for all on-going rover technology development. Currently, fully autonomous rovers for future missions are being developed, but due to limited resources, the work has not reached a stage of maturity where it can confidently be included in the next flight project. Technology for autonomous mobility exists at various locations, including several NASA Centers ARPA, and Carnegie Mellon University. Overall, the development of small automated science instruments and of m~cromanipulators with capabilities similar to Flight Telerobotic Servicer-cIass systems and suitable for use in small rovers is lagging behind the rest of the spacecraft systems. Small, capable stereo-vision systems do not exist. Unless technology advances are macle, the capability of future, proposed programs such as the Mars Surveyor program consisting of orbiters, ground stations, and microrovers, will be limited by the inability of the rovers to perform important exploration tasks, such as sample collection' preparation, and even limited analysis. Major enhancements in the ability of microrovers can be made by investing in the development of small calibrated science instruments and small robotic manipulators capable of extended operations and improved analytical skills on the planets. Fully autonomous microrovers capable of independent exploration and reporting could turn a network of small rovers into a powerful research tool. This would minimize the impact of the time delay between the Earth and Mars on mission planning and scientific results. The ultimate long-term success of these remote explorers will result from the aggressive application of autonomous operation in an unstructured environment. Since current rover designs use solar arrays to generate power, the ability to fully implement a truly productive, planetary microrover exploration system is now directly related to the amount of sunlight available at the planet to be explored. For the Mars 71
72 Technology for Small Spacecraft Pathfinder program, the limited power available to Rocky from its solar power generation system seriously limits its effectiveness, especially at higher latitudes of Mars. In the future, lightweight, low-cost radioisotope power systems could be enabling technologies for planetary exploration with microrovers. Other technology developments recommended throughout this report, such as miniaturized guidance and control and communication components; high-capacity, lightweight computers; and advanced materials and structures, could also contribute to the effectiveness of these small robotic spacecraft. One way to achieve planetary surface technology development objectives in the near term while accomplishing significant scientific objectives, would be to conduct experimental work on the surface of the moon. This could be done with much shorter flight times than those required to travel to Mars, provide higher solar power, and permit greatly expanded communications. SMALL ROBOTS IN LOW EARTH ORBIT The Space Station is a major investment in space infrastructure, in which a limited number of humans must be provided with systems that will help improve their efficiency. The proper application of automation and robotics can improve the return on this investment by freeing the crew from repetitious tasks and allowing for more direct involvement of ground-based researchers in mission execution via teleoperations. Within the research environment of the Space Shuttle and the Space Station, small intravehicular activity robots such as the German ROTEX on the 1993 Space Laboratory mission can turn a limited flight opportunity into a productive research project. The automation of human-tended teleoperatec! space-based investigations is within the technical capability of university and industry investigators. To date only the Germans have demonstrated its utility in space. Small robotic spacecraft could reduce, and in some cases eliminate, the need for extravehicular activity and ShuttIe-related operations. In the vicinity of the Space Station, small, free-flying robots could be programmed for autonomous or teleoperated inspection of critical Space Station systems as an integral part of repair and maintenance. The concepts and technology base exist within NASA, the universities, ant! industry to develop autonomous systems that are efficient and fault-tolerant with respect to human safety needs. These new, small space robots can become unique, relatively Tow-cost tools for the crews of the Space Station, and they could help bring research productivity more in line with earlier (1980) expectations involving larger crew complements. Perception research is needler} for recognizing robot tasks ant! positioning, controlling, ant! safeguarding devices used in extravehicular and intravehicular activities. Task-contro! research to achieve robotic competence via task-control techniques, task diagnostics, and fault recovery strategies should be conducted. In abolition, research on mechanical systems is needed to create miniaturized, low-power, reliable robot components and complete servicing robots. Finally, flight opportunities to evolve competent servicing robots should be made available. Several NASA centers, including Ames Research Center, GSFC, JPL, and the Johnson Space Center, are working on task
Robotics, Automation, and Artificial Intelligence control techniques such as robot task planners, sequence safeguards, user displays, and interfaces. Much progress has been made in this area since the 1987 National Research Council report (NRC, 1987~. Small, robotic free flyers capable of performing autonomous, as well as teleoperated, inspection and maintenance are being researched at the University of Maryland, Johnson Space Center, and JPL. The only project with a planned launch is the University of Maryland's Ranger Robot, which is not scheduled to interact with a manned spacecraft during its flight. Such robotic systems employ single-string designs and software. The integration of autonomous inspection and maintenance into the manned program has not occurred. Work to date is limited to concepts and laboratory demonstrations. Initiating the required research and development now could make systems available for insertion incrementally during the projected 10-year life of the Space Station. OTHER AGENCY PROGRAMS Beyond NASA, there is much automation and robotic work supported by DOE, DoD, ARPA, BMDO, the National Science Foundation, and industry. For the most part, these programs are focused on point solutions for repetitive problems such as manufacturing or hazardous-waste cleanup, rather than on changing tasks and space environments that are characteristic of most NASA applications. ARPA is supporting an unmanned ground vehicle development for teleoperated and autonomous vehicle navigation for the U.S. Army and U.S. Marine Corps. The current support of space- based automation and robotics research and development is almost entirely funcled by NASA. Much of this development can apply to small space systems. FINDINGS AND PRIORITIZED RECOMMENDATIONS Recommendations for NASA technology development for spacecraft subsystems that are also applicable to small robotic spacecraft are addressed in other sections of this report. The recommendations included here apply to technologies specific to small robotic spacecraft or peculiar to the robotic technology. Technology work related to autonomous operations in unstructured environments should be supported and expanded. 2. Autonomous systems and artificial intelligence should be developed for application to microrovers. ~ A research and development program focused on miniaturizing robotic devices, science instruments, and associated computing power should be developed. 73
74 Technology for Small Spacecraft 4. Robotic spacecraft systems incorporating the most advanced autonomous systems and artificial intelligence technology currently available should be developed for demonstration in space on small spacecraft and on the Space Shuttle. The technology should be applied to the development of a free-flying robotic spacecraft for inspection, maintenance, and research support on the Space Station. Beyond the NASA-unique advantages, the continued investment in automation, artificial intelligence, and robotics is in the broader national interest. Such systems have already changed the way we work and how goods are manufactured.