TELEROBOTIC FLIGHT EXPERIMENT
Joseph D. Graves
Space Systems Laboratory
Department of Aerospace Engineering
University of Maryland
College Park, MD 20742
In the future the demand for space operations is projected to increase dramatically, requiring the development of innovative ways to accomplish work in space. One method under consideration to meet this demand is the use of robotic systems, with capabilities ranging from simple teleoperation to complete autonomy. The Ranger Telerobotic Flight Experiment is designed to demonstrate in the near term the ability of a free flying telerobotic system to perform many required operational tasks including on-orbit refueling, instrumentation package replacement, and deployment of failed mechanisms such as antennae and solar arrays. In addition, this experiment will include the development and operation of a functionally equivalent neutral buoyancy vehicle, to quantify the utility of the neutral buoyancy environment for simulation of on-orbit robotic operations. By combining current robotic technology with a free-flying spacecraft bus, Ranger embodies a new class of highly capable space vehicles that will help meet the demand for future space operations.
The Space Systems Laboratory has a unique background which has led to the development of the Ranger mission. Since its inception at MIT in 1976, the SSL has focused on space operations, emphasizing complete, realistic tasks performed in high-fidelity neutral buoyancy simulation. Development of an extensive data base on EVA structural assembly led to the Experimental Assembly of Structures in EVA (EASE) flight experiment, which flew on STS 61-B in late 1985. The flight hardware was completely designed and constructed in the SSL, built almost entirely by students, with only one full-time staff person.
Since 1983, using neutral buoyancy simulation, the SSL has been performing extensive development and testing operations on space telerobots. The Beam Assembly Teleoperator (BAT) has been used for end-to-end assembly of the same structure flown on STS 61-B, and has further been used for studying Space Station external robotic operations, telerobotic rescue of incapacitated EVA crew, remote servicing of Hubble Space Telescope, and potential cooperative roles for EVA and telerobots in a single integrated work site. These experiments were conducted both in the Neutral Buoyancy Simulator (NBS) at Marshall Space Flight Center and the Neutral Buoyancy Research Facility (NBRF) at the University of Maryland.
BAT, as well as its derivative vehicles, have produced one of the few data bases in existence on advanced telerobotic space operations, and have generated the experience base necessary for the accomplishment of the Ranger mission.
The main focus of the Ranger mission is to demonstrate in the near term the capabilities and limitations of free flying telerobotic servicers for space operations. The mission is designed to shed light on some of the most important issues in space robotics and teleoperation. A task set representative of current and planned space operations has been chosen, and will be attached to the upper stage of the launch vehicle. Ranger's performance on these tasks will demonstrate to potential users of this technology the level of capability that presently exists.
The mission also represents a new way of doing business. Ranger is an extremely low cost, high return flight experiment. This is made possible by accepting some increase in mission risk. To reduce cost, the SSL is streamlining the development process by utilizing selective certification of components where it makes sense. Class S components are being used only in a limited number of critical systems, with MIL-SPEC or ruggedized commercial components being used elsewhere. This simplification is made possible by the short mission duration. The design life of the Ranger vehicle is only 45 days. To achieve an extremely aggressive development schedule, Ranger is being designed by a small localized team. This greatly reduces the delays which naturally occur when multiple contractors are linked primarily by paperwork. This new paradigm in design philosophy has caused NASA Headquarters to view Ranger as a programmatic experiment as well as a robotic experiment. Ranger is currently scheduled for launch in the first quarter of 1997 on a Lockheed LLV2 launch vehicle.
The free flying telerobotic spacecraft, known as the Ranger Telerobotic Flight Experiment (Ranger TFX), is the main component of the mission. This vehicle, although highly experimental for this initial mission, represents a complete operational system capable of final phase rendezvous, docking, and servicing activities.
The payload of Ranger is the four manipulators used for telerobotic operations. These are centralized in the manipulator module, which is a small cubical structure at the front of Ranger to which all four manipulators are mounted. The vehicle configuration was influenced strongly by the previous work conducted in the SSL with BAT and other free flying telerobots.
Although most EVA tasks are designed for one arm performance, two dexterous arms are used by Ranger. Supporting activities such as tool handoff implicitly require a second manipulator. A second manipulator can also be used to stiffen the dock to the work site when large forces are required.
A rigid grapple to a point near the workspace on the target satellite is desirable. This permits local reaction of work forces back into the satellite structure. This has led to the inclusion of a grapple manipulator.
A wide range of camera views is necessary. Simple tilt and pan units have proven inadequate for complicated manipulation tasks. Many researchers have found that wrist mounted cameras are useful for manipulator grasping tasks in which the end-effector is "flown" into position with the assistance of a visual target. SSL experience has shown that for more general tasks a tool-orthogonal view is often preferred. This requirement has led to the inclusion of an independent camera manipulator.
A narrow forward-mounted manipulator base enhances versatility. Restricted work volumes are common on existing space systems. Broad shoulders are not desirable in tight quarters. Also, when using multiple dexterous manipulators, overlap of workspaces is required for maximum effectiveness. Ranger's four manipulators are based on a 1-foot cubic manipulator module at the forward end of the vehicle.
Although most NASA programs have assumed that the telerobot is stationary, or positioned by an RMS-type manipulator, experiments in the SSL have repeatedly demonstrated the utility of a mobile base carrying manipulators for servicing. Free-flight is critical to the success of a small servicing system such as Ranger, since it must achieve rendezvous and grapple of the target satellite to initiate servicing operations. For this reason, free-flight maneuvering is a key element of the Ranger mission. Ranger will include cold gas thrusters and magnetic torquers for attitude control and local maneuvering. Also, although not planned for this experimental mission, Ranger is designed to incorporate an Orbital Maneuvering System (OMS) engine for extensive changes in its orbit.
A task set representative of current and planned space operations has been chosen for the Ranger mission. These tasks will allow the vehicle to demonstrate the full range of capabilities required of a robotic system for servicing a failed satellite.
The first part of the mission task set will be a task board consisting of standard EVA fasteners and interfaces such as J-hooks and electrical connectors. The task board will also contain a set of basic robotic performance metrics currently performed in laboratories, such as peg-in-hole and contour following tasks.
The second subset will address robotic servicing tasks currently planned for the International Space Station and other future missions. In part of this session, Ranger will perform an Orbital Replacement Unit (ORU) change out using a standard H-handle robotic interface. The final task subset involves planned and unplanned EVA servicing. Here, Ranger will demonstrate several tasks that have performed by or have been planned for EVA astronauts. The selected tasks include selected HST-type servicing operations, spacecraft refueling, and several others.
To facilitate development of Ranger TFX, a functionally equivalent Ranger Neutral Buoyancy Vehicle (Ranger NBV) has been developed and is currently operational in the SSL.
The NBV is kinematically identical to the TFX vehicle, and shares the same control and avionics architectures. Extensive NBV operations are currently being conducted in the NBRF allowing validation of control software and electronic systems.
A high fidelity underwater mock-up of the TFX task hardware is currently under construction. During the time leading up to the launch, Ranger NBV will repeatedly perform the tasks planned for the mission. This will allow the mission plan to be repeatedly refined, and problems to be identified before the mission begins. Also, during the flight, Ranger NBV will be used for anomaly resolution by recreating the problem in the NBRF while Ranger TFX is waiting for the next telerobotic test period. This approach is similar to the one used by astronauts for mission training and anomaly resolution.
Ranger NBV will also help facilitate the training of the flight crew. Several operators will be trained for each task, along with pilots for the rendezvous and docking phase of the mission. The neutral buoyancy vehicle will be controlled from the same facility as Ranger TFX. This will allow the crew to gain operational experience with the actual flight controls prior to launch.
During the time leading up to the launch, Ranger NBV will develop an extensive data base on the performance of these tasks. This data, when compared with actual flight results, will help to quantify the differences between the neutral buoyancy and space operations.
After the mission, Ranger NBV will stand in its own right as a powerful research tool for robotic space operations. The vehicle uses high fidelity sensors and actuators, which will give the level of operational precision necessary to further refine the science of neutral buoyancy simulation.
In order to reduce communications time delay, and to simplify system interfaces, a line of sight communications strategy with a single ground station has been chosen. Ranger will use 4 omnidirectional antennae to communicate with a high gain tracking station on earth. This system allows a command uplink of 0.5 MBits/sec, and a digital stereo video and telemetry downlink of 4.0 MBits/sec.
With this scenario, the mission consists of two main operational modes. The telerobotic testing or "high activity" mode occurs when Ranger is in direct line of sight with the ground station. It is during these portions of the mission that task operations will be carried out. The electrical power recharge or "low activity" mode occurs when Ranger is out of sight of the ground station. During low activity periods, the attitude control system will keep the solar arrays pointed at the sun allowing the batteries to recharge in preparation for the next high activity period.
When considering task operations, it is desirable to make the high activity periods as long as possible. Studies in the SSL indicate that high activity periods of less than 15 minutes are not useful for real-time telerobotic operations, and times greater than 25 minutes are best. This causes higher orbits to be desirable. When this fact is traded against launch vehicle costs, an optimum (within the Ranger program fiscal constraints) is achieved with a circular orbit altitude of about 900 nmi. with an inclination of about 45°ree;. With a ground station located at the NASA Wallops Island facility, or at the University of Maryland, this orbit produces about 5-6 useful high activity periods per day, with an average length of about 22 minutes.
After insertion into orbit, Ranger TFX will initiate communications with the control station at the SSL, and will perform basic system self-tests to establish vehicle functionality. Ranger will then deploy its solar arrays, and the attitude control system will establish sun pointing.
Upon establishing control of the Ranger vehicle, the mission will proceed through six phases. The difficulty and risk of the test operations increase with each successive test phase with the greatest risk at the end of the mission.
In the first phase, the manipulators will be unstowed and tested one at a time. Dexterous arm dynamic calibration will be accomplished by monitoring joint torques and accelerations with the arm in a variety of positions. Force-torque sensors in the dexterous and grapple manipulators will then be calibrated. Teleoperation of each arm then will be tested one at a time in single joint, resolved rate, and position control modes. Following this, the camera manipulator will perform a visual survey to determine the condition of the Ranger vehicle.
After completion of manipulator checkout and calibration, the grapple manipulator will reach back to the launch vehicle, and attach to a grapple fixture. This operation will be monitored by the video manipulator. Upon positive confirmation of grapple, Ranger will command the release of the launch restraints, and will then use the grapple manipulator to swing itself around to face the launch vehicle. It is in this configuration that Ranger will perform most of the tasks planned for the mission.
Phase two will consist of tasks performed using the task board mounted on the upper stage of the launch vehicle. This task board will allow Ranger to practice some of the task elements to be used in later parts of the mission such as J-hooks, electrical connectors, and structural connectors. Ranger will also perform basic robotic performance metrics such as peg-in-hole and contour following.
In phase three, Ranger will perform tasks currently planned for future space robots. These tasks have been designed to be "robot friendly", with simple interfaces and indicators. The most important of these is the ORU changeout. Many systems on the International Space Station have been designed to use these robotic interfaces, therefore successful demonstration of these tasks will provide important validation to the space community.
Phase four will encompass tasks that have been performed by or have been planned for EVA astronauts. These tasks have not been designed to be performed by robots, and will present a much greater challenge than the tasks in the previous phase. The Hubble Space Telescope provides a rich, well understood EVA task set from which several tasks have been selected. Another important task planned for this phase is the demonstration of on-orbit refueling. Successful demonstration of this robotic capability will have important implications for spacecraft currently in geosynchronous orbit that are functioning properly, but are running out of attitude control or station-keeping fuel. Another task planned for this phase will involve the deployment of a failed mechanism. Many spacecraft have failed, or have been forced to operate at a limited capacity, due to a simple deployment mechanism failure. In many cases, the mechanism is still functional, but thermal effects or friction have caused it to seize, requiring the application of a small amount of additional force to break it free.
Phase five repeats the tests performed in phase one in order to recalibrate the manipulators and gather data on changes in performance due to environmental and dynamic effects. This information is important when considering the correlation of the previous tasks with those performed with Ranger NBV. This information will also be valuable for future space manipulator designs.
Upon completion of the first five phases of the mission, Ranger will release the target vehicle and will begin the free-flight portion of the mission. To successfully service an ailing spacecraft, Ranger must be able to achieve final phase rendezvous and docking. This phase is designed to demonstrate this capability. Ranger will use its cold gas thrusters to perform a small distancing maneuver, and then will return and grapple to the target vehicle. This will be repeated several times, increasing the distance each time. Finally, Ranger will perform a fly-around and inspection of the upper stage of the launch vehicle, before ending the mission by grappling to the target vehicle.
Ranger TFX is currently in the final stages of design, and long lead items are being procured. Ranger NBV is now operational with two dexterous arms and a passive grapple arm. Ranger NBV has proven to be a valuable tool for both simulating space operations, and providing useful data for Ranger TFX. ORU changeout tasks performed with Ranger NBV have provided useful data return on hardware design, as well as dexterous arm operations. An underwater mockup of the target vehicle and task set is currently under construction, with end-to-end servicing simulation scheduled to begin in Fall 1995. Fabrication and testing of the spacecraft will continue through calendar year 1996 with spacecraft delivery occurring in early 1997.
The ambitious nature of this schedule is made possible by the fact that a major portion of the flight vehicle is identical to the neutral buoyancy vehicle, allowing hardware and software duplication in many areas, particularly in the manipulators. Neutral buoyancy results have thus far validated the utility of the free-flight approach to servicing, and grappling operations are still being tested.
The Ranger mission will provide a large data return at a very low cost, and upon completion will represent a powerful new capability available to the space community.
This work is funded under NASA Grant NAGW-4034; the Technical Monitor is Mr. David Lavery, whose encouragement and support is gratefully appreciated. The author would also like to acknowledge Dr. David Akin, Dr. Craig Carignan, Gardell Gefke, and Rob Cohen who have also made contributions to this paper.