FIDO prototype Mars rover field trials, Black ... - Wiley Online Library

11 downloads 0 Views 1MB Size Report
as test of the ability of robotic mobility systems to conduct field science. R. E. Arvidson,1 ... the test site commanded the rover remotely from the Jet. Propulsion ...
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. E11, 8003, 10.1029/2000JE001464, 2002

FIDO prototype Mars rover field trials, Black Rock Summit, Nevada, as test of the ability of robotic mobility systems to conduct field science R. E. Arvidson,1 S. W. Squyres,2 E. T. Baumgartner,3 P. S. Schenker,3 C. S. Niebur,1 K. W. Larsen,1 F. P. Seelos IV,1 N. O. Snider,1 and B. L. Jolliff 1 Received 5 February 2001; revised 20 December 2001; accepted 19 April 2002; published 30 August 2002.

[1] The Field Integration Design and Operations (FIDO) prototype Mars rover was deployed and operated remotely for 2 weeks in May 2000 in the Black Rock Summit area of Nevada. The blind science operation trials were designed to evaluate the extent to which FIDO-class rovers can be used to conduct traverse science and collect samples. FIDO-based instruments included stereo cameras for navigation and imaging, an infrared point spectrometer, a color microscopic imager for characterization of rocks and soils, and a rock drill for core acquisition. Body-mounted ‘‘belly’’ cameras aided drill deployment, and front and rear hazard cameras enabled terrain hazard avoidance. Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) data, a high spatial resolution IKONOS orbital image, and a suite of descent images were used to provide regional- and local-scale terrain and rock type information, from which hypotheses were developed for testing during operations. The rover visited three sites, traversed 30 m, and acquired 1.3 gigabytes of data. The relatively small traverse distance resulted from a geologically rich site in which materials identified on a regional scale from remotesensing data could be identified on a local scale using rover-based data. Results demonstrate the synergy of mapping terrain from orbit and during descent using imaging and spectroscopy, followed by a rover mission to test inferences and to make discoveries INDEX TERMS: 5455 that can be accomplished only with surface mobility systems. Planetology: Solid Surface Planets: Origin and evolution; 5464 Planetology: Solid Surface Planets: Remote sensing; 5470 Planetology: Solid Surface Planets: Surface materials and properties; 5494 Planetology: Solid Surface Planets: Instruments and techniques; 6225 Planetology: Solar System Objects: Mars; KEYWORDS: Mars, rovers, field trials, FIDO, MER, remote sensing

1. Introduction [2] This paper uses data and results from ‘‘blind’’ field trials conducted at the Black Rock Summit, Nevada, in May 2000 using the Field Integration Design and Operations (FIDO) prototype Mars rover (Figure 1). Specifically, the paper focuses on testing a paradigm for robotic, rover-based exploration that has as its core the synergistic use of orbital, descent, and rover-based observations to develop, test, and update hypotheses about the nature of surface materials and the geologic history of landing sites and surrounding areas. During the May 2000 tests a science team unfamiliar with the test site commanded the rover remotely from the Jet Propulsion Laboratory (JPL) in Pasadena, California, simulating 27 Martian days, or ‘‘sols,’’ of rover activities. The trials were designed to simulate the types of complex remote operations expected with such missions as the 1 Department of Earth and Planetary Sciences, Washington University, St. Louis, Missouri, USA. 2 Department of Astronomy, Cornell University, Ithaca, New York, USA. 3 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California, USA.

Copyright 2002 by the American Geophysical Union. 0148-0227/02/2000JE001464

FIDO

original 2003/2005 Mars Sample Return Missions, the 2003 Mars Exploration Rover Mission, and the Mobile Science Laboratory expected to be on the surface toward the latter part of this decade. The trials included operations associated with traverse science and sample collection done in ways that characterize the site and sample representative materials. [3] FIDO was built at JPL and equipped with instrumentation designed to simulate selected elements of the Athena Payload [Squyres et al., 1998; Haldemann et al., 2002] (Figure 1). Prior to the Nevada experiments, FIDO was deployed numerous times in the JPL Mars Yard and twice during initial engineering trials located at Silver Lake, California [Arvidson et al., 2000a], to understand how to operate the vehicle efficiently in geologically realistic terrains. Thus, by the time the rover was deployed at the Black Rock Summit site the focus was on science operations designed to test the ability of rovers to conduct traverse science and collect samples. [4] This paper builds on the rover-based science and operations delineated during the Pathfinder mission [Golombek, 1997], field tests at the Pisgah Volcanic Field, California, using the Rocky 7 rover [Arvidson et al., 1998], and results from the FIDO [Arvidson et al., 2000a] and Marsokhod rover deployments at Silver Lake, California

2-1

FIDO

2-2

ARVIDSON ET AL.: FIDO PROTOTYPE MARS ROVER FIELD TRIALS

2. Athena and FIDO Payloads

Figure 1. FIDO deployed in JPL Mars Yard. Mast stands 1.94 m above surface. Arm is deployed onto red rock and Color Microscopic Imager data are being acquired.

[Stoker et al., 2001]. The Black Rock Summit trials differed from previous work in that FIDO, its payload, and the rover command and control system were similar to those to be used during the 2003 Mars Exploration Rover Missions and what was expected to be used during the original 2003/2005 Mars Sample Return Missions. The autonomous operation capabilities associated with FIDO are also prototypical for what is expected during the Mobile Science Laboratory Mission, including onboard hazard avoidance during traverses and deployment of instrument arms and associated instruments. [5] This contribution is a companion to papers describing FIDO and its payload [Haldemann et al., 2002], the procedures for fusion of orbital, descent, and rover-based data to maximize extraction of scientific information about the field test sites [Moersch et al., 2002], and a detailed analysis of the geologic setting of the Black Rock Summit site by the remote science team, based only on analyses of orbital, descent, and rover-based data [Jolliff et al., 2002]. Additional companion papers deal with an engineering experiment in which FIDO and its sister rover, K9, were used to jointly explore a portion of the Black Rock Summit area [Stoker et al., 2002], joint use of rover-based reflectance and laser-induced breakdown spectroscopy (LIBS) for remote mineralogical and elemental characterization of targets [Wiens et al., 2002], and experimental studies focused on joint photogrammetric analyses of descent and rover-based images to pinpoint or localize rover locations in specific coordinate systems [Li et al., 2002].

[6] The international Mars exploration program will include measurements from the Martian surface as well as eventual return of samples. The goal of this program is to achieve a broad understanding of the nature and history of the planet, including evolution of the climate and life and the role of water. Surface mobility using rovers will be a major part of these future surface science and sample return missions. For example, the 2003 Mars Exploration Rover Mission will consist of two rovers equipped with the Athena Payload [Squyres et al., 2001] and designed to operate at two sites for at least 90 sols. The original Athena Payload as it was selected for the Mars Sample Return Mission [Squyres et al., 1998] was designed to focus on two major aspects of surface operations: (1) conducting long-distance traverses in which remote-sensing and in situ data are collected and analyzed to determine the geologic history of the survey sites and (2) identifying, approaching, sampling, characterizing, and caching rocks and soils that are representative of the materials exposed at the site. In designing the Athena Payload, particular emphasis was placed on including instruments with measurement capabilities that allow identification of aqueous minerals and thus provide evidence of ancient climatic conditions and possible prebiotic or biotic activity. [7] The original Athena Payload selected for Mars Sample Return is summarized in Table 1. For remote sensing it included a mast-mounted high-resolution stereo panoramic multispectral imaging system (Pancam) and MiniTES, an emission spectrometer operating in the 5 – 29 mm spectral window. MiniTES is a point spectrometer that gathers thermal data as individual spectra or as rasters for key targets identified using Pancam data. The payload also included four instruments to be placed on the end of a deployable arm for in situ analyses of rocks and soils. These were Alpha Particle X-Ray, Mo¨ssbauer, and Raman Spectrometers, and a Microscopic Imager. Use of all four instruments together would have provided detailed elemental, mineralogical, and textural characterization of rock and soil targets. The final capability included in the original payload was the MiniCorer, which could core rocks and collect soil samples. The in situ instruments would also have been used to characterize the ends of cores and to verify core presence before caching in sample containers. [8] The Mars Sample Return Mission was canceled in February 2000. Subsequently, a revised version of the Athena Payload was chosen in July 2000 for the 2003 Mars Exploration Rover (MER) Mission. The Athena Payload on MER includes Pancam, MiniTES, the Mo¨ssbauer Spectrometer, the APXS, and the Microscopic Imager. It also includes a Rock Abrasion Tool (RAT) to remove rock coatings before measurements are accomplished. The MER Mission is designed to conduct mast-based remote-sensing and in situ analyses over a distance of up to 1 km during the nominal operational period of 90 sols per rover, with a possible 90-sol extended mission (Table 3). Finally, the Mobile Science Laboratory expected to be operating on the surface toward the end of the decade will feature an 10 km traverse capability, together with remote-sensing instrumentation to characterize sites and find targets, drills to access subsurface materials, and an extensive array of in situ

ARVIDSON ET AL.: FIDO PROTOTYPE MARS ROVER FIELD TRIALS Table 1. Summary of Full Athena Payloada Instrument Pancam

MiniTES

Key Parameters Mast-Mounted 12 bands (0.4 – 1.0 mm) for stereoscopic imaging with 0.3 mrad IFOV; 9.2 by 18.4 FOV emission spectra (5 – 29 mm, 10 cm 1 resolution) with 8 or 20 mrad FOV

Purpose detailed imaging of surface for geologic and topographic characterization mineralogical mapping of key targets identified using imaging data; thermophysical properties of surfaces

Arm-Mounted In-Situ Package Cm alpha particle elemental abundances sources, solid-state for rock and soil targets alpha and X-ray detectors, 4 cm FOV 57 Fe spectrometer in identification of ironbackscatter mode; bearing minerals and Co/Rh source and iron oxidation states Si-PIN diode detectors; 2 field of view 1.5 cm Raman 532 nm laser, 20 mm Mineral detection and Spectrometer spot size, 532-620 nm abundance mapping for spectral coverage rock and soil targets Microscopic 30 mm/pixel close-up imaging of Imager monochromatic imager texture and mineralogy (1024  1024) with 6 of surfaces mm depth of field Alpha Particle X-Ray Spectrometer (APXS) Mo¨ssbauer Spectrometer

MiniCorer

a

244

Rover-Mounted drill with 0.8 cm bit capable of coring to 2.5 cm depth

Obtain rock cores and soil samples. Allow core ends to be examined with in situ package. Cache samples.

Note: Engineering elements such as Navcam and Hazcam not shown.

instrumentation to determine soil and rock chemistry and mineralogy. [9] The FIDO science payload was selected to be exemplary of the types of instruments and capabilities associated with the Athena Payload, although cost and time constraints dictated simpler systems on FIDO as compared to the Athena Payload (Figure 1, Table 4) [Haldemann et al., 2002]. Engineering instruments on board FIDO include front and rear hazard avoidance cameras to monitor the terrain in the immediate vicinity for hazards during traversing (Figure 1). The onboard hazard avoidance software is designed to analyze the hazard avoidance image data and have the rover go over relatively small obstacles or circumnavigate those that are judged by the algorithm to be too hazardous to traverse over. ‘‘Belly’’ cameras are mounted under the front of FIDO to help support and document instrument deployments. A mast is included that houses three instruments that comprise the remote-sensing component of FIDO. Pancam is a false-color infrared stereo imaging system capable of surveying the terrain at high spatial resolution for scientific purposes. The Navcam is used for planning traverses and has low spatial resolution, wide field of view, and broad stereo baseline as compared to Pancam. An Infrared Point Spectrometer (IPS) is boresighted with Navcam and acquires spectral radiance data over wavelengths from 1.3 to 2.5 mm. Although operating in

FIDO

2-3

the reflected rather than the emitted portion of the spectrum, IPS is similar to MiniTES in that it allows testing of concepts related to the joint use of high spatial resolution imaging and low spatial resolution hyperspectral data. An arm is included with a Color Microscopic Imager (CMI) and a Mo¨ssbauer Spectrometer [Klingelho¨fer et al., 1996]. [10] FIDO was also equipped for the May 2000 trials with a prototype MiniCorer that could acquire rock cores 0.5 cm in diameter by up to 1.7 cm long [Haldemann et al., 2002]. The cores could be extracted and examined with the CMI to ensure core presence and integrity. Once presence of a core was confirmed, it could either be ejected or kept and placed in a caching tube. The MiniCorer used on FIDO during the May 2000 tests was a breadboard sampling tool not designed for rugged field use. For example, it did not have a central guide rod to keep the bit onto a target. It also had very preliminary deployment mechanisms designed for proof of concept rather than rugged use. Nevertheless, it was deemed important to demonstrate drill deployments onto rock surfaces, even if acquisition of cores was difficult. Therefore test objectives related to the MiniCorer for May 2000 focused on successful drill deployment and not core acquisition. [11] Finally, for the Black Rock Summit field trials it was impossible to use the Mo¨ssbauer Spectrometer due to logistical difficulties associated with obtaining permission for on-site use of the radioactive source associated with the instrument. The reader is referred to Mo¨ssbauer Spectrometer results from the Rocky 7 field trials in 1997 and the work conducted in Silver Lake with FIDO in 1999 [Arvidson et al., 1998, 2000a].

3. Precursor Engineering Trials [12] Several engineering trials were conducted at the JPL Mars Yard during 1999 and early 2000 to develop and test scenarios for efficient uses of the rover during traversing and sampling. For the most part these tests were focused on objectives related to meeting requirements associated with the 2003/2005 MSR Mission that was active at the time (Table 2). In addition, many of the objectives are consistent with the 2003 MER Mission requirements, with the exception of sampling (Table 3) and with capabilities expected for the Mobile Science Laboratory. The Mars Yard is a 20 m by 20 m area covered by basaltic boulders and cobbles (diameters ranging from 1 m to several centimeters) strewn on a coarse-grained sand deposit. The boulders and cobbles cover 15% of the surface, consistent with what is observed on Mars [Golombek, 1997]. [13] The Mars Yard trials included a combined science and engineering team (Core Operations Team, or COT) who commanded the rover remotely, using only data from the rover for analyses and planning. Tests started with acquisition of Pancam and Navcam panoramas to identify rock targets for in situ analyses and drilling, along with defining the waypoints for the rover to use to reach the targets. After analyses of the panoramas, uplink sequences were generated and transmitted to the rover using the onboard wireless Ethernet system. FIDO then initiated a traverse, using onboard hazard avoidance techniques to get safely from waypoint to waypoint. Results showed that traverses could be planned for at least 20 m and that the best approach was

FIDO

2-4

ARVIDSON ET AL.: FIDO PROTOTYPE MARS ROVER FIELD TRIALS

Table 2. Mars Sample Return Mission 2003 and 2005 Level 1 Requirements and FIDO Test Objectives Requirements Conduct rover-based remotesensing and in situ observations to characterize site, including mineralogy and chemistry. Collect at least 45 rock cores/fragments and up to 10 soil samples that characterize site diversity and deliver to Mars Ascent Vehicle. Traverse at least 1 km integrated over 90 primary sample acquisition mission and postMAV launch 90 sol exploration and discovery mission.

Use Pancam, Navcam, IPS to characterize landing site and define traverses to drill targets. Approach to within 1 m of targets using Navcam data for waypoint selection. Characterize targets using Pancam, IPS.

to come within 1 – 2 m of a rock target and then to command the rover to turn in place to face the rock target. Acquisition of Navcam, Hazcam, and Bellycam data at the end of the traverses became the standard practice in order to ascertain rover position and the presence of local hazards and to plan the fine-scale motions needed to move over the rock target for arm and drill deployments. It was found that typically two to three command cycles were needed to get the rover from its starting position to a position over the rock target and ready for arm or drill deployment. Both the arm and the drill were deployed numerous times, including successful acquisition of CMI data and drill cores. [14] The Mars Yard trials served to prepare the COT for an extended period of engineering tests at the Silver Lake, California, site in 1999 [Arvidson et al., 2000a]. The Silver

Table 3. Mars 2003 Exploration Rover Level 1 Requirements and FIDO Test Objectives Conduct rover-based remotesensing and in situ observations to characterize site, including morphology, topography, mineralogy, and chemistry Conduct in situ observations of selected rock and soil targets. Traverse up to 1 km integrated over 90 sol primary mission.

Instrument

FIDO Test Objectives

Use Pancam, Navcam, Hazcam, Bellycam to define fine-scale maneuvering over targets and acquire CMI and MBS data. Drill rock targets, verify core presence, cache, and deliver to Mars Ascent Vehicle. Determine ability of rover to accomplish objectives within given data volume and command cycle windows.

Requirements

Table 4. Summary of FIDO Payload

FIDO Test Objectives Use Pancam, Navcam, IPS to characterize landing site and define traverses. Approach to within 1 m of targets using Navcam data for waypoint selection. Characterize targets using Pancam, IPS. Use Pancam, Navcam, Hazcam, Bellycam to define fine-scale maneuvering over targets and acquire CMI and MBS data. Determine ability of rover to accomplish objectives within given data volume and command cycle windows.

Mast

Pancam

Navcam

Key Parameters Mast-Mounted 4 degree-of-freedom deployable mast located on rear deck of rover; places mast instrument package up to 1.94 m above surface false color IR (0.65, 0.75, 0.85 mm), stereo (12 cm baseline) imaging system, 0.35 mrad/pixel IFOV, 10 FOV panchromatic stereo (12 cm baseline) imaging system, 1.5 mrad/pixel IFOV, 40 FOV

Infrared Point 1.3 – 2.5 um spectrometer Spectrometer with 13 cm-1 resolution, 9 mrad/pixel FOV, bore(IPS) sighted with Navcam

Field Use deploy Pancam, Navcam, IPS remote-sensing package for acquisition of data for scene characterization and sequence planning detailed imaging of surface for geologic and topographic characterization characterization of morphology and topography of surface for use in sequence planning mineralogical mapping of key targets identified using imaging data

Arm-Mounted In Situ Package 4 degree-of-freedom arm deploy in situ analysis with Color Microscopic package (CMI, MBS) Imager and Mo¨ssbauer onto surface Spectrometer Color 30 um pixels, 1.5 cm FOV close-up imaging of Microscopic texture and Imager mineralogy of surface 57 Fe spectrometer in identification of ironMo¨ssbauer bearing minerals and Spectrometer backscatter mode; Co/Rh source and Si-PIN diode iron oxidation states detectors; field of view 1.5 cm2 Arm

MiniCorer Bellycam

Hazcam

Rover-Mounted drill with 0.5 cm bit capable of coring to 2 cm depth 110  90 FOV, 3 mrad/pixel IFOV, 12 cm baseline; mounted on rover underbelly same imaging system as Bellycam, mounted on rover body front and back

obtain rock cores image surface for MiniCorer deployments characterization of morphology and topography of surface for use in sequence planning

Lake site allowed the COT to test rover operations on alluvial fans, along beach deposits, and within an incised channel system with walls that exposed weathered diorite covered with caliche-cemented alluvium. These tests validated the ability of the COT to command the rover to conduct successfully long distance traverses (60 m per command cycle) while acquiring remote-sensing data, in addition to approaching rock targets, making in situ measurements, and deploying the MiniCorer and acquiring cores. These tests were conducted from a command trailer located at the site but removed visually from the rover operations areas. Thus the tests were not fully blind and allowed the COT to approach the site after FIDO completed its operations to see what was accomplished and what was not executed. In addition, the Mo¨ssbauer Spectrometer on the arm was deployed, and data were acquired during the Silver Lake tests [Arvidson et al., 2000a]. Lessons learned were

ARVIDSON ET AL.: FIDO PROTOTYPE MARS ROVER FIELD TRIALS

FIDO

2-5

Figure 2. Black Rock Summit FIDO test site geologic map based on Quinivilan et al. [1974]. Carbonate and rhyolitic tuff units are exposed as a horst block in the Basin and Range Province. Basalt flows and ash deposits in the vicinity represent the eastern edge of the Lunar Crater Volcanic Field. FIDO trials were conducted on a hill to the south of a long basalt flow. The site was chosen to be able to view the 100-m cliff or wall formed by differential erosion between the flow and other rocks and to be able to sample basaltic cobbles, tuffs, and carbonates representative of the regional geology by traversing at most tens of meters. North is to the top. Trials were conducted at the Black Rock Summit area so the basalt flow flowed downhill to the east. fed back into revisions of operational scenarios, improvements for the command and control system (Web Interface for Telescience, or WITS), and improvements in FIDO systems. With the completion of the Silver Lake engineering trials the stage was set for blind trials at the Black Rock Summit site, using operational constraints expected for Mars rover missions.

4. Black Rock Summit Field Trial Constraints [15] Rover operations on Mars will be fundamentally constrained by the short turnaround time for science and engineering teams to process and analyze data, plan the next uplink, generate sequences, and transmit the uplink to the vehicle. Simulation of the management and decision-making processes for the May 2000 field trials was based on what is expected for MSR and MER. Specifically, the Core Operations Team (COT) at JPL consisted of a Science Operations Working Group (SOWG) that examined quicklook data and defined activities for the rover to accomplish and a Rover Operations Team (ROT) that generated and transmitted commands to accomplish the tasks. The COT was sequestered at JPL and had no prior knowledge of the

site other than information derived from analysis of orbital and descent data, that is, Airborne Visible and Infrared Imaging Spectrometer (AVIRIS), IKONOS, and helicopter data described in a subsequent part of this paper. [16] Communication between a command trailer located at the site and the rover was by wireless Ethernet (sometimes using repeaters). Communication between the command trailer and JPL was via communications satellite connections. The FIDO tests were compressed in that no attempt was made to simulate wall-clock time that will be associated with actual Mars operations. For example, first the SOWG typically received and processed data, conducted quick-look analyses, and defined the next set of desired activities. The SOWG and ROT then worked together to generate the next set of sequences. Next, the ROT worked with a representative of the SOWG to generate a command upload that was consistent with SOWG desires and limitations associated with realistic rover operations based on an engineering assessment of traverse hazards inferred from descent and rover-based images, time for operations, and allowable downlink data volumes. All of this work was accomplished within an hour or two of receipt of data. Thus the trials represent a situation even

FIDO

2-6

ARVIDSON ET AL.: FIDO PROTOTYPE MARS ROVER FIELD TRIALS

more time-constrained than expected for actual operations. To ensure efficient rover operations, a Field Team was also formed to provide on-site care of the rover and associated systems, together with documentation as to what the rover accomplished. [17] Mars rover missions will also be constrained by available power for operations, data volumes per sol, and the number and length of uplink command opportunities per sol (e.g., once per sol for MER). The 2003/2005 MSR and 2003 MER rover designs call for up to 100 m of traverse capability per sol. This activity is one of the most power consumptive uses of the rovers. FIDO in all of its tests stayed within this limit by restricting traverse distances to