\header ENTRY_TIME = 1997 APR 08 16:49:22 SYSTEM_NAME = SCIENCE AUTHOR = STEVE PETERS, BORIS SEMENOV INSTITUTION = JPL/CALTECH START_TIME = 1996 DEC 18 16:10:00 *STOP_TIME = 1996 DEC 19 00:40:00 *TARGET = NULL *DAY = NULL \text ROCKY-7 LAVIC LAKE TEST FIELD NOTES (third day of the test) DATE: Wednesday, 18 December 1996 PLACE: Lavic Lake AUTHOR: Steve Peters NOTE: times below are from both Steve Saunders' and Steve Peters' watches which differed by approximately 2 minutes TIME NOTE ---------------------- -------------------------------------------------- Dec 18, 1996 8:10 PST arrive. Paul at JPL is trying to test his shuffle-and-place. Walk to starting place, discuss scenario, directions to Diana at JPL to control; Dec 18, 1996 8:35 PST Return. Make crater for Diana to investigate. Scenario is we have landed x meters east (?) of three impact craters, Diana investigates then go y meters theta to a rayed crater (1st three exist, last was dug) dig will be done at prepared crater; Dec 18, 1996 9:10 PST rover placed at start, pointing north (arm south); Dec 18, 1996 9:15 PST booting rover; Dec 18, 1996 9:21 PST ray crater; Dec 18, 1996 9:24 PST Volpe patch complete, initialize wheels, arm done; Dec 18, 1996 9:25 PST begin panorama pan counterclockwise from south monochrome; Dec 18, 1996 9:37 PST mast stow commences after panorama; Dec 18, 1996 9:38 PST complete; Dec 18, 1996 9:40 PST cannot ping helios... not sure if satellite problem or network problem; Dec 18, 1996 9:42 PST helios ping successful, panorama in WITS, satellite down; Dec 18, 1996 9:50 PST chesapeake taken off line to reduce traffic over satellite link; Dec 18, 1996 9:58 PST hot swap batteries; Dec 18, 1996 10:04 PST restarting WITS server; Dec 18, 1996 10:06 PST sending rover to west 30 m (operator error); Dec 18, 1996 10:10 PST rock jammed in left steering wheel; Dec 18, 1996 10:12 PST rock removed with screwdriver total drag roughly 4 m; Dec 18, 1996 10:17 PST same wheel caught basalt; Dec 18, 1996 10:18 PST finished traverse .45 rad tilt < bearing was ~246 magnetic; Dec 18, 1996 10:22 PST taking panorama; Dec 18, 1996 10:25 PST not getting right image reseated connector; Dec 18, 1996 10:33 PST stowing mast; Dec 18, 1996 10:34 PST Jeff & Ray walked to site and selected rock to investigate ("astronaut-assisted rover operations"); Dec 18, 1996 10:36 PST Diana can see 1st panorama at JPL, now processing latest panorama; Dec 18, 1996 10:41 PST hot swap batteries, rock has been designated in panorama; Dec 18, 1996 10:45 PST send command to go to rock (T9); Dec 18, 1996 10:46 PST started turn will go to pointy rock and take navigation image; Dec 18, 1996 10:49 PST motion stops (~5.5 m travel); Dec 18, 1996 10:56 PST issue relative turn 90deg, take another nav image; Dec 18, 1996 10:58 PST think see rock in image measurement of move: 29.1 m (between panoramas) straight line (theory error is affected by curves, dragged wheel); Dec 18, 1996 11:03 PST calib wheels to obtain sun sensor measurement, measured drag was 4.05 m; Dec 18, 1996 11:08 PST start move toward rock jerky motion; Dec 18, 1996 11:09 PST stopped 2" from rock accidentally sent command to rover; Dec 18, 1996 11:12 PST motion ends; Dec 18, 1996 11:14 PST begin move back to previous location at rock T9; Dec 18, 1996 11:16 PST cinder caught in same wheel; Dec 18, 1996 11:18 PST stopped extremely close to rock point of rock would interfere with mast deployment; Dec 18, 1996 11:20 PST looking at new image 20 cm distance from rover origin to rock; Dec 18, 1996 11:29 PST moved rover back 30 cm, worked fine; Dec 18, 1996 11:32 PST power downPST Dec 18, 1996 12:15 PST walked to site, took "family picture", decided immediate plan: take overhead science image with mast camera of rock, move vehicle, take nav image for placement; Dec 18, 1996 12:34 PST start rover power; Dec 18, 1996 12:40 PST calib starts; Dec 18, 1996 12:46 PST start mast move to observe rock T9 from above; Dec 18, 1996 12:52 PST decide to move ("rotate") rover; Dec 18, 1996 12:54 PST stowing mast make multiple low-level sequence started sequence, only single navig moves work in SCE 2nd command sequence had wrong sign for delta x, so movement in wrong direction; Dec 18, 1996 1:25 PST changing batteries; Dec 18, 1996 1:30 PST turned 135 deg; Dec 18, 1996 1:31 PST move -1.42 m; Dec 18, 1996 1:32 PST turn to face, take navig image; Dec 18, 1996 1:40 PST start close-up imager sequence touched stereo-pair box corner; Dec 18, 1996 1:47 PST complete; Dec 18, 1996 1:47 PST 310 deg magnetic; Dec 18, 1996 1:49 PST start panorama 2 pm AMES people arrive; Dec 18, 1996 2:02 PST panorama complete plan is: back up ~1 m then go 20 m toward Ray crater, take another panorama, then straight to Ray; Dec 18, 1996 2:14 PST sequence stepped to allow measurement of rover position between move; Dec 18, 1996 2:15 PST first move was in wrong direction see that forgot to zero out panorama origin physically pick up vehicle and place ito position of panorama; Dec 18, 1996 2:23 PST chamging batteries; Dec 18, 1996 2:27 PST calibrated wheels (to updated orientation from sun sensor); Dec 18, 1996 2:30 PST completee measurement of origin, T9, step sequence; Dec 18, 1996 2:33 PST move complete, measure position; Dec 18, 1996 2:35 PST 7 m move starts (last of this sequence); Dec 18, 1996 2:38 PST completed move; Dec 18, 1996 2:39 PST start next move (previous was not last!); Dec 18, 1996 2:42 PST sequence complete, measuring position plan is to take panorama; Dec 18, 1996 2:43 PST start panorama; Dec 18, 1996 2:55 PST finishing panorama (was interrupted by some failure) single stepping; Dec 18, 1996 2:58 PST stowing mast Greg processing mast image into mini panorama; Dec 18, 1996 3:03 PST calib wheels, start move caught cinder -- dragged same wheel; Dec 18, 1996 3:14 PST battery hot-swap rock was stuck, having passes the new bumper and got caught on strut; Dec 18, 1996 3:15 PST start move (17 m, previous was 20 m); Dec 18, 1996 3:18 PST move complete; Dec 18, 1996 3:46 PST completed very long move when command 30 m (measured 31.9); Dec 18, 1996 3:52 PST have started next traverse ~20 m; Dec 18, 1996 3:57 PST done; Dec 18, 1996 4:04 PST 20 m traverse starts; Dec 18, 1996 4:08 PST sun half set behind mountain; Dec 18, 1996 4:12 PST stopped plan to send 50 m; Dec 18, 1996 4:13 PST started 30 m move; Dec 18, 1996 4:23 PST cinder stuck in same wheel, manually removed; Dec 18, 1996 4:24 PST fender blocked (other wheel); Dec 18, 1996 4:25 PST caught in other wheel, manually removed, move completed; Dec 18, 1996 4:27 PST started on next traverse 30 m command; Dec 18, 1996 4:35 PST cinder caught on "other" wheel; Dec 18, 1996 4:40 PST done 205 m measured from last panorama; \header ENTRY_TIME = 1997 APR 08 16:51:08 SYSTEM_NAME = SCIENCE AUTHOR = STEVE PETERS, BORIS SEMENOV INSTITUTION = JPL/CALTECH START_TIME = 1996 DEC 19 16:58:00 *STOP_TIME = 1996 DEC 19 19:44:00 *TARGET = NULL *DAY = NULL \text ROCKY-7 LAVIC LAKE TEST FIELD NOTES (fourth day of the test) DATE: Thursday, 19 December 1996 PLACE: Lavic Lake AUTHOR: Steve Peters NOTE: times below are from both Steve Saunders' and Steve Peters' watches which differed by approximately 2 minutes TIME NOTE ---------------------- -------------------------------------------------- Dec 19, 1996 8:58 PST panorama has been taken ~4 m in front of "Ray" crater. Paul from JPL is on the phone; Dec 19, 1996 9:05 PST Paul ready to send command, Pete adding patch to disable collision detection; Dec 19, 1996 9:07 PST conference call -- on speaker phone to class; Dec 19, 1996 9:16 PST set of tasks has been selected by class: 2 moves [untelligible], dig, dump, downlink image; Dec 19, 1996 9:25 PST rover rebooted while doing image check during dig; Dec 19, 1996 9:40 PST rover has been rebooted, wheels & arm calibrated after being repositioned at panorama origin. disable-vision patch entered; students have prepared move, move, image scenario; Dec 19, 1996 9:45 PST cell phone outage reconnected; Dec 19, 1996 9:50 PST sequence complete, Greg moving image so it can be viewed over web; Dec 19, 1996 9:52 PST viewing close up image (nav image); Dec 19, 1996 9:57 PST remote student operation complete; Dec 19, 1996 10:30 PST panorama far from trailer already complete, batteries swapped, Diana to command rover from JPL; Dec 19, 1996 10:31 PST rover initialized to origin of new panorama; Dec 19, 1996 10:35 PST adding Lavery's IP address to the .htaccess; Dec 19, 1996 10:45 PST measured rock of interest to be at x=-4.5, y-11.5 m; Dec 19, 1996 11:08 PST calib wheels, begin 2 moves then science image; Dec 19, 1996 11:11 PST stopped; Dec 19, 1996 11:18 PST change batteries; Dec 19, 1996 11:31 PST begin mast deploy to place close up imager on ground in fron of rover therover placed; Dec 19, 1996 11:44 PST mast stowed; \header ENTRY_TIME = 1997 APR 10 23:54:38 SYSTEM_NAME = SCIENCE AUTHOR = UNKNOWN, PROVIDED BY CURT S. NIEBUR INSTITUTION = UNKNOWN START_TIME = 1996 DEC 15 12:00:00 *STOP_TIME = 1996 DEC 15 22:00:00 *TARGET = INSTRUMENT_1 *DAY = NULL \text ROCKY-7 LAVIC LAKE TEST FIELD NOTES ("zero" day of the test) DATE: Sunday, 15 December 1996 PLACE: Lavic Lake AUTHOR: UNKNOWN, provided by Curt S. Niebur Drove to Lavic Lake site from JPL. High winds forced the use of an alternate route, so the site was not reached until after dark. The trailer and RV were stationed at the site for the night. \header ENTRY_TIME = 1997 APR 10 23:55:12 SYSTEM_NAME = SCIENCE AUTHOR = UNKNOWN, PROVIDED BY CURT S. NIEBUR INSTITUTION = UNKNOWN START_TIME = 1996 DEC 16 08:00:00 *STOP_TIME = 1996 DEC 16 22:00:00 *TARGET = INSTRUMENT_1 *DAY = NULL \text ROCKY-7 LAVIC LAKE TEST FIELD NOTES (first day of the test) DATE: Monday, 16 December 1996 PLACE: Lavic Lake AUTHOR: UNKNOWN, provided by Curt S. Niebur Weather note: Sunny, clear, cold; with a gentle yet steady and bitter wind. The satellite dish was set up and the satellite found; however, communication with JPL could not be established. It was speculated that high winds may have knocked the rooftop dish on Bldg. 72 out of alignment. The first test was of the sunsensor, using the correct parameters (latitude, etc.) for the site. The rover was initially aligned with North, then manually turned in 90 degree increments. The readings seemed reasonable (360, 120, 210, and 296). Initial mast tests revealed problems with the right camera. The nav cameras returned consistently good images. The next test involved sending the rover on a 10 m traverse over a dense cinder field, to ensure the perception system would not interpret the cinders as obstacles. The rover traversed a straight line for 8.3 m, then stopped after turning (as if toward goal). Before it could be determined whether the rover had finished, the battery died (it had been on for 50 minutes). The trailer was moved to Tuesday's test scenario configuration. New speculation on the still-nonexistent connection to JPL was that when the trailer was disconnected from the Ethernet, the line was left unterminated in the MarsYard (this, I believe, turned out to be the problem). Despite not being able to send rover commands from JPL, it was decided to run through tomorrow's test scenario. The rover was moved 200 m away from the trailer, near the rock cairn marking 180m from a specified small crater. On the first mast image, both cameras worked, but filter wheel selection had been forgotten. It was noted that the mast was swaying significantly in the wind. In the second mast image, for reference, the rock cairn was picked out. It was measued to be 3 m away from the rover, and was located halfway vertically in the image. At 11:50am, Paul Backes was successful in pinging assateague from JPL, showing that satellite connection had been established. Since the mast cameras were not calibrated, the images could not be processed to be available in WITS for Paul to use at JPL. After taking mast imagery, the rover was sent due South 10 m. Judging from the way the rover turned, South was off from the desired direction (toward the crater) by 10-20 degrees to the right. However, the rover did not drive after turning, and contact was lost. Contact was regained after power cycling the rover, but battery status was still indeterminate. (It had been 33 minutes since the last battery swap.) At this point, the battery for the cordless phone died--since there was no spare, communication between the trailer and those attending the rover was difficult. The rover batteries were swapped--speculation was that the batteries died so quickly due to the cold weather. A human relay was set up for communication. Instead of restarting the test scenario, after the rover was rebooted, Rocky7 was sent on a 10 m traverse, due South (~30 deg to the right off of the line to the crater). After ~5 m, the rover rebooted itself--possibly because a cinder became jammed in the wheel strut. Post-lunch, the rover rebooted itself several times during traverses. Speculation was that the batteries may have been damaged by the freezing weather overnight. That possible cause was eliminated when the rover rebooted itself while tethered. A full panorama was taken with the rover tethered, with no error; the rover seemed to reboot solely during traverses. Further experimentation to pinpoint the error--such as holding wheels immobile, executing a long traverse with the rover up on blocks, etc.--failed to cause the reboot error to recur. The scientists arrived at lunchtime, and spent the afternoon exploring alternate sites at the lakebed. Other logistics notes: The hotel misplaced the sandwich order; groceries to replace the missing lunch were bought on the way to the site in the morning. This was possibly a cheaper solution for lunches on site. The spouts for the gas cans had been forgotten at JPL, so new gas cans with spouts were bought. Sand screws to secure the satellite dish did not work; the screws (actually screws meant as dog ties) broke in the hard ground. The portable toilets worked out well; to my knowledge the RV tanks did not have to be dumped during the field tests despite having a group of up to 16 attending the tests. \header ENTRY_TIME = 1997 APR 10 23:55:47 SYSTEM_NAME = SCIENCE AUTHOR = UNKNOWN, PROVIDED BY CURT S. NIEBUR INSTITUTION = UNKNOWN START_TIME = 1996 DEC 17 08:00:00 *STOP_TIME = 1996 DEC 17 22:00:00 *TARGET = INSTRUMENT_1 *DAY = NULL \text ROCKY-7 LAVIC LAKE TEST FIELD NOTES (second day of the test) DATE: Tuesday, 17 December 1996 PLACE: Lavic Lake AUTHOR: UNKNOWN, provided by Curt S. Niebur Weather note: Still cold, the sky was overcast, which may have had some effect on the sunsensor. To alleviate the on-site communication problem, walkie-talkies were obtained. As a side note, a rear window of one of the Suburbans shattered when the door jammed. The spare battery for the video camera did not work; fortunately, the battery was the same as those used for the cel phones so another spare could be found. The test scenario was run. After taking a full panorama, Greg manually processed the mast imagery so it could be used as the wedge view in WITS for the scientists. Rocky7 was sent on a traverse, but it rebooted after several steps. The batteries were swapped (after running 38 minutes; they were measured at 13.8 V). The rover was physically moved back to its starting position and resent on the traverse. The rover rebooted again after ~ 8 m. The mast was used to try to image the rover tracks. The first image was blank; the filter selection had been forgotten again. The image was retaken with filter setting 1; the right image was bad (as usual, half dark and half light). Images were taken with all filter settings, and the right camera continued to return faulty images. Speculation was that the camera itself was failing. The final stereo pair produced a good right image. Since the tracks were difficult to make out on the hard ground, a second wedge at a different angle was attempted. At first, the mast was sent accidentally to -pi radians (waist joint); the mast position was reset to 0 before taking the image (no filter; right image was good). Another single, no filter image was taken at an inclination of 60 deg off of vertical; again the right image was good. At this point the battery timer indicated time to swap, so the mast was stowed. The rover tracks were still found to be too difficult to pick out in the mast images. Rather than restarting the test scenario, the rover was rebooted in place, and a half-panorama (forward) taken. Scientists designated a rock target in WITS. The rover was sent toward the rock, but unfortunately the wrong coordinate frame was used. The traverse was aborted and the rover was physically moved to its previous location. The rover lurched briefly forward and backward, then executed the traverse. Despite avoiding a nonexistent obstacle (ground reflection?), the general direction of the traverse was directly toward the rock. The rover rebooted itself before the end of the traverse. Since it had been 35 minutes since the last battery swap, the batteries were changed again. The rover was moved over to the specified rock, in preparation for placing the close-up imager against it. The scientists used nav camera imagery to designate a target on the rock, and then the close-up image sequence was sent, which executed successfully. The scientists requested an overhead image of the rock, using the mast, which was also successful. Next, the mast was moved to image a second rock, farther away. The mast swung unexpectedly back and forth around the waist joint before moving to position. This first attempt proved too steep (i.e. the rock was not in the field of view), but the second attempt worked. The mast was stowed after capturing the image, and the batteries were hot-swapped (after running 35 minutes) for the first time. The procedure worked. The rover was moved physically to the second rock, and the previous sequence repeated: grabbed nav image, designated a target, deployed mast, took a close-up image, then stowed the mast. This time, a target on the front face was chosen to test mast placement ability, although this choice risked hitting the stereo camera box on the concave face of the rock. The mast approached slowly, and ended up ~2mm short of the rock face. Everything was powered down then to refuel the trailer generator. After swapping the rover batteries and restarting the trailer equipment and the rover, the test scenario continued with a dig operation at the rim of the crater. Four digs were attempted: the first three failed because of the hard ground (the third dig was with partially prepared soil). The soil was carefully groomed for the fourth attempt with fine soil was brought from a nearby site; the operation succeeded. Next, Paul Backes at JPL sent a dump command and took a science image with the nav camera. The image included a coin and a rock hammer for scale comparison. Paul then sent a dig and a dump command, which were both successful, and which finished the JPL remote part of the day's activities. The mast was deployed for rover self-inspection using two views, one each of the rover front and right sides, and then restowed. The batteries were hot-swapped again. Rich Volpe lowered the parameter for dig depth by 2.5 cm, to see whether the dig operation would be more successful on unprepared soil. The rover was physically moved to a nearby site, and the dig operation attempted. However, the ground proved too hard, and the operation lifted the rover. It seemed that Rocky7 recovered gracefully, restowing its arm, but there was no subsequent response from the rover. There was no evidence of a reboot. Rocky7 was carried closer to try to re-establish contact, but there continued to be no response, even with a tether to the serial port. The rover rebooted fine after being power cycled, and contact was re-established. At this point, the planned testing for the day was finished, so control of Rocky7 was given to Robert Ivlev in order to try to solve the rover's auto-reboot problem. As a side note, designating a waypoint via WITS from mast imagery was done for the first time today. Robert Ivlev visually inspected Rocky7's circuitry, looking for loose connections, etc., and found none. He measured the supply voltages; all were nominal. Rich Volpe implemented in software a bypass of the vision request during navigation, to establish whether the autoreboot behaviour was caused by the perception system. (The stimulus is "fsm manual navig BYPASS_OBSDETECT = active.) The rover was sent due South 50 m to test Rich Volpe's mod. The rover rebooted, but it was discovered that it was still running the vision system. The code had accidentally been changed in the wrong place: "newvision" rather than in "lavic". The error was corrected, and the rover was resent on its 50 m clear traverse. This time, the rover completed the long traverse successfully. It travelled on a straight line, very near (if not right on) due South, 49.8 m. The batteries were hot-swapped (after running 53 minutes), and the rover was sent back due North. A 50 m traverse had been intended, but the rover was inadvertently sent to (50,0,0) rather than back to the origin. The rover tracks were not as straight this time, perhaps because the sun was setting, causing errors in heading computed by the sunsensor code. The rover completed the 100 m traverse, without obstacle detection, in 39 minutes. According to its telemetry, its final position was (49.954,3.637). Its actual position was not measured accurately, but was paced out as approximately 100 m from where it began the traverse. The rover was next commanded back to the origin, but it responded by turning in circles, behaviour consistent with losing the sun behind the horizon. It did, however, resume a straight traverse when a flashlight was substituted for the missing sun. In parallel, SACS was used to take a full panorama at 25 deg inclination, as well as a few wedges at a shallower angle (20 deg) to look at a crater approximately 10 m away from the camera. The same crater was imaged again after being seeded with black rocks to test whether increased texture would improve range map accuracy. SACS was then moved 5 m along the direct line toward the crater, and images were taken with and without the rocks. The crater could easily be picked out in the elevation map from 5 m but not from 10 m; the rocks seemed to have little or no effect on the results. *The field test continued for at least one more day. I was not in attendance, however, so my notes end here. \header ENTRY_TIME = 1997 APR 10 23:56:45 SYSTEM_NAME = SCIENCE AUTHOR = UNKNOWN, PROVIDED BY CURT S. NIEBUR INSTITUTION = UNKNOWN START_TIME = 1996 DEC 15 12:00:00 *STOP_TIME = 1996 DEC 19 22:00:00 *TARGET = INSTRUMENT_1 *DAY = NULL \text DATE: Tuesday, 17 December 1996 PLACE: Lavic Lake AUTHOR: UNKNOWN, provided by Curt S. Niebur NOTES FOR NEXT TRIP: Be sure to read these notes (and those from the Nov. desert trip) before leaving on the next field trip! Be certain, just before leaving, that Ethernet is terminated in the MarsYard! Bring: walkie-talkies spare battery for cordless phone additional cel phone for trailer AAA batteries gas cans with spouts cotton swabs & alcohol ...and everything on November's list