NASA Langley Research Center, Hampton, Virginia
Controls-Structures Integration (CSI) technology involves the use of active and/or passive systems to enable or enhance the stability and pointing accuracy of disturbed systems. The application of this technology to the Shuttle Remote Manipulator System (RMS) is aimed at improving RMS operational timelines by actively damping the flexible response of the RMS. This is done through the addition of end-point accelerometers to the RMS which sense the low frequency flexible modes of interest in three axes. This signal is then fedback to a closed-loop control law which commands the joint servos to actively damp the response following normal payload maneuvers and Shuttle thruster firings. This Active Damping Augmentation or ADA, offers potential benefits to a variety of manipulator systems including systems that might be installed on the redesigned space station or other spacecraft.
The dynamics of the RMS, which are low frequency and lightly damped, are largely due to flexibility of the two long links, joint compliances, and, to a lesser extent, flexibility of the Shuttle interface. These easily excited dynamics cause vibratory oscillations following normal maneuvers and payload handling operations which are slow to decay. Because of the vibrations, maneuvering of the RMS with attached payloads is intentionally slow and deliberate. To avoid exciting undesired dynamic response, astronauts operating the RMS are trained to use hand controller inputs which reduce RMS start and stop transients, and to wait between hand controller inputs for transient motions to decay. The result is increased payload deployment and handling times and reduced operational flexibility.
Early analysis efforts determined the feasibility of controlling the flexible dynamic response of the RMS using minimum hardware modifications. The results established the need for additional end-point sensors, in the form of accelerometers, to measure the fundamental bending mode of the RMS which dominated its tip motion. The approach to active damping augmentation which has evolved out of this earlier work is shown in Figure 1. The established strategy for implementing ADA compensators in the RMS software was to maintain the way the RMS "feels" to a trained operator, to localize the modifications to the greatest extent possible, and to adhere to the nominal health and safety monitoring of the RMS (including the payload dependent joint rate limits).
ADA has been applied to the RMS with three different attached payloads ranging in mass from very light to heavy. These payloads include an astronaut in the Manipulator Foot Restraint (MFR), the Shuttle Pallet Satellite (SPAS), and the Hubble Space Telescope (HST). The dominant low frequency mode of the RMS with these attached payloads ranges from 0.33 Hz for the 550 lb. MFR, to 0.056 Hz for the 23,900 lb. HST.
Multi-input/multi-output linear state-space models were derived for each of the attached payloads in a set of three different point-design arm confirgur-ations. The models were synthesized using system identification techniques and time response data from the real-time, astronaut training, Systems Engineering Simulator (SES) at the NASA Johnson Space Center. The system identification methodology used in allowed a state-space model and a non-optimal observer to identified directly from the time response data. The model, augmented with cor-responding observer, was then used in ADA compensator design.
Two types of ADA control laws were designed. The first were point-design compensators in which the RMS geometry was fixed and system nonlinearities (e.g., flexibility of structural members, joint servo motor dynamics, gearbox compliance, and data transfer delays) were modeled as white noise. The second type was a single 'global' controller which addressed the kinematic nonlinearities of large joint angle maneuvers. Several different control law design methods were used to establish candidate ADA and control laws. Each of the approaches started with control laws designed to operate at the respective point-design configurations and worked toward a design that achieved stable performance over a set of possible geometric configurations.
Numerous preliminary performance evaluations of the ADA point-design and global control laws were made to address maximum valid arm motion ranges, stability margins, gain scheduling, loads reductions, effects of RMS nonlinearities, and computer signal time-delays. Following these evaluations, the ADA compensators were implemented and tested in the SES. Several mod-ifications were made to the SES to accommodate these tests.
Human-in-the-loop tests of ADA compensators were conducted in the SES's aft flight deck mock-up. This mock-up provides functional displays and controls for human-in-the-loop interaction with a computer simulated environment. Electronic scene generators give the operator visual cues that would be available on-orbit.
Astronaut evaluations of the RMS with the attached SPAS payload were conducted in the SES in September 1992. During this evaluation, the astronaut operators recommended investigation of ADA benefits to heavier payloads where oscillations make maneuvering and positioning of payloads more difficult. Based on this recommendation, as well as relevant mission requirements and engineering judgment, active damping of the RMS with attached heavy (HST) and very light (MFR) payloads was subsequently conducted in October 1993.
For each payload, astronaut operators independently evaluated the performance of a single ADA control law following single-joint and coordinated six-joint translational and rotational maneuvers, as well as ADA disturbance rejection of Shuttle thruster firings. Quantitative results were also derived from recorded simulator data.
The astronauts' assessment of RMS ADA for each of the three payloads was generally favorable. The different operational styles were reflected in the range of comments received. While some of the operators felt that ADA's operational benefit was limited, others called it "a big improvement" and said that the RMS "definitely damps out faster". Damping of the very light (MFR) and mid-weight (SPAS) payload responses were viewed as more dramatic than with the heavier (HST) payload. This was due, in part, to the significantly smaller joint rate limits imposed on the HST payload (0.18 ft/sec as compared to 0.36 ft/sec and 0.77 ft/sec for the SPAS and MFR, respectively). The potential of ADA to aid in maneuvering payloads close to obstructions or during precision tasks was noted as significant.
The performance of an ADA compensator derived for the low hover arm configuration with the SPAS payload is shown in Figure 2. The plot of y-axis acceleration demonstrates the improved damping of the RMS following the manual mode maneuver as well as disturbance rejection of the thruster doublet. Additional demonstrated benefits of ADA include measurable reductions in vibration decay times and peak response following (single joint and six-joint coordinated) RMS maneuvers and Shuttle thruster firings, reduced RMS loads, improved payload positioning steady-state error, and transparency to the operator.
Future space missions will rely on robotic manipulators to perform a variety of tasks such as the complex construction of a space station. RMS ADA has contributed to the research base needed to understand the dynamics and control of space-borne manipulators and to meet evolving mission requirements.
RMS ADA has been successfully demonstrated in a real-time training simulator for a range of payload configurations. Significant improve-ment in damping performance has been achieved through the addition of simulated accelerometers near the end of the RMS. While such end-point sensors are not yet available on the Shuttle arm, increased demands for high precision and improved payload handling capabilities will necessitate active damping and the associated sensors on future space-borne manipulators.
Point of Contact:
Michael Gilbert
Mail Stop 230
NASA Langley Research Center
4b West Taylor Street
Hampton, VA 23681-0001
804-864-2839
m.g.gilbert@larc.nasa.gov
Telerobotics Program
page
Please email the site webmaster
with any comments, criticisms or corrections
for this page.
Maintained by: Dave Lavery
Last updated: May 10, 1996