2.2.2 Lunar Rovers

This segment of the program supports the development of robotics to satisfy the planned requirements for exploration of the surfaces of the Moon. These plans call for robotic reconnaissance and surveying systems preceding the eventual human missions to these bodies. During such missions robots will explore potential landing sites and areas of scientific interest, place science instruments, gather samples for analysis and possible return to Earth, gather and transmit video imagry, and provide images required to generate "virtual environments" of the Lunar surface. The robotic systems required for these operations will require high levels of local autonomy, including the ability to perform local navigation, regulate on-board resources, and schedule activities, all with limited ground command intervention. The objectives of the tasks within this segment of the program are to develop these abilities, as well as conduct research into mobility systems, miniature mechanisms, planning, and on-board navigation. Specific applications are to programs planned by the Space Science and Exploration user communities, as well as other commercial lunar resource utilization opportunities.

Rover Technologies - Antarctic Meteorite Robots

Lunar Rover Demonstration - Desert Trek

Technology Transfer Roadmap Details

Rover Technology - Antarctic Meteorite Robots

Objectives

The purpose of the Meteorite Search task is to develop and demonstrate innovative mobile robot navigation technology that enables vehicles to systematically search for surface meteorites in planetary analog sites such as Antarctic blue ice fields. The general technical objective of the Meteorite Search task is to enable rovers to operate reliably, over long durations in rugged, natural, unstructured environments. The approach falls under the lunar paradigm, in the sense that power and communications bandwidth are orders of magnitude greater than for near-term Mars missions, and in the sense that low temperatures are a technology driver.

The specific technical objective of the Meteorite Search task is to demonstrate and validate multi-rover mobile search systems capable of reliable field operations in the remote, hostile, and low-temperature conditions of Antarctica.

Detailed technical objectives are:

Demonstrate robot operation over 10^6 cycles per outing

Demonstrate sensing and perception to find surface meteorites, and to record meteorite position within 1 m

Demonstrate planning algorithms for efficient search for meteorites, and for coordinating motion of multiple robots in formation

Demonstrate a scientist interface enabling remote geologists to oversee search operations and annotate data

An important non-technical objective of the work is to commercialize the demonstrated technology, enabling a contracted search service.

Approach

1. Mobility

The task will evolve an existing mobility system for Antarctic operations. Candidates include the Nomad, and snowmobile derivatives. Two of the key drivers are rocky icy terrain and low temperatures.

The task will harden rover components, including electronics and wheels. Further, it will design and build power and thermal control systems, will extend the safeguarded teleoperation navigation system to handle additional sensors, and will develop communication links for command, telemetry, and video.

2. Sensing and Perception

The task will develop sensors and detection algorithms for finding surface meteorites. Candidate sensors include vision, magnetometers, radar, eddy current, infrared cameras, and portable Mossbauer backscatter spectrometers. In addition, it will develop sensors and detection algorithms for finding the boundary between blue ice and white snow. This boundary defines the search area.

The task will develop position estimation techniques for tagging find coordinates. These will use sensors such as differential GPS, skyline tracking, and IMU integration, and fusion of their results.

The task will develop techniques for combining images acquired from multiple viewpoints into a seamless mosaic.

3. Planning

Work in planning will develop algorithms for planning the motions of multiple robots searching a given area in formation. Research will pay particular attention to the issue of error recovery, for example, regaining the formation after a disruption.

4. Interfaces

The task will develop a one-person operator interface capable of operating a cadre of robots. This interface will be portable both in terms of easily movable hardware, and in terms of multi-platform software using a language such as Java.

Research will develop a multi-person interface that allows scientists to view telemetry and video from the search party. The interface will include tools for data logging and data annotation.

Focus and Directions:

FY97

Dec 96: Convene workshop to formulate external participation and directions for task.

May 97: Design review

Jul 97: Demonstrate search planners covering 400 sq meters during Atacama field trial

Jul 97: Demonstrate image data mosaicing covering 400 sq meters during Atacama field trial (Level 1)

Jul 97: NSF approval

FY98

Oct 97: Mission readiness review

Dec 97: Sensor data acquisition in Antarctica using at least 2 imaging and 2 non-imaging sensors, with at least 10 meteorites and at least 1,000 meteorite-like but indigineous specimens (Level 1)

Jan 98: Vehicle component testing in Antarctica. Components to be selected from mobility, power, thermal control, and communications. Independent variables to include temperature, wind velocity, terrain type, terrain slope, and snow density.

FY99

Dec 98: Single robot system controlled by single operator searches 10 sq km Antarctic blue ice field, with interaction by tens of remote scientists (Level 1)

FY00

Dec 99: Multi-robot team controlled by single operator searches 50 sq km Antarctic blue ice field, with interaction by hundreds of remote scientists (Level 1)

Mar 00: Multi-robot search for 1 month of austral winter

Point of Contact:
Eric Krotkov
Carnegie Mellon University
Pittsburgh, PA 15201
(412) 268-3058
epk@cs.cmu.edu

Lunar Rover Demonstration - Desert Trek

The purpose of the Lunar Rover Demonstration - Atacama Desert Trek is to demonstrate enabling capabilities for high-performance planetary exploration by mobile robots. The project, part of the CMU Lunar Rover Initiative, will demonstrate and validate these capabilities in terrestrial analogs to telerobotic planetary exploration, and provide critical technologies to NASA for near-term planetary missions and to the private sector for planetary enterprise. The Central focus of this task is to successfully deliver its level one milestone: completion of a 200 km field trial traverse across Chile's Atacama desert, demonstrating robust locomtion, navigation, visualization, and communication capabilities.

Objectives:

The technical objective of the Lunar Rover Demonstration task is to demonstrate a comprehensive mobile robot mission capability for "hands-off" rover operation. The primary objective is long-duration operation, with a target of two months of testing. Assuming 60 days at 8 hours per day, and an average travel speed of 0.15 meters per second, the traverse will cover up to 250 km. The rover will operate in lunar analog terrain, which implies a desert setting (including crater and boulder features in a barren, arid environment). The specific setting will be selected based on cost and availability.

The general technical objective of the Atacama Desert Trek task is to conduct a long-duration, long-distance, untended field demonstration of Nomad, a self-sufficient mobile robot, in planetary analogous desert terrain.

The specific technical objective of the Atacama Desert Trek task is to enable Nomad to operate autonomously as well as respond to human supervision from a US-based control center, with time delays simulating those encountered in a planetary mission. The Atacama Desert Trek will be a public event. Nomad's safeguarded teleoperation and panospheric visualization will combine with a virtual environment interface to allow novice operators such as researchers, scientists, and partner organizations (science centers, schools, and commercial sponsors) to operate Nomad from remote control center(s) and follow its progress over the Internet.

The Atacama Desert Trek will also demonstrate technologies with immediate relevance to the Antarctic Robotic Meteorite Search project, described above.

Approach

During the field trial, the technical approach will be to confirm or refute the following hypotheses:

Four-wheel drive/four-wheel steered locomotion will enable a 200km traverse on planetary analog terrain. Success will be measured by the distance travelled and the percentage of days without mechanical mobility failure.

Autonomous driving capability enables longer traverses than direct teleoperation in time-delay planetary excursion scenarios.

Data from dead reckoning sensors, absolute position sensors, and potentially skyline imagery, can be effectively fused to improve robot position and attitude estimation in a planetary setting.

Active pointing of a directional antenna, even with occasional line of sight loss, provides higher bandwidth communications from a mobile robot, measured in overall number of bytes transferred.

Ultra-wide field of view (panospheric) cameras improve teleoperation and impart a realistic remote experience from a mobile robot, by delivering more imaging control to the remote driver.

Operator interfaces that provide a mix of robot-built terrain models and robot-deployed imagery are appropriate for planetary driving.

Milestones

1994 Demonstrated natural terrain traverse using an existing rover and key components including stereo, a user interface suitable for novices, and control via a low-bandwith command stream.

1995 Designed and built a novel rover, and demonstrated a 10 km outdoor traverse with it using lunarworthy perception.

1996 Demonstrated a 43 km traverse featuring onboard computing in diverse and planetary analog terrain.

FY97, Q1 Demonstrate component technologies: mobility, navigation, visualization, inertial measurement and communications. Hold CDR for Atacama Desert Trek.

FY97, Q2 Integrate components into single system - controllers/interfaces for all subsystems implemented on the real time computer: antenna pointing, serial communications, laser range sensor, GPS, IMU, odometry, compass, advanced steering/deployment system. Small-scale testing of combined systems. Demonstration of subsystem interaction (e.g., position estimation with antenna pointing).

FY97, Q3 Pre-trial experiment; full-up test (all components but satellite link) and reviews. Live imagery from panospheric sensor integrated with user interface. Mobility characterization of fully integrated vehicle.

FY97, Q4 Desert Trek field trial (described above) - Nomad will be teleoperated by novice and expert users located in the United States. A 5-6 person away team will deploy Nomad in the Atacama Desert. The team will remain on site to ensure successful deployment, set up communications relay stations as needed, and operate Nomad remotely when programmatic requirements or satellite downtime requires it. The 200 km traverse, component failure mode tests, and Meteorite technologies will be demonstrated.

FY97, Q4 Documentation of lessons learned, including publication of program products and presentations of its findings.

Point of Contact:
Red Whittaker
Carnegie Mellon University
Pittsburgh, PA 15201
(412) 268-6559
red@ri.cmu.edu