Opportunity

NASA/JPL-Caltech/Cornell Univ./Arizona State Univ. Public domain (NASA / US government work).

Overview

Opportunity landed in Meridiani Planum on 25 January 2004 with a planned 90-sol mission [1]. It drove 45.16 km over 14.375 years of operation, which is the record for off-Earth wheeled vehicle odometry [10]. Its twin Spirit landed in the opposite hemisphere and operated until 2010; through March 2005 the two had driven 3158 m and 4161 m respectively [2]. The two vehicles are identical in design.

Opportunity landed inside a small crater whose walls exposed sulfate-rich sedimentary rock containing hematite concretions. Later work at the rim of Endeavour crater identified clay minerals formed in water of near-neutral pH.

Specifications

Parameter	Value	Source
Mass	185 kg	[1]
Wheelbase length / width	1.4 m / 1.2 m
Width across solar panel	1.8 m
Height with mast deployed	just over 1.5 m
Ground clearance	at least 0.3 m
Wheels	six driven, four corner wheels steered, 0.25 m diameter, rocker-bogie	[3]
Solar arrays	triple-junction GaInP/GaAs/Ge	[4]
Array output at start of mission	approximately 140 W

Mission profile

Parameter	Value	Source
Landing	Meridiani Planum, 25 January 2004 UTC	[10], [15]
Planned mission	90 sols	[1]
Last communication	10 June 2018	[15]
Total traverse	45.16 km over 14.375 years, the off-Earth wheeled odometry record	[10]
Twin	Spirit, Gusev crater, 3 January 2004, concluded 2010	[15]
Combined raw images returned by both rovers	more than 300,000	[15]

The NASA mission page dates the landing to 24 January 2004 [15]; the landing was at 05:05 UTC on 25 January, which is 24 January in Pacific time. Dates here are UTC. The same page is the only source held here for the combined image return, which it carries in the caption of an infographic rather than as a figure with a source of its own [15].

The arm carries a microscopic imager, Moessbauer spectrometer, alpha particle X-ray spectrometer, and a rock abrasion tool for grinding through weathering rinds to unaltered material [3].

Mobility

Meridiani Planum is flatter and less rocky than Gusev, which allowed long drives but exposed Opportunity to wind-formed ripples of loose sand that are geometrically unremarkable and therefore invisible to a hazard detector reasoning about shape [2]. Aeolian sand of that kind has a cohesion at or below 1 kPa and a friction angle near 30 degrees, against 4 to 25 kPa for cemented material, so the bearing capacity difference between a ripple and the surrounding plain is an order of magnitude with no corresponding difference in relief [9].

On sol 446 Opportunity executed 50 m of blind driving into what the imagery showed as an unremarkable pile of sand, and the wheels turned through 50 m of rotation while the vehicle advanced about 2 m [2]. Extraction took 39 sols. During recovery, commanded motions of 2 m with the wheels steered for a 23 degree heading change produced actual progress on the order of 1 mm on many sols [2]. Driving was terminated on sol 484 when visual odometry failed to converge three times, indicating the vehicle had finally moved far enough that the tracked features left the field of view.

The mechanism is the regenerative slip-sinkage coupling: rotation without advance displaces material rearward and downward, enlarging sinkage, contact patch and bulldozing resistance, which further reduces net thrust [5], [6]. Slip ratio, defined from commanded wheel surface velocity against measured vehicle velocity, approaches 1 as this proceeds, and slope traversal adds lateral slip on top of it. The Bekker pressure-sinkage relation with the Janosi-Hanamoto shear-displacement relation is the formulation used to model it, though measured wheel forces in dry granular media are reproduced more accurately by resistive force theory and by continuum plasticity.

The extraction procedure was built around visual odometry. After each commanded 2 m of driving, visual odometry estimated actual motion; if it confirmed little or no movement more commands were sent, and once it measured non-trivial motion or failed to converge, all driving stopped so the site could be examined [2]. The feature detector had been tuned for feature-rich natural terrain and often could not find enough features in bare sand, but the same pliability that trapped the rover made its own wheel tracks sharply visible and supplied the features needed for convergence.

Power and energy

Solar arrays with rechargeable batteries for the night. Array performance is the primary constraint on both allowed landing latitude and available surface power, and it is degraded by two separate dust processes: suspended atmospheric dust that modifies the spectrum and reduces intensity with optical depth and time of day, and deposited dust that obscures the cells cumulatively [4]. Wind gusts repeatedly cleared the arrays and restored power; that effect was not credited in the power budget and accounts for much of the mission duration. Removal requires a surface shear stress at or above the saltation threshold, which is not schedulable, while deposition continues whenever suspended dust is present, and background column optical depth stays below 1 through the aphelion half of the year with regional activity recurring at L_s 210 to 240 and 320 to 340 [7], [8].

Energy dominates winter operations: during low battery states of charge the mission canceled all telemetry transmission except the afternoon Odyssey relay pass [1]. An instrument deployment device heater that failed on drew power continuously, forcing development of a deep sleep mode that eliminated overnight drain at the cost of being unable to wake for early morning relay passes; it was introduced in July 2004 and used for the rest of that winter.

Thermal

Parameter	Value	Source
Enclosure	composite honeycomb exoskeleton lined with aerogel, closed by the equipment deck	[1]
Aerogel composition	99.8 percent air
Surface peak daytime temperature	-15 C	[4]
Surface nighttime minimum	-127 C
Transponder and SSPA allowable flight temperature	50 C (protoflight 60 C)	[1]
UHF transceiver allowable flight temperature	55 C (protoflight 70 C)

The cruise-stage heat rejection loop was severed at Mars arrival by design, so surface thermal control is passive plus heaters [1]. The diurnal swing the enclosure must absorb is large: at low latitude the air temperature range across a sol commonly exceeds 80 K, and measured air temperature at landed sites spans 140 to 300 K across the year [7]. The active constraint is on the hot side: because the transponder, the two solid-state power amplifiers and the UHF transceiver sit close together on the rover electronics module, temperatures climb through the day as transmitters run three or four times in succession.

Compute and avionics

Element	Specification	Source
Processor	RAD6000, 20 MHz	[2]
Bus to transponder	MIL-STD-1553, low power	[1]
Power distribution	rover power distribution unit
Inertial measurement unit	Litton LN-200
IMU attitude drift	under approximately 3 deg/h	[2]
Visual odometry image size	256 x 256 NavCam stereo pairs
Visual odometry time per step	up to three minutes, averaging nearly three

The 20 MHz RAD6000 is the same class of part flown across the MER program [3], and its cost per visual odometry update dictated the opposite design choice from terrestrial practice: rather than update frequently so that inter-frame motion stays small and tracking is cheap, MER images as rarely as possible and makes the tracker tolerant of large feature displacements [2].

Non-volatile storage degraded across the mission until Opportunity was operated without usable flash, downlinking all data before each shutdown.

Autonomy

Constraint	Value
NavCam field of view / height	45 deg / 1.5 m above ground
Required overlap between successive stereo pairs	at least 60 percent
Maximum straight or arcing drive per step	75 cm
Maximum heading change per step when turning in place	18 deg
Convergence rate, Opportunity	95 percent, 828 of 875
Convergence rate, Spirit	97 percent, 590 of 609
Maximum slip ratio measured	125 percent
Smallest displacement resolved	2 mm
Drive rate with visual odometry running continuously	approx 10 m/h

Constraints from [2].

Motion outside the per-step bounds forfeits the update for that step [2]. The bounds follow from the 45 degree camera field of view and the 1.5 m camera height, which together fix how much ground leaves the frame per meter driven. Because each update is expensive, visual odometry was commanded selectively: for short drives under about 15 m on slopes typically steeper than 10 degrees, when a wheel was being dragged, or when driving through sand [2]. Attitude was well maintained by the IMU, so visual odometry corrected position only during the first two years.

Purgatory changed the drive strategy permanently. Blind drive segments over sandy terrain were capped, initially at 5 m, after which a slip check, a short 20 cm step measured by visual odometry, verifies forward progress; if progress falls below a set fraction, driving stops [2]. Slip checks halted drives before they dug in on sols 501 and 603, allowing driving to resume the following sol rather than costing 39. The slip check bounds how deeply the rover can bury itself while retaining most of the speed of blind driving. It is a fault protection threshold on slip ratio, not a soil model: the underlying strength contrast that produces the runaway is not observable from the surface the rover is standing on [5].

Non-geometric hazards are not detectable onboard: a patch of loose sand has no distinguishing shape [2]. Predicting them requires soil state, and neither cohesion nor friction angle is observable from the rover’s own imagery or from orbit at drive scale [9]. Where such a hazard was suspected, human drivers defined keep-out zones manually, and the mobility sequence polled the corrected position against that list and halted on approach. Statistical mobility prediction over deformable terrain, which propagates uncertainty in soil parameters through a wheel-soil model to bound expected slip before a drive, was developed afterwards and was not available to MER [6].

Communications

X-band uses a small deep space transponder and solid-state power amplifier inside the warm electronics box, feeding either a steerable flat-panel phased array high-gain antenna or a near-omnidirectional low-gain antenna on the deck [1]. Bulk return is UHF relay from a 19 cm monopole through a CMC Electronics transceiver matched to the unit on Mars Odyssey.

Link	Parameter	Value
X-band LGA	Minimum uplink / downlink	7.8125 bps / 10 bps [1]
X-band HGA	Best uplink / downlink at shortest range	2 kbps / 28.8 kbps [1]
UHF	Rates supported either direction	8, 32, 128, 256 kbps [1]
UHF	Forward link as flown	fixed 8 kbps [1]
UHF	Return link per pass	128 or 256 kbps [1]
Relay geometry	Passes above 20 deg elevation	approx 1.8 per sol per orbiter, 2 to 8 min [1]

Return rate was chosen per pass between 128 and 256 kbps for maximum volume, and 256 kbps came into regular use in the extended missions [1]. Odyssey runs CCSDS Proximity-1 with Go-Back-N retransmission and reaches about 97 percent throughput despite up to 15 loss-of-lock gaps per pass. Mars Global Surveyor used the older Mars Balloon Relay protocol, transmitting only 13.3 to 13.8 s out of every 16 s and losing about two frames every 16 s at 128 kbps, so it carried only low-priority data and was dropped for Opportunity after sol 171 [1]. Through September 2005 about 92 percent of returned data went via Odyssey, 5 percent via MGS and 3 percent over the direct X-band link, over which interval the rover had driven 3158 m across 172 driving sols [2].

Modes of operation

Most sols include a five-minute carrier-only downlink, a beep when stationary and a honk when driving, confirming that the morning master sequence uploaded and started when a full direct-to-Earth session cannot be afforded in power, thermal or activity terms [1]. The master sequence is uplinked blind each morning, without confirmation that the command sweep succeeded, so the timing of the returned beep is the acknowledgement. Beeps are found first by open-loop fast Fourier transform and are detectable down to 12 dB-Hz [1].

Driving and imaging cannot be scheduled during a UHF pass, because pre-launch testing showed they can interfere with the UHF receiver [1]. Deep sleep is the low-power overnight mode described above. Communication windows are parameterized, including high-priority windows that override the default low-rate downlink during anomalies. Driving modes are parameterized in the same way: blind, visual-odometry-corrected, and blind with periodic slip checks, selected per segment against the terrain the previous sol’s imagery showed [2].

Ground operations

Operated from JPL. Relay passes are requested and sequenced onto the orbiters one to two weeks ahead, with unrequested passes left in a default configuration; tactical planning then works sol by sol over the next pass or two, optimizing UHF return against rover attitude, available energy and when the data is needed on the ground [1].

Vehicle pose is treated as a communications parameter. Data volume forecasts identify desirable parking yaw angles, handed to mobility planning as an end-of-drive constraint, and on occasion the rover was turned between two afternoon overflights purely to put the next orbiter in a high-gain region of the antenna pattern [1]. Where tilt was large enough to invalidate the zero-tilt forecast, the predictor was rerun with the estimated attitude.

Rover drivers carry two responsibilities unusual for a teleoperated vehicle: choosing camera pointing so that the terrain in frame contains enough trackable features for visual odometry to converge, and writing sequences that check convergence at every step [2]. Both follow from the 60 percent overlap and feature-count requirements above.

Long duration forced continuous adaptation. A steering actuator failed, fixing one wheel. A heater failed on, drawing power every sol [1]. Flash memory degraded until the rover operated without non-volatile storage. A planet-encircling dust storm in June 2018 raised column optical depth far above background; during the comparable MY34 event the global mean reached about 4 with local values of 5 to 10, and dust was lofted to 80 km [8]. Opportunity stopped responding and did not recover.

Technologies developed

MER flew visual odometry on another world for the first time. It began as a position-knowledge tool and became the basis of a family of operational safeguards: slip checks that bound how deeply a rover can dig itself in, keep-out zone enforcement against a corrected rather than a dead-reckoned position, and bounded position error while dragging a dead wheel [2]. Every subsequent Mars rover carries visual odometry, and Curiosity’s slip-check and full modes are direct descendants of the modes invented here in response to Purgatory and to the rock lodged in Spirit’s wheel on sol 339.

The flight record it produced, 95 percent convergence over 875 stereo pairs on Opportunity, slip measured to 125 percent and displacement to 2 mm, is the dataset against which subsequent wheel-soil and slip-prediction models are validated [2], [5], [6].