TUBA CITY
Marsokhod Field Test
The purpose of the Tuba City experiment was to perform a simulation of a rover mission to Mars.
A key feature of this experiment was that the site was chosen as an example of the type of terrain that might be found on Mars which would be relevant to determining the past history of the climate, and the finding evidence of past life. Science team choose the locations to send the rover to achieve science goals, and interpreted the observations from the rover as they were received. From the performance of the science teams, we are better able to evaluate the effectiveness of rover observations for achieving an accurate scientific understanding of the Martian surface. Using these results, we are further able to recommend additional capabilities would substantially enhance science performance of rovers on Mars.
The engineering objectives of the mission were to simulate, at the highest possible fidelity, the operation of a rover on Mars. Some of the engineering hardware on the rover was not representative of flight systems, but the mission operational environment was closely analogous to a system that could be used in flight. Systems which were not flight relevant did not effect rover operations. Some of the onboard software used to operate the rover (such as visual servocontrol) would be highly useful for flight systems. These systems were tested and performed well under conditions that simulated a flight mission.
The outreach objective of the experiment was to involve and inform the rural Native American community who live in the area surrounding the field site in the Marsokhod rover experiment, and in the Mars exploration program in general. Using their computers and the Internet, students controlled the rover as it navigated through a sparsely-vegetated area during the field test
Field Experiments with the Marsokhod Rover:
Daniel Christian, David Wettergreen, Maria Bualat, Kurt Schwehr, Deanne Tucker, Eric Zbinden
The site is located on the Navajo Indian reservation in the painted desert in Northern Arizona. The site was found by a team who visited a number of sites on the Navajo reservation. ( The site survey team consisted of Ron Greeley, Carol Stoker, Jeff Moore, and Dan Christian). The survey team was looking for a site which had evidence of sediments deposited by liquid water. We further sought a site in which fossil evidence of past life could be found. Field experiment logistics provided further constraints in that the site had to be accessible by road. The site had to be close to an area in which supporting field facilities could be set up. We needed to be able to maintain communications link with the rover during its traverse via line of sight communications within a range of 1 km. We needed to be able to deploy and retrieve the vehicle at the sight easily so that it could be housed and serviced overnight. Furthermore, it was important to have a remote site in which public access to the field site could be controlled.
The chosen site met all of these requirements. The site had very good exposures of terrain which indicated a varied climate history and had an unusally good collection of features showing evidence of aqueous deposition containing a fossil life record. Furthermore, it provided good access for supporting logistics, but was in an area rarely visited at the time of year of the field test. A topographic map of the area shows some of the site's features.
To provide the science team with adequate information about the site, we chose one location in the area to be used as a simulated landing site. A series of aerial photographs were taken centered on the landing site to form a set of simulated descent images. In addition to that, standard imaging products at resolution of 10 cm per pixel were obtained via the USGS mapping service. Finally, images from the NASA U2 were obtained that provided a simulated satellite image. Note that in these aerial photos, the white strip shows the simulated landing site. The length of the strip is 3 meters.
Since the field site was on land belonging to the Navajo Indian Nation, we were required to obtain their official permission to use the site. Obtaining this permission was a two step process. First, the local tribal chapter house, with juristiction over the land in question, passed a resolution by public vote allowing us to use the land. Obtaining this permission involved getting a resolution drafted by the officers of the tribal chapter and submitting it for public comment and vote at a chapter meeting. To gain support from the local community, it was valuable to show them how supporting this resolution would be of benefit to their community. The key benefit to the community was believed to be educational activities associated with the mission.
In order to provide educational activities in association with the mission, we formed a collaboration with the Northern Arizona University Native American Institute. Lawrence Gishey of that organization, helped draft the resolution supporting our use of the land that was presented to the tribal chapter. This resolution was passed by public vote of the local tribal chapter.
The second step involved obtaining a use permit from the Navajo Nation. We submitted a use permit application to the Navajo Nation, which was subjected to review by various offices within that organizaion including the agricultural council, the fish and wildlife commission, the historic preservation office, etc. These offices judged that the experiment would not have deleterious impact on their land. Most of these offices reviewed our written documentation and determined without further study that our activity would not have a negative impact. The historic preservation office recommended that we have a archeological survey performed at the site to verify that we would not disturb cultural or historic artifacts by using the site. After this survey was performed, and the report from it concluded that no cultural artifacts would be disturbed in the area, the Navajo Nation granted the permit. The entire process to obtain permission took about 6 months.
The payload selected for the rover was designed to simulate as closely as time and budget permitted the characteristics of a Mars rover mission. We took Mars rover missions currently being planned as guidance for mission design. Missions to be simulated included the Mars Pathfinder mission, scheduled to land on Mars in July 1997, and two other missions under considerations and planning for a possible 2001 launch. These latter two missions were the joint US-Russian Mars Together mission, and a Mars rover mission concept under development for a proposal to NASA's Discovery program. The payload was chosed to reflect (but not exactly duplicate) key features of each of those missions.
The mast camera on the rover was designed to simulate the capabilities of the Imager for Mars Pathfinder (IMP) camera. This camera design was chosen because we wanted to simulate key features of the Pathfinder mission by using a panorama from a camera of this nature to target rover observations, we wanted to evaluate the use of multispectral image cubes in association with time-limited rover observations and finally, multispectral imaging is an important capability for mineral identification on Mars.
The rover arm and multifunction end effector was designed to duplicate the capabilities of an arm/end-effector planned for Mars Together. However, a final design of the arm for that mission was not available to us, and we opted to produce our own design that could achieve multiple co-located instrument placements in a single arm placement. The end effector we designed was capable of placing up to three instruments, and had a fourth position for a sample scoop so that analyses could be performed and samples collected from the same location.
The field experiment was supported by equipment at the field site which included a motor home housing a small computer laboratory including an SGI Indy workstation and a Sun workstation. These computers allowed commands to be sent to the rover and were networked with other workstations at Ames via a remote RLAN network. A portable satellite dish was set up at the site, operating at a 112 Kbps bandwidth. The satellite dish was supplied by Lymon Brothers. When not in operation, Marsokhod was housed in the covered bed of a moving van. Power to the field camp was supplied by one 5 kw electrical generator. A 600 w portable generator was used for the rover.
The daily operations schedule was as follows. Each day of the 6 day science mission, the Marsokhod was under the control of the Ames/California team from (9AM - 4PM PST. Two hours before Marsohod operations started, the field team went out to the field camp from the nearby town where they were lodged brought the camp to full operational status, and deployed Marsokhod at a predetermined location. The location for deployment varied from day to day but it was either at the simulated landing site or at the point where operations stopped on the previous day.
At the end of the operational period, Marsokhod was packed up and put into the covered van space for overnight storage. Necessary maintanance was performed, and the field camp was closed for the night.
The field team consisted of three engineers (D. Christian, E. Zbinden, C. Mina) and a ground truth field geologist [three people participated in this: David Nelson, Jim Rice, and Michael Kraft]. For the week of setup preceeding remote operations, the project leader (C. Stoker) accompanied the field team to coordinate logistics and interact with the local community leaders as necessary. For the two days of the Pathfinder mission simulation, H. Eisen, a Mars Pathfinder engineer joined the field team. In addition, each day 6-10 high school students from the local area joined the field team as field assistants. These students observed rover operations, took notes, and interacted with any members of the local community who came by as spectators. Only the field team were allowed in the area of rover operations during the remote control periods. We orignally planned for the students to help with crowd control, but due to the remotelness of the area, we had very few visitors and crowd control was not needed.
During the six days of rover operations, there were three different groups of scientists with different operational approaches. The three teams and their operations are described below:
The teams were organized so that there was a clear chain of command and clear methods for making decisions and communicating them with the engineering team. This organizational structure was refined during the mission in response to feedback from the teams.
Team assignments were as follows:
The science team was only allowed to communicate to the operations team via the science team leader to the operations officer. This kept science team out of the operations area. This was rather difficult because our operations (Science and all) was confined to a single room.
In addition, the science team was divided into two groups, once responsible for "trafficability issues" and one responsible for "science analysis". After the first two days, this division was abandoned because it was found to serve no useful function. Instead, the science team all kept trafficability issues in mind in traverse planning.
Pathfinder Team Members:
Carol Stoker
Bob Reid
Andy Mishkin
George McGill
Hank Moore
Natalie Cabrol
Edmund Grin
Ron Greeley
This team consisted of scientists associated with the Pathfinder mission. In addition, Andy Mishkin, a member of the Pathfinder JPL mission operations team , participated in mission operations. The science team was not provided with descent imaging.
The goal of the mission operations was to simulate significant aspects of the Pathfinder mission, including:
5.2.2 Mission Operations:
The mission started at 8AM with a prebriefing of the team given by Ron Greeley. Ron acted as science team leader for the Pathfinder simulation (Days 1 and 2). The first data provided were a complete panorama of the terrain in the clear filter (from the stereo camera). The tilt angle of the panorama was not optimized to allow viewing of the near field, so an additional panorama was taken to image the near field. From these panoramas, the team was asked to select a single location to send the rover for closeup observations. At the end of daily operations, a science team debrief was held. An attempt was made to solve problems identified by the science team overnight.
In the afternoon of day 2, the science team was allowed to see the
simulated
descent images to orient themselves to the site. All the team felt
that mission planning would have been significantly enhanced if these images
had been available at mission start.
Click on image for full size.
| Science Images | Navigation Images | Data Volume | Traverse | Science Sites |
|---|---|---|---|---|
| 78 | 183 | 41,727,146 Bytes | 33 Meters | 4 |
Team 1 Command Cycles | Nov 4 | Not available | ||
| Nov 5 | 39 | |||
Mars Together Team Members:
George McGill
Steve Saunders
Jack Farmer
Natalie Cabrol
Edmund Grin
Virginia Gulick
Carol Stoker
Rags Landheim
5.3.1 Mission Goals:
This mission simulation was designed to be more relevant to the Mars Together Mission. The key goals of this mission were:
5.3.2 Mission Operations:
A few of the scientists involved in the Pathfinder simulation stayed on for the Mars Together Mission, and thereby had already achieved a greater level of training and understanding of the site than the new people coming in. Training was as big factor in the performance of a team.
The mission simulation started at 8:00 AM with a prebriefing and orientation.
The team was provided with the descent imaging products and documents
describing the rover, the data formats, and an introduction to software
products available. The team was also able to use image panoramas, including
a color panorama, generated during the Pathfinder simulation to plan a
traverse to achieve science goals. The rover was situated at the simulated
landing site at the start of the mission, and the team was asked to define
a list of science goals, define science targets to achieve these goals,
and direct the traverse to these targets.
On the second day of Mars Together operations, Sasha Eremenko operated the rover for the entire day. He was trained by watching the previous day's rover operator for two hours, and then was given about an hour of assistance by a trained operator. We found that the rover operations system could be learned with one day of training (or less).
Click on image for full size.
| Science Images | Navigation Images | Data Volume | Traverse | Science Sites |
|---|---|---|---|---|
| 49 | 258 | 19,320,513 Bytes | 132.8 Meters | 6 |
Team 2 Command Cycles | Nov 6 | 71 | ||
| Nov 7 | 53 | |||
Team 3 Members:
Steve Squyres
Mike Carr
Dave DesMarais
Carol Stoker (Support)
Jeff Moore (Support)
Natalie Cabrol
Edmund Grin
5.4.1 Mission Goals:
The team was provided with descent imaging in advance of the mission. From these they deduced the following: "The landing site appears to consist of sedimentary country rock with multiple bedding units. Some units have been tilted and eroded to a peneplain. These units are overlain non-conformably by other bedded units. The region has been volcanically active with intrusive dikes and pipes as well as flat lying units that may be either intrusive sills or extrusive flows. Tectonically, the region is cut by a single throughgoing, linear fault of unknown displacement. Fluvial and aeolian erosion and deposition have been recently active, producing relict stream beds and dust covered surfaces."
The goal of the mission, as stated by the team leader (Squyres) was to test this hypothesized general framework within 500 m of the landing site by examining suspected examples of each of the processes posited. At each site, data will be gathered to characterize each of the local rock units by spectra, texture, any bedding margins, and degree and type of weathering. Dust and aeolian and fluvial deposits will be characterized by spectra, grain size distribution, and flow direction indicators. Fractures and faults will be characterized by strike and dip and displacement. The objective is to characterize as many different units as possible and to allow for correlation between units both from rover data and orbital images.
5.4.2 Operational Objectives:
Team 3 consisted of three members of the Mars Science Working Group. This team (unlike the other two) had been sent the simulated descent imaging information in advance and, from this, they had designed a traverse before they started their operational period. Key operational objectives of team 3 involved using the arm for observations and performing long traverses and then trying to determine their location using dead reckoning by comparing images at the end of the traverse with the descent images. Because the team was involved in preparing a proprietary proposal, they elected not to disclose details of their operation and so no detailed operational summary is given here. In their traverse, they obtained data at 5 stations and traversed a total of 106 m distance.
Click on image for full size.
| Science Images | Navigation Images | Data Volume | Traverse | Science Sites |
|---|---|---|---|---|
| 58 | 193 | 21,173,423 Bytes | 106 Meters | 6 |
Team 3 Command Cycles | Nov 8 | 35 | ||
| Nov 9 | 63 | |||
The team was debriefed by Stoker, who asked them to summarize what they
believed the stratigraphic record implied about the geologic history. The
three agreed that the data implied the area was sandstone formed from wind-deposited
material. They had not seen any evidence of aqueously-deposited sediments.
After confirming that they were sure of this interpretation, this team
was shown photographic evidence of aqueous deposited sediments they had
missed, including the dinosaur tracks.
All science teams were provided with workstations with which they could view and process images using Photoshop® and xv. Hard copy images were also printed out in color and in black and white. The preference for hard-copy vs workstation views of images varied with the teams. Team 1 preferred working with hard copy and was unhappy when hard copy access was slow. Team 2 used some hard copy but used workstations extensively. Team 3 used workstations exclusively, and did not use hard copy at all.
Color images and color panoramas were prepared by IMG staff and were made available by posting an enlarged panorama on the wall within 24 hours of receipt of the image. Stereo panoramas could be viewed in stereo on a workstation by two methods: 1) a red-blue stereo which used red-blue glasses, and 2) a stereo projection using Stereographics stereo glasses. The latter allowed full color panoramas to be viewed in stereo. There was also software -- a program called ilStereoView by SGI -- that allowed a user to pan around in the stereo panorama while viewing it in stereo.This latter capability was very popular with the science team.
A large monitor was set up which allowed the science team to view the rover operators panel, which showed the vehicle's position and state in a terrain model, and showed the latest image aquired. This helped the science team keep track of what was going on with rover operations and what was the latest data aquired. In addition, an image log was kept by the and posted by the operations officer where the science team could clearly see it.
The following is a summary of mission operations for the mission simulations. To get the detailed operation commands sent to the rover, and detailed information about the image files, refer to the mission operators' logs.
Mission Operations started at the simulated landing site which is shown as a 3 m long strip pointed North on the gridded photograph.
The first objective was to take a complete panorama which included a mockup of the Pathfinder windsock. Panoramic Images of this site are:
Next, a set of windsock images were taken.in each of the camera filters. Some examples follow. To get the complete windsock took two image frames (top and bottom).
Next, from analysis of the panoramic images, the science team selected a target to move the rover to. The rover was moved using the visual tracking, and an image was taken at the destination at full resolution. The target (TGT 1 on rover path) was only a short distance from the initial starting point.
Rover path from Nov. 4-5 shows the location of TGT 1.
Finally, images of the target area, as the rover was driving around at the spot were taken with the front pallet camera. Images taken driving around in this area are:
To prepare for Nov. 5 operations, a color panorama (previously taken and stored) was printed out for the science team to analyze. They were told to decide on 3 more science target stops based on using this color panorama.
The first operation was to obtain a stereo pair image of Target 2 for context.
The rover was then moved to target 2. The target outcrop was then imaged in stereo at full resolution, and in red and blue filters at half resolution.
Target 3 is a lag deposit on the surface. The objective at Target 3 was to analyze it before and after disturbing it with the rover wheels. The rover was moved to target 3, the area was imaged in RGB filters, then the rover executed a 360 degree turn in place (churning the area up) and another RGB sequence of images was taken.
Images at target 3 are:
The next objective was Target 4. Target 4 is a ledge forming unit which appears to have layering. The objective was to obtain close up images of the outcrop, and to do an arm placement and scoop some fines at the base of the outcrop.
First a clear filter context image of target 4 was taken from the location of target 3. Also, a blue and red half frame image were taken at this location. Then the rover traversed to the target. Once at the target, a RGB filter set of images (red, green, blue) were taken. Then a front pallet image was taken in an attempt to see under the ledge, but it was heavily in shadow. Finally, the rover was moved to sediment at the edge of the outcrop and the arm was placed in the sediment, which was imaged using the arm camera. The scoop was used to pick up some sediment, and the same area was imaged with the arm camera after scooping.
End of Pathfinder mission operations.
The rover was started at the landing site. The first objective was to find and traverse to target 1. After locating the target, images were taken of the target in blue (camera 13) and IR (camera 19) filters. The IR filter is out of focus.
Additional stereo images were taken around the target area, including:
The rover was then moved to Target 2, rock arcs west of landing site. This material was imaged in stereo.
Images were then taken of target 3 for geologic context from this perspective.
Traverse was completed to Target 3 and it was imaged in mono and stereo.
The rover moved on towards target 4, but did not get there by the end of daily operations.
The rover resumed traverse to target 4. Target 4 is blocky material on middle of K9. Cameras obtained this high resolution image for navigation purposes at S.W. corner of K7 on gridded descent image on bearing of 190deg.
It then continued the traverse to target 4. Upon reaching target 4, it took a high resolution stereo image of blocky material on the ground at -14 ° tilt.
It then took a context stereo image of Target 5 from Target 4 at 1 ° tilt and 138 heading.
A bumper camera image of target 4 was taken. The science team wanted to see these before making a decision about where to do an arm placement at this site. This image has not been located.
Next the arm was placed on a surface selected by the science team, and 4 arm camera images were taken: a clear filter (scale bar is 5mm), a red color, a green color, and a blue color image.
The rover then traversed to Target 5. Target 5a is the SE corner of L10 on the 10 m descent image. The objective is to get a full resolution stereo of slope to NE of gap at target 5. It was expected that several images would be needed to get the entire slope.
The traversed continued on to target 5B2, which is a light-colored rock in the gap in the SW 1/4 of M10 on the 10 m descent image. The front of the rover was brought to the light rock for an arm placement. Two stereo images were taken of the slope due E of rover. Two were needed to get the entire slope. Images are:
The arm was then placed on the light-colored rock at 5B and arm camera images were taken in clear, red, green, and blue channels. We then attempted to scoop from the surface of this rock, but had no effect (red filter image) on the hard surface.
End of Mars Together Mission
| TOTALS (Data Volume in Bytes) | NOV 4 | NOV 5 | NOV 6 | NOV 7 | NOV 8 | NOV 9 | ||
| Number of Images | 109 | 139 | 131 | 170 | 65 | 182 | ||
| Image Data Volume | 6,759,555 | 16,471,614 | 6,686,130 | 12,452,071 | 6,722,173 | 12,054,482 | ||
| Panoramas | 7 | 6 | 0 | 1 | 1 | 3 | ||
| Panorama Data Volume | 16,568,807 | 1,165,750 | 0 | 182,312 | 2,112,154 | 1,843,183 | ||
| Cumulative Data Volume | 23,328,362 | 18,398,784 | 6,686,130 | 12,634,383 | 8,834,327 | 13,897,665 | ||
| Total Data Volume: | 83,779,651 Bytes | |||||||
| TOTALS (Data Volume in Bytes) | NOV 4 | NOV 5 | NOV 6 | NOV 7 | NOV 8 | NOV 9 | ||
| Science Images | 25 | 53 | 14 | 30 | 18 | 40 | ||
| Science Data Volume | 20,934,187 | 14,257,448 | 4,057,174 | 9,466,335 | 7,275,701 | 10,867,490 | ||
| Navigational Images | 91 | 92 | 117 | 141 | 48 | 145 | ||
| Navigational Data Volume | 2,394,175 | 4,141,336 | 2,628,956 | 3,168,048 | 1,558,626 | 3,030,175 | ||
| Science Images | 180 | ||||||
|---|---|---|---|---|---|---|---|
| Science Data Volume | 66,858,335 Bytes | ||||||
| Navigational Images | 634 | ||||||
| Navigational Data Volume | 16,921,316 Bytes |
Carol Stoker
Last modified: Sun March 22, 1998