2.2.4 Attached Servicers

This segment of the program is focussed on the development of space robotics for on-orbit servicing by systems attached to supporting structures such as the Space Shuttle or Space Station. The purpose of this segment of the program is to focus the development of component technologies into applications and environments which will demonstrate their utility and additional capability when incorporated into operational systems. These technologies include virtual reality telepresence, advanced display technologies, proximity sensing for perception technologies, and robotic flaw detection. The target applications include such tasks as repair of small satellites, ground-based control of robotic servicers, and servicing of external space platform payloads. Each of these areas have been identified by the potential space robotics user community as applications where space robotics will be necessary to satisfy their planned requirements. This user community includes Space Station Alpha, Mission to Planet Earth, and the Space Transportation System.

 

Robotics Technologies for ISSA Maintenance

Increased Autonomy for ISS Dexterous Robotics

Robonaut Robotic Surrogate

Technology Transfer Roadmap Details

Robotics Technologies for ISSA Maintenance

(Note: In FY 1997, this task was transferred to the Office of Space Flight robotics program.)

Objective

The objective of the Space Station Robotic Maintenance Technology Transfer project is to develop and evaluate robotic technology with respect to its potential for meeting the International Space Station's (ISS's) needs. This is a step in the larger goal of identifying, developing, testing and integrating robotic technology into the ISS program.

The intent of the testing covered by this plan is to bridge the gap between testing in a developmental laboratory and integration into the ISS program by testing in a facility which emulates the ISS systems and their operating environment. Information learned from such testing during the early development of robotic technology can be factored back into its design to improve its applicability to the ISS needs. Testing in this environment for more mature technologies can provide needed exposure to the ISS community by demonstrating its usefulness for the ISS.

In parallel with the testing process conducted at the Johnson Space Center, the Jet Propulsion Laboratory will develop component technologies for ISS maintenance which will be transferred to the ARMSS test facility for evaluation.

The Automated Robotic Maintenance of Space Station (ARMSS) facility located in building 9C at NASA/JSC is used for this testing. The ARMSS has two 7-DOF Robotics Research model 1607 manipulators which are configured to emulate the ISS Special Purpose Dexterous Manipulator (SPDM). The ARMSS also has a full scale Space Station Pre-Integrated Truss (PIT) segment plus several cameras which provide a realistic Space Station operating environment for ORU replacement tasks. Improvements to the lighting facilities to better emulate on-orbit conditions are anticipated. The Ground Control Console (GCC) in the Dexterous Robotics Laboratory (DRL)will be the baseline for remote control of the ARMSS during testing. These basic facilities will be augmented by hardware and software from the developers of the technology to be tested.

FY97 Test Plan

The following evaluations will be performed in FY97.

Remote Surface Inspection:

The Remote Surface Inspection (RSI) technology under development by JPL will be tested first. It is based on an Integrated Sensor End-Effector (ISSE) which uses machine vision to visually inspect for surface flaws along with eddy current, proximity, temperature, gas/vapor and force sensors. JPL will supply an ISSE along with controller hardware and software and integrate it into the ARMSS. JSC will develop realistic Space Station scenarios and use them to test this inspection capability. Metrics for evaluating RSI performance will be established as part of the preparation for initial integration and testing.

Calibrated Synthetic Viewing:

An improved version of the Calibrated Synthetic Viewing (CSV) system developed at JPL that used models of objects being manipulated and realtime video data to provide synthetic views of the manipulator workspace will be tested. This technology is intended to enable the conduct of manipulator operations in spite of obstructions that prevent a direct view of, for example, the interface between objects being mated. Upgraded software will be supplied by JPL and installed in the ARMSS and the GCC.

An earlier version of this system was tested in FY96. As in the previous test, the test scenario will involve changing out a Remote Power Conditioning Module (RPCM), an ISS orbital replacement unit (ORU). The surrounding structure will be "mocked up" to represent a typical ISS structure and ISS procedures will be used for the changeout. Initially, the operator will changeout the ORU using live video as a baseline. Then the task will be repeated using only synthetic views after they have been calibrated by the CSV system. Multiple test subjects, all with experience in robot manipulator operations, will be used in the tests. JPL operators with extensive experience in CSV will be used as available. Based on the initial testing, the goal for FY'97 is to reliably position/align an ORU for insertion with errors not exceeding 1/4 in. and 3 deg. and take no more time than twice the duration of the task when performed with traditional telerobotic methods.

While testing of Robotic Control Technologies for Space Station Maintenance was originally proposed, it is not currently planned since this technology is not expected to be ready for testing before the end of FY97. An additional test of CSV is being explored in place of the Robotic Control Technology.

Future Plans

Future plans include:

1. Restesting RSI after making performance improvements identified during initial integration and testing.

2. Investigating the potential of a Shuttle/Station flight experiment evaluation of CSV by installing it in the MDF/MRMDF with additional displays to allow one person to operate the RMS using it while another monitors operation using the normal displays and out the window views or vice versa.

3. Develop and infuse new technology into the evaluation process.

JPL has a strong track record and technology base in robotics for the Space Station, and has developed several enabling robotics technologies such as: dexterous manipulator control, contact control, and the user-robot interface[5]. The Remote Surface Inspection (RSI) task that was performed at JPL during 1990 - 95 is a good representation of the development and integration of these component technologies. JSC, on the other hand, is the lead center in human-related robotics, and its charter includes all aspects of robotics technologies for low-orbit platforms such as the Space Station.

The technical approach adopted in this task is to develop and demonstrate robust contact and motion control technologies needed to perform the Space Station maintenance operations. The robust control schemes will be formulated as "outer feedback loops" so that they can easily be implemented in practice without disrupting the inner position control loops baselined for the Space Station robotic manipulators. The following describes the approach for each technology development.

Robust Contact Control: The control of contact transitions in dexterous robotic manipulators represents one of the most important challenges to the successful implementation of Space Station robotics at the present time. When the end-effector executing a free-space motion impacts a reaction surface, the environmental dynamics seen by the end-effector undergoes a gross and abrupt change instantaneously, posing a challenging transition control problem. Contact transitions can be notoriously difficult to control. These events involve all of the well known problems of force control, with the added difficulties of nonzero approach velocities and discontinuous dynamic characteristics. If a robotic manipulator is to interact effectively with its environment, however, it must frequently make and break contacts with external objects. Compliant fingertips or end-effectors have been used in the past to eliminate some contact transition problems, but many difficult tasks remain. Even with soft fingertips, contact transitions tend to excite instability or result in undesirably high impact forces. These phenomena can be particularly bothersome when manipulating with sensor-laden fingertips, devices that are sensitive to "glitches" in contact forces.

Our particular goal in this task element is to achieve smooth, stable transitions between motion and force control. We want to avoid instability and large force spikes during the controller transition, while increasing the contact force from zero to the desired level as rapidly as possible. Two approaches will be investigated to enhance the robustness of contact control tasks; namely, nonlinear end-effector compliance and adaptive force regulation. Each approach is now described briefly:

Nonlinear End-Effector Compliance: In the conventional compliance control approach, the contact force is measured by a wrist-mounted force/torque sensor and is fed back to modify the reference position command through a predefined constant gain. The effect of this force feedback loop is to reduce the apparent stiffness of the contact surface so that the reference position can be used as a command to control the contact force. In the proposed approach, we explore the use of nonlinear feedback gains where the gain is a function of the contact force, instead of having a fixed predefined value. An innovative feature of the proposed approach is to use sigmoidal functions for feedback gains. The goal is to use these nonlinear force feedback gains to reduce the impact forces and to accommodate the abrupt changes that occur in the end-effector dynamics at impact.

Adaptive Force Regulation: Previous research on force control has focused primarily on linear fixed-gain controllers. For a contact surface with a known constant stiffness, a linear fixed-gain controller can be designed to achieve a desirable force response with small or zero error, low overshoot, and rapid rise time. However, the same controller typically exhibits sluggish response in contact with softer surfaces, and becomes unstable when contacting harder surfaces. In other words, because the stiffness coefficients of different contact surfaces can differ substantially, a fixed-gain controller design based on a nominal surface stiffness leads to a non-uniform dynamic performance and often instability. In this task element, this problem will be alleviated, to a large extent, by employing adaptive elements in the force control scheme. These elements can compensate for stiffness variations and yield a stable and uniform response. Even when the contact surface stiffness is constant and known, adaptive controllers can result in superior command tracking and disturbance rejection performances compared to linear fixed-gain controllers.

The nonlinear end-effector compliance and adaptive force regulation schemes will considerably enhance the robustness of robotic manipulators for executing contact task, which is the basic requirement for successful execution of many robotic maintenance operations on the Space Station.

Robust Motion Control: The intent of this task element is to enhance the robustness of motion control scheme for the Space Station robotic manipulators in the face of parameter uncertainties, payload variations, and unexpected disturbances, as identified in the Space Station Requirements Document. These unpredictable conditions are very likely to occur during the operation of the Space Station manipulators. For instance, the robotic manipulators will have payloads with widely different mass properties, from low mass electronic boards to massive ORU's. It is essential that such gross payload variations do not cause loss of stability or degradation of performance of the manipulator control system. In this task element, nonlinear and adaptive compensation techniques will be developed for enhanced robustness. The performance of the system with and without these compensators will be demonstrated and compared to quantify the benefits.

Dissemination of Technology

Testing and evaluation reports will be provided to the technology developers, the TRIWG, the ISS Program, the JSC Engineering Directorate, Missions Operation Directorate and Crew Office and also made accessible on the Internet in an appropriate place in the NASA telerobotic technology pages.

In addition to demonstrations, a short video tape of each technology evaluated will be produced. The video will highlight the benefits and limitations of the technology as tested. Small portions of this video could be made available on the Internet.

Milestones:

Oct 1996 - RSI Integration into ARMSS
Dec 1996 - RSI Test plan and procedures developed
Mar 1997 - RSI testing on ARMSS complete
Mar 1997 - RSI evaluation report delivered
Mar 1997 - CSV Integration into ARMSS
Apr 1996 - CSV Test plan and procedures developed
Jun 1997 - CSV testing on ARMSS complete
Jun 1997 - CSV evaluation report delivered
Jul 1997 - Shutdown ARMSS for overhaul and relocation
Q2 1997 - Baseline the existing contact transition and contact force control performance.
Q3 1997 - Develop nonlinear contact transition control scheme.
Q4 1997 - Implement and demonstrate nonlinear transition control capability in RSI lab (Level One).
Q2 1998 - Develop and demonstrate adaptive contact force control scheme in RSI lab.
Q3 1998 - Transfer new contact transition and contact force control schemes to JSC.
Q4 1998 - Demonstrate and quantify benefits of new transition and force control schemes for performing ISS contact tasks at JSC (Level One).
Q2 1999 - Develop robust motion control scheme.
Q4 1999 - Demonstrate robust motion control for ISS maintenance tasks at JSC (Level One).

Point of Contact:
Kenneth Baker
Johnson Space Center
Houston, TX 77058
(281) 483-2041
Kenneth.baker1@jsc.nasa.gov

Increased Autonomy for ISS Dexterous Robotics

In FY 1997, this task was transferred to the Office of Space Flight robotics program.

Point of Contact:
Edith Taylor
Johnson Space Center
Houston, TX 77058
(713)483-1527
etaylor@gp301.jsc.nasa.gov

Robonaut Robotic Surrogate

(Note: In FY 1997, portions of this task were transferred to the Office of Space Flight robotics program.)

The objective of this task is to develop a robotic astronaut surrogate to perform high-payoff EVA tasks and provide "minuteman" responsiveness for EVA contingencies. According to the current International Space Station (ISS) EVA Concept of Operation Document (JSC-33048), the majority of an EVA sortie will be spent on worksite setup and close-out. To maximize overall EVA productivity and to increase crew safety, a robotic astronaut surrogate could perform these "overhead" tasks in place of an astronaut. Also a robotic astronaut surrogate could be used as a "helper" astronaut working alongside a suited EVA astronaut, freeing up the second EVA astronaut to perform a "parallel" task, thereby reducing the total EVA time required for the job. Another needed capability missing from the current EVA concept of operation is the "minuteman" responsiveness to EVA contingencies. The current space suit technology requires the astronauts to undergo hours of pre-breath and checkout before each EVA. This level of responsiveness may not be adequate for some contingencies. An attractive alternative is to station a robotic surrogate outside the spacecraft, so that a crew member can "go EVA" without delay via immersion telepresence.

The planned Space Station robots (SPDM and SSRMS) are poor surrogates for an astronaut because they (1) require special alignment targets and grapple fixtures on everything they handle; (2) are too large to fit through tight EVA access corridors; and (3) do not possess adequate speed and dexterity to handle small and complex items, or soft and flexible materials, or even common EVA interfaces. Furthermore, the teleoperator controls for these robots, which usually consist of flat panel displays and joystick-like hand controllers, are grossly inadequate for coordinating the high level of dexterity inherent in complex EVA tasks. Therefore, to meet the demanding task requirements, a highly dexterous telerobot with an intuitive human control interface must be developed. The Robonaut will be designed to meet these demanding requirements.

Robonaut has applicability to all future EVA missions, including Shuttle, Space Station, and planetary missions. The most immediate application will target the International Space Station, as directed by the Code M-sponsored External Work System Program.

Technical Approach:

System Concept

The Robonaut concept (shown above) calls for two 7 degrees of freedom (DOF) arms, two 12 DOF multi-finger robotic hands, a 6+ DOF "stinger tail", and a 2+ DOF stereo camera platform. The robotic arms are capable of dexterous, human-like maneuvers, and are designed to ensure safety and mission success. The robotic hands are designed to handle common EVA tools, such as an ORU handling tool (a.k.a. "ice cream scoop"), to grasp irregularly shaped objects, and to handle a wide spectrum of tasks requiring human-like dexterity. For stabilization, the Robonaut will have a "stinger tail" that can plug into WIF sockets conveniently located around the ISS. Potentially, the Robonaut can be carried by the Crew Equipment Translation Aid (CETA) to various ISS work sites, or can be picked up by the SSRMS to perform various "end-of-arm" tasks. The Robonaut will be teleoperated by an IVA crew member using telepresence equipment, such as a head-mounted display (HMD), tracker sensors, virtual reality gloves, or force-reflective arm and hand masters. The stereo camera platform will provide stereo video images to the human operator and to the on-board vision system.

The Robonaut development will be divided into four phases. Robonaut I will focus on developing an integrated dexterous arm-hand module and evaluate the module against a set of one-handed EVA tasks. The concept of telepresence control will also be evaluated along with the arm-hand module. Robonaut II will focus on developing a full torso system with two arm-hand modules and an articulated stereo camera head. Robonaut III and IV will be derivatives of the Robonaut II system, each with added features for ISS and planetary surface mobility, respectively.

A key component of the overall Robonaut concept is the integrated dexterous arm-hand module, which is the main focus of Robonaut I. Different from most laboratory robotic arms and hands, the Robonaut I arm-hand design will not only address functional and operability issues, but will also attempt to address space flight issues, such as fault tolerance, safety, thermal extremes, vacuum, radiation, vibration, and mass and volumetric constraints.

The dexterous hand will have four fingers and a thumb in a human hand-like arrangement. The thumb, index, and middle fingers will each have three independent degrees of freedom. The ring and little fingers will each have one degree of freedom. The dexterous hand will also have a palm degree of freedom which allows the palm to arc (as viewed from the end of the fingers). Extensive human hand kinematics analysis has revealed that the palm degree of freedom is essential for grasping and using hand tools. The hand will be driven by twelve miniature brushless motors tightly packaged in the forearm. The forearm will also house two motors that drive the wrist P-Y degree of freedom. As a safety measure, the hand will be non-backdriveable to eliminate any chance of inadvertent release. For fault recovery, an EVA crew member can release the robotic hand through one-handed manual backdrive mechanism. The fingers will also have a compliance feature to minimize the impact force stemming from inadvertent contact with the surrounding structures.

The dexterous arm closely approximates that of an EVA crew member (approximately 24 inches of reach). The current arm design has a R-P-R-P-R-P-Y configuration (R-roll, P-pitch, Y-yaw), with elbow self-motion capability to optimize strength and avoid collisions. The final P-Y joints will be part of the hand-wrist mechanism. For safety and thermal control, the arm will be covered by soft non-outgas materials.

Technical Challenges

Several technical challenges will be addressed during the Robonaut I development. The Robonaut arm-hand module will be designed to securely handle a variety of EVA tools and interfaces currently too difficult or impossible for the baseline Space Station robots. Integrating and packaging the functionality of human arm and hand within the volumetric constraints will also be a challenge. Clever mechanisms and advanced sensors and actuators will have to be developed and utilized. The arm and hand designs will also have to consider safety and environmental issues associated with space flight.

Robonaut I: Integrated arm-hand module (FY97-98)

The module will have a 12 DOF dexterous hand, 2 DOF wrist, and 5 DOF dexterous arm, all arranged in a human-like configuration to emulate, as a minimum, a suited crewmember's reach and dexterity, as according to NASA STD-3000, EVA Hardware Generic Design Requirements Document (JSC-26626), and U.S. On-Orbit Segment Specification (SSP-4162).

The dexterous hand will have 20 pounds of grip force and will weigh less than 10 pounds total, including the wrist and forearm; the hand will move through its full range of motion in less than 1.0 second; all fingers and wrist are shock-mounted to absorb inadvertent impact or overloads; manual grip release mechanism will be incorporated for fault recovery. The dexterous arm will have 24 inches of reach from shoulder to wrist, which will be 1/4 of the SPDM length; the most proximal 5 DOF will weigh less than 50 pounds; the arm will have EVA-equivalent strength and will absorb up to 3 Joules of contact energy. All electronics will be contained within the arm-hand module and a 1.0 cubic foot volume at the shoulder. The arm-hand module will operate in the temperature range of -40 to 100 C in a vacuum environment; the electronics in the shoulder volume will operate in the temperature range of -40 to 85 C.

With appropriate form of control, the Robonaut I is expected to be able to work with EVA tether hooks, install and remove portable foot restraints (PFR), pull Velcro tabs to remove multi-layer insulation (MLI), mate and demate electrical connectors, grasp and actuate ORU handling tools, and complete these tasks in less than twice the time it would take an astronaut. The IGD will demonstrate the arm-hand module under telepresence control performing a subset of the tasks described above.

Robonaut II: Two arm-hand modules and a camera platform (FY99-00)

Robonaut II will consist of two symmetrical Robonaut I arm-hand module mounted on a torso in an anthropomorphic configuration. The camera platform will have more than 2 DOF for pan and tilt, and will provide an egocentric view of the robot's surrounding consistent with or better than that of an EMU helmet.

With appropriate form of control and with the assistance of an RMS-like robot, the Robonaut II is expected to be able to do all that tasks that Robonaut I can perform plus install antenna boom assembly, changeout ORUs, reposition articulated portable foot restraints (APFR), setup portable work platform (PWP), relocate EVA tool stowage device (ETSD), remove some of the unique launch restraints on ULC/SLP, and complete these tasks in less than twice the time it would take an astronaut.

The IGD will demonstrate the Robonaut II system under immersion telepresence control performing a subset of the tasks described above.

Robonaut III: Robonaut II with a stabilizer arm (FY01)

Robonaut III will consist of Robonaut II plus a stabilizer arm to interface with, as a minimum, the worksite interface sockets (WIF). The IGD scenario is being planned.

Robonaut IV: Robonaut II with a surface mobility system (FY02)

Robonaut IV will consist of Robonaut II plus a planetary surface mobility system. The IGD scenario is being planned.

FY 97 Activities

FY 97 activities will focus on development, design, fabrication and testing of the hand-wrist module; arm Kinematic design; joint design; pitch and roll joint fabrication and testing; and perception technology development.

Hand-wrist module development activities begin with completion of prototype engineering analysis and drawings. A functional plastic model will then be fabricated. Analysis of the plastic model will lead to final drawing revisions prior to beginning fabrication of the first fully functional metal hand-wrist module. Finger fabrication will be completed first. Finger testing will be conducted to characterize and verify performance. Fabrication of the remaining wrist components will follow, with integrated hand-wrist module characterization and performance verification completed by years end.

Arm development activities begin with Kinematic analysis of the proposed configuration. The analysis will consider single and dual arm operations for the specific EVA tasks defined in the "Robonaut, External Work Systems Technology Development Proposal". The analysis will seek to optimize the Robonaut arm design for reach, strength for a typical EVA sortie. Additional analysis will be conducted addressing power and thermal dissipation performance for the sorties.

Arm joint design activities will focusing on developing space hardened components that will converge with the top-down requirements to produce a mission relevant robot that is technically feasible. Manipulator components tested and selected for Robonaut include thermal vacuum rated harmonic drives and lubricants, redundant optical position sensors, thermally compensated torque transducers, DC brushless motors with skewed cores, and Tefzel cable harnesses with Kevlar fabric coverings.

Following joint design activities, fully functional pitch and roll joints will be fabricated and tested. Testing will cover characterization and performance verification.

Early development of perception technologies, which will be incorporated into Robonaut II, will begin in FY97. Stereo vision technologies will be developed to detect obstacles within a spherical region of space in front of the stereo vision cameras, calculate distance to an object, and to provide stationkeeping and object tracking with respect to a moving object. Optical correlation techniques will be enhanced to improve pattern recognition of "natural" environmental features and determine object pose without the aid of specially designed targets. More robust means of calculating precise pose estimations with monocular vision techniques will also be explored. These technologies will be first utilized in the AERCam II Integrated Ground Demonstration scheduled for the end of October 1997.

Major Milestones:

FY 1997:

The FY97 activities will be focusing on dexterous robotic hand development and initial dexterous arm design.

2/97 Hold the first design review of the dexterous robotic hand

4/97 Conduct single finger test to verify the design parameters

9/97 Conduct arm joint test to verify the arm design parameters

10/97 Integrate the dexterous robotic hand and conduct functional tests

12/97 Hold FY97 task review and demonstration

FY 1998:

The FY98 activities will be focusing on the dexterous arm development completion and the integration of the dexterous arm and hand.

Point of Contact:
Edith Taylor
Johnson Space Center
Houston, TX 77058
(713)483-1527
etaylor@gp301.jsc.nasa.gov