Skip to content
Space Robotics

Compression and Prioritization

A surface rover produces far more data than its downlink can carry in a sol, on a link whose rate is set by antenna geometry, the deep space network schedule and Mars-Earth range rather than by anything onboard. Every camera on Spirit and Opportunity produced a 1024x1024 image at 12 bits per pixel, and the average returned rate was 1.13 bits per pixel [1]. That ratio, roughly ten to one, is what made a Mars imaging mission possible on a link that returns about 56 Mb/sol. MSL’s pre-launch telecom plan assumed two 15-minute UHF passes per sol and 125 Mb per pass [27]. Reducing the byte count is only half the problem; the other half is deciding, sol by sol and pass by pass, which of the products worth sending goes first when not everything will fit. What is compressed, how hard, and in what order it is sent are three separate decisions, made in the flight compressor’s design, in ground-set byte quotas, and in a priority table the operations team revises for every communication window.

Families of approach

Section titled “Families of approach”

Five families address the problem, and they differ in what they measure, what they cost and whether they have flown.

  • Lossy and lossless image codecs. ICER, LOCO and LZO on MER [7], Rice and JPEG on Mars Pathfinder [2], JPEG on MSL [10]. All have flown.
  • Standardized codecs. CCSDS 122.0-B-2 wavelet image compression and CCSDS 121.0-B-3 Rice lossless compression [5][6]. The Rice coder is the basis of the Pathfinder compressor; the image standard has not flown on a Mars surface vehicle.
  • Reduction before compression. Companding, thumbnails, subframes, pixel averaging, row sums, column sums and histograms on MER [1], and onboard summarization products for microscopy [17]. The MER techniques have flown; the microscopy pipeline ran only in a field campaign.
  • Content-based prioritization. Onboard analysis of what an image contains, used to rank products for downlink: OASIS and its predecessors [13][14][15], SYNOPSIS [18], caption-based ranking [16]. WATCH on MER is the flown example of discarding frames without events [24]; AEGIS on Curiosity uses content analysis for targeting rather than downlink order [23].
  • Ground-set priority and transport. Priority tables, data product metadata and commanded reprioritization on MER [4][12], and file and bundle protocols such as CFDP and delay-tolerant networking [26][25][30].
  • Link and ground-resource scheduling. Relay pass planning, safe mode communication windows, Deep Space Network allocation, and onboard product selection under joint processor and bandwidth limits [31][32][33][34][35][36].

The flown image compressors

Section titled “The flown image compressors”

MER carries three compressors: ICER for lossy and progressive compression, a modified LOCO for lossless, and LZO for engineering data [7]. The cameras that feed them share one 12-bit CCD design behind seven optical assemblies [21], and the Microscopic Imager data are compressed after onboard processing and before packetization [20]. Sojourner’s Mars Pathfinder used a Rice compressor and JPEG [2]. One Pathfinder color mosaic used a 24:1 ratio on the red filter image [9] and 48:1 on the green and blue images [9].

ICER is wavelet-based and progressive: the compressed stream is organized so that reconstructions of successively higher quality can be produced as more of it arrives [1]. It partitions the image into independent error-containment segments so that a lost packet damages only its segment. Its parameters are the wavelet filter, the number of decomposition stages, the number of error-containment segments, the byte quota and a minimum loss parameter [1]. The progressive stream gives an operations team a byte-quota knob tied directly to output size.

LOCO exists alongside it because although ICER can compress losslessly, LOCO’s simple predictive approach is several times faster at similar effectiveness [1]. On six Mars Pathfinder test images at 12 bits per pixel, lossless rates averaged 8.78 bits/pixel for ICER, 8.84 for LOCO, 9.10 for JPEG 2000 and 9.36 for the CCSDS Rice compressor; the Rice implementation actually used by Pathfinder was worse still, ranging from 9.38 to 10.61 bits/pixel across the five single-frame images [2]. On the same images companded to 8 bits, ICER and LOCO both averaged 5.02 bits/pixel against 5.22 for JPEG 2000 and 5.52 for CCSDS Rice [2].

What was actually selected in flight

Section titled “What was actually selected in flight”

As of 7 February 2004 [1]:

CompressorImage typeImagesCombined area, MpixelVolume, MBAverage rate, bits/pixel
ICERregular51322867.9387.31.13
ICERthumbnail607524.93.81.27
LOCOregular, lossless489124.592.76.24
LOCOreference pixels111436.59.22.12
noneregular44.26.012.00

Source: [1], Table 1. Four images in the entire set were returned uncompressed. Lossless LOCO at 6.24 bits/pixel costs about five and a half times the volume of the ICER default, so it was reserved for cases where fidelity was the point [1].

Parameter selection varied by camera and by purpose. Panoramic camera full frames on Spirit used 3 or 4 decomposition stages, wavelet filters A or E and 1 to 20 segments, and averaged 1.12 bits/pixel over 1014 images; navigation camera full frames used 12 to 20 segments and averaged 0.97 bits/pixel over 235 images, the lowest rate of any full frame category [1]. Navigation images are not science products, which is why they were compressed hardest. Microscopic Imager frames took 24 to 32 segments and averaged 1.34 bits/pixel on Spirit and 2.36 on Opportunity [1]. Hazard camera full frames used 16 to 32 segments, the highest counts, because they are the images a lost packet can least afford to damage [1].

Two mechanisms reduce volume before compression runs. Images are companded from 12 to 8 bits per pixel to improve fidelity in dark regions, which is worthwhile because the sensor noise is signal dependent [1]. Every acquired image also produces a 64x64 thumbnail by averaging 16x16 blocks, ICER-compressed to about 650 bytes, which is about 0.005 bits per pixel of the original; thumbnails preview full images that may not be sent for some time. Beyond that, MER sends subframes of interest at full quality with the whole frame at much lower quality, performs pixel averaging, and sends row sums, column sums or histograms alone.

What it cost on the flight processor

Section titled “What it cost on the flight processor”

Timing for the MER implementation of ICER was measured under VxWorks on a 20 MHz RAD6000 identical to the flight units [2]. Three results hold regardless of the absolute scale: compression is faster when fewer compressed bytes are produced, the time at a bit rate of zero is essentially the cost of the wavelet decomposition alone, and the choice of wavelet filter and the number of decomposition stages have little effect on speed. The published figure gives compression time against output rate for a 1024x1024 test image with one error-containment segment, but its vertical axis is labeled in milliseconds per pixel, which cannot be literal at this image size, so the absolute value is not usable from the open document [2].

The byte quota does not give exact control over output size. ICER checks the quota only between compression of segments of subband bit planes, and does not account for the backlog still held in the segments’ interleaved entropy coders, so the output almost always exceeds the quota when the quality goal is not reached [2]. The overshoot depends on the segment count through two competing effects: more segments mean a larger total coder backlog, but also more frequent quota checks, so overshoot is minimized at a moderate segment count. MER controlled quality with the byte quota alone, setting the minimum loss parameter to zero, and transmitted the entire bitstream rather than truncating to the exact target, because truncation produces small quality variations between segments [1].

Standards

Section titled “Standards”

CCSDS publishes two relevant Blue Books. Lossless Data Compression, CCSDS 121.0-B-3, is a Rice-based coder in two stages: a preprocessor that decorrelates and maps prediction errors to nonnegative integers, and an adaptive entropy coder that picks the shortest of several code options for each block of 8, 16, 32 or 64 samples, with sample resolution up to 32 bits [6]. Only a unit-delay predictor is specified, any other is application-specific, and the stream carries no sync marker, so one channel bit error can corrupt the rest of a compressed sequence and recovery falls to the transport layer and the reference sample interval, which is capped at 4096 blocks [6]. This is the Rice algorithm in the Pathfinder comparison above.

Image Data Compression, CCSDS 122.0-B-2, applies a three-level two-dimensional 9/7 wavelet transform, in a floating point variant for lossy work and an integer variant that alone supports strictly lossless coding, followed by a bit-plane encoder [5]. Input depth is limited to 25 bits for the integer path and 27 unsigned or 28 signed bits for the float path. Rate is controlled per coded segment by a byte limit and a quality limit, whichever is reached first, and a segment can be truncated further at any point at the price of quality [5]. Neither document reports a compression ratio, throughput or power figure for any encoder [5][6]. No Mars surface vehicle has flown the CCSDS image compression standard; the flight compressors are mission-specific.

ICER-3D extends the ICER approach to hyperspectral imagery by compressing across the spectral dimension as well as the spatial ones [3]. It has not flown as a compressor on a space vehicle.

For transport rather than coding, CCSDS 727.0-B-5 defines the CFDP file delivery protocol with acknowledged and unacknowledged modes, and leaves timer and retry limits to the implementer [26]. MER chose not to use a protocol stack for retransmission and reprioritization, and managed files and metadata directly instead [12]. The DINET experiment ran the delay-tolerant Bundle Protocol on the Deep Impact spacecraft at 24 million km for 27 days and delivered every bundle despite service lapses and an injected 3.125 percent segment loss; priority queuing gave a delivery acceleration of 1.10 to 1.12, and only one of the ten nodes was a spacecraft [25].

Later missions

Section titled “Later missions”

MSL kept the mission-specific pattern but moved back toward a standard codec for its Mastcams and Descent Imager: most Mastcam and MARDI images are downlinked as lossy JPEG, compressed onboard when the scientific need for full fidelity does not outweigh the downlink volume it costs, with a lossless mode available separately for the cases that need it [10]. Images are stored uncompressed in onboard memory and the compression is specified when the product is chosen for downlink [10][19]. That is a narrower design than MER’s: one general-purpose codec with a selectable quality setting, rather than a wavelet compressor tuned per camera and per segment count.

Compression of non-image data is a separate, quieter problem that the imaging literature does not cover. On Mars 2020, GZIP compression of non-image telemetry removed more than two thirds of the volume actually generated in the first 216 sols, taking 88.8 Gibit of source data down to 27.5 Gibit on the link, against a daily total data return of 1.5 to 2 Gibit per sol [11]. That is compression applied to housekeeping and engineering products, not camera frames. The M2020 UHF return link tops out at 2048 kbit/s with LDPC encoding [11].

Onboard prioritization by content

Section titled “Onboard prioritization by content”

A second route to more science per bit is to rank or discard data by what it contains. Rover traverse distance grew faster than downlink capacity, which is the stated motivation of the OASIS line of work at JPL [13]. OASIS finds rocks in navigation and hazard images, extracts albedo, Gabor-filter texture and shape measures, and ranks images by three criteria: match to a scientist-specified target signature, novelty, and representativeness by clustering [15]. The 2004 paper merges the three scores with an adjustable non-linear function so that no one criterion captures the downlink set, and couples the analysis to onboard replanning with two reactions, stop and call home, and request another measurement [15]. The 2003 description proposes validating the ranking against expert geologists’ pair comparisons [14]. Neither paper, nor the 2002 poster, reports a detection rate, a runtime or a data saving [13][14][15]; OASIS is a described system, not a measured one in these sources.

Three deployed capabilities sit at the edge of this family. WATCH was uploaded to both MER rovers in the R9.2 flight software load in summer 2006; it screens Navcam frames and saves only those containing a detected dust devil or cloud, which cut image volume by about 70 percent [24]. It was tuned toward false alarms because a missed event cannot be recovered, and 76 percent of detections were near-field ground, so only about 21 percent of 533 features were even candidate dust devils [24]. AEGIS selects ChemCam targets after a drive without waiting for Earth: it was used 54 times between sols 1343 and 1663, placed all LIBS points on the preferred material for 44 of 47 targets against 10 of 42 for modeled blind targeting, and ran in 150 to 450 seconds on the RAD750 [23]. AEGIS reduces the cost of choosing a measurement, not the volume of a downlink queue.

Newer work has not flown. SCOTI trains a VGG-19 and LSTM captioning network on about 1,250 expert captions of Curiosity images and ranks images by scientist text queries; it is a proof of concept with no flight processor run and no downlink saving measured [16]. SYNOPSIS is a C library integrated into cFS and F Prime that ranks products by a science utility estimate and a diversity descriptor, and in a replay of archived ChemCam data from sols 2828 to 2837 under a hypothetical degraded link it accumulated utility faster than a first-in first-out queue; runtime on flight-like processors is stated as future work [18]. The Ocean Worlds Life Surveyor study reduced 1.23 GB of holographic and fluorescence microscopy to summary products at average compression ratios of 1888 and 2247, in about 108 and 110 seconds on a 2.1 GHz laptop-class processor, not flight hardware [17].

Section titled “How downlink priority is set”

Priority on MER is a property of the data product and it is set and revised from the ground. Each communication window command carries the durations for the real-time and recorded data priority tables along with start time, duration, antenna, rates and hardware configuration, and the window executes inside the communications behavior part of the flight software rather than through the sequence engine [4]. Every downlinkable item is a data product, a file with a metadata instance recording its source, type, creating command, validity time and size, and the ground uses that metadata to retransmit, reprioritize or delete [12]. Reported deficiencies of the as-flown design include a single retransmit range per product [12]. Data products also carry a priority in their metadata file alongside collection time [8].

Onboard data organization places the highest priority playback data, fault and warning event reports and spacecraft health reports, earliest in a direct-to-Earth pass [4]. That ordering interacts with receiver lockup: lockup takes about 1 minute for high gain antenna rates but varies from pass to pass, so the project must plan a lockup allowance before the valuable playback starts. Too short and sent data is lost; too long and less data fits. The comm window lockup parameter was changed from 1 minute to 3 minutes for downlink rates of 3160 bit/s and higher, and 3 minutes consumes 10 percent of the return capacity of a 30 minute window; it was reduced to 2 minutes for some windows in the later extended missions [4].

Reprioritization is a routine operations activity. UHF reports were generated for every Odyssey pass at low priority in the extended mission by attaching a report sequence to the start of window preparation; to actually return one, data management issued a command raising its priority. Reports were reprioritized every seventh sol so telecom could spot-check relay performance, could be raised individually for interesting passes, and were marked for automatic deletion after 7 sols [4].

Priority also serves fault diagnosis. During the Spirit sol 18 flash anomaly a high priority communication window overrode the 10 bit/s fault default to give 40 or 300 bit/s on the low gain antenna, and the repeating event reports that came down at that rate are what first pointed at the flash memory [4][8]. The Remote Agent design for Deep Space 1 set its downlink priority table differently for autonomous and partially autonomous operations, but that paper was written before the flight experiment and reports no measured result [22].

Relay, networking and scheduling

Section titled “Relay, networking and scheduling”

The capacity that priority is spent against is itself planned. Perseverance relay planning precomputes the data volume of every overflight at every allowable rate and duration, then filters by passes per sol, minimum and maximum volume and latency to the next decisional pass; LDPC coding commissioned on the Mars Relay Network in 2021 returned 70 percent more data on four MAVEN overflights than convolutional coding predicts, a figure measured against prediction rather than paired passes [31]. For MSL, InSight sharing orbiter view periods and the non sun synchronous MAVEN and TGO overflights moved the data arrival time against the roughly 10:00 LMST uplink, so what reaches the ground before the tactical shift became a scheduling constraint; the tools are described as designed, with no before and after efficiency figure [32].

Safe mode shows the same trade in power terms. Curiosity’s fault protection windows set when the rover wakes and which antenna it uses, and because it is power negative while the compute element is on, window length trades communication against recharge: nine Martian hours of daily safe mode communication gained 22 percent state of charge per sol at landing, and by 2023 only 6.6 hours were affordable at a gain of 5.5 percent [33]. These are model outputs checked against trending, not measurements of a safe mode entry [33].

On the ground side, the Deep Space Network schedules about 500 activities a week for about 40 users and has been oversubscribed by nearly 60 percent at peak; a ranked mission priority list was rejected for it in favor of user preferences with automated optimization, tested on synthetic preferences and one field-test week [35]. A demand access prototype lets a spacecraft claim shared antenna time with a beacon tone, and was evaluated only in simulation with three lunar spacecraft [34].

Delay-tolerant networking carries priority on the wire. The CCSDS Bundle Protocol profile makes the Extended Class of Service block mandatory, with 256 ordinal priority levels within expedited service, so time-critical telemetry can be separated from bulk science on one link; the document gives no throughput or storage figure [30]. On the ISS, ground-equivalent hardware running ION delivered about 15 Mb/s downlink against a 20 Mb/s allocation and lost no data across a 10 minute loss of signal or a gateway reboot, but the on-orbit service had been checked out only with bping and CFDP when reported [29]. The fprime-dtn prototype integrates ION into the F Prime framework and has not run on flight hardware, so it reports no memory, CPU or throughput figure [28].

Two studies apply the same prioritization argument to other resources. A proposed HyspIRI scheduler treats onboard processing and downlink as capacity resources and selects which products to generate against a roughly 1 Gb/s instrument and 15 Mb/s link; it was never flown as described and reports no scheduler runtime [36]. For multi-robot mapping, ranking loop closure candidates by predicted uncertainty reduction allowed computing 0.5 percent of candidates on four underground datasets and cut median localization error 75 percent against a baseline, but observability ranking did worse than random in one environment [37].

What is not established

Section titled “What is not established”

The MER compression numbers are drawn from a fixed operational snapshot: the vehicle-wide totals are as of a single date in February 2004, and the earlier per-image test that justified choosing ICER over LOCO and JPEG 2000 used six Pathfinder images, five of them the same scene from one camera, with the authors stating outright that the filter ranking may not hold for other image types [1][2]. No compression time figure from the flight processor survives with a usable absolute scale; the published curve is mislabeled and only the qualitative relationships it shows are usable [2]. The ICER-3D hyperspectral results are airborne AVIRIS data, not a flight compressor running on any space vehicle, and its own authors show it losing to a simpler compressor in the lossless case [3]. Neither CCSDS standard gives a compression ratio, throughput or power figure for any encoder [5][6].

No source cited above gives a byte-quota overshoot statistic, a rate-distortion or image-quality number for any flown compressor, or a packet loss rate for MER’s downlink; the interaction between segment count and overshoot is described only qualitatively [1][2]. The priority and lockup mechanics are reconstructed from a single mission’s operations account, written by the team that ran it, so the generalization from MER’s tables and lockup allowances to other missions’ downlink scheduling is not demonstrated here [4]. The MER data management paper reports no throughput, latency or onboard cost for its system [12].

For content-based prioritization, OASIS has no reported accuracy or runtime [13][14][15], the WATCH detector has no recall figure because no ground truth exists for the flight imagery [24], and AEGIS’s blind-targeting comparison is modeled rather than executed [23]. SCOTI and SYNOPSIS have not run on flight hardware [16][18]. The DTN result rests on one spacecraft node and reports two of four metrics meeting criteria only after post-hoc exclusions [25]. Whether any content-based method reduces the bytes returned per sol on a flown rover is not shown by these sources. The relay, DSN and DTN results rest on predictions, simulation or ground-equivalent hardware, and none reports a flight measurement of bytes returned per sol under a priority scheme [28][29][31][34].

References

  1. Kiely, A. and Klimesh, M. (2004). Preliminary Image Compression Results from the Mars Exploration Rovers . IPN Progress Report. Source
    BibTeX
    @article{kiely2004preliminary,
  title = {Preliminary Image Compression Results from the Mars Exploration Rovers},
  author = {Kiely, A. and Klimesh, M.},
  journal = {IPN Progress Report},
  volume = {42-156},
  institution = {Jet Propulsion Laboratory},
  year = {2004},
  url = {https://ipnpr.jpl.nasa.gov/progress_report/42-156/156I.pdf}
}
  2. Kiely, A. and Klimesh, M. (2003). The ICER Progressive Wavelet Image Compressor . IPN Progress Report. Source
    BibTeX
    @article{kiely2003icer,
  title = {The ICER Progressive Wavelet Image Compressor},
  author = {Kiely, A. and Klimesh, M.},
  journal = {IPN Progress Report},
  volume = {42-155},
  institution = {Jet Propulsion Laboratory},
  year = {2003},
  url = {https://ipnpr.jpl.nasa.gov/progress_report/42-155/155J.pdf}
}
  3. Kiely, A., Klimesh, M., Xie, H. and Aranki, N. (2006). ICER-3D: A Progressive Wavelet-Based Compressor for Hyperspectral Images . IPN Progress Report. Source
    BibTeX
    @article{kiely2006icer3d,
  title = {ICER-3D: A Progressive Wavelet-Based Compressor for Hyperspectral Images},
  author = {Kiely, A. and Klimesh, M. and Xie, Heping and Aranki, N.},
  journal = {IPN Progress Report},
  volume = {42-164},
  institution = {Jet Propulsion Laboratory},
  year = {2006},
  url = {https://ipnpr.jpl.nasa.gov/progress_report/42-164/164A.pdf}
}
  4. Taylor, J., Makovsky, A., Barbieri, A., Tung, R., Estabrook, P. and Thomas, A. G. (2014). Mars Exploration Rover Telecommunications . Deep Space Communications. Source
    BibTeX
    @incollection{taylor2014mars,
  title = {Mars Exploration Rover Telecommunications},
  author = {Taylor, Jim and Makovsky, Andre and Barbieri, Andrea and Tung, Ramona and Estabrook, Polly and Thomas, A. Gail},
  booktitle = {Deep Space Communications},
  series = {DESCANSO Design and Performance Summary Series},
  publisher = {Jet Propulsion Laboratory, California Institute of Technology},
  chapter = {7},
  year = {2014},
  url = {https://descanso.jpl.nasa.gov/monograph/series13/DeepCommo_Chapter7--141030.pdf}
}
  5. CCSDS. (2017). Image Data Compression . Consultative Committee for Space Data Systems, CCSDS 122.0-B-2, Blue Book. Source
    BibTeX
    @techreport{ccsds2017image,
  title = {Image Data Compression},
  author = {{CCSDS}},
  number = {CCSDS 122.0-B-2, Blue Book},
  institution = {Consultative Committee for Space Data Systems},
  year = {2017},
  url = {https://ccsds.org/Pubs/122x0b2e1.pdf}
}
  6. CCSDS. (2020). Lossless Data Compression . Consultative Committee for Space Data Systems, CCSDS 121.0-B-3, Blue Book. Source
    BibTeX
    @techreport{ccsds2020lossless,
  title = {Lossless Data Compression},
  author = {{CCSDS}},
  number = {CCSDS 121.0-B-3, Blue Book},
  institution = {Consultative Committee for Space Data Systems},
  year = {2020},
  url = {https://ccsds.org/Pubs/121x0b3.pdf}
}
  7. Reeves, G. E. (2005). An Overview of the Mars Exploration Rovers Flight Software . IEEE International Conference on Systems, Man and Cybernetics. Source
    BibTeX
    @inproceedings{reeves2005overview,
  title = {An Overview of the Mars Exploration Rovers Flight Software},
  author = {Reeves, Glenn E.},
  booktitle = {IEEE International Conference on Systems, Man and Cybernetics},
  volume = {1},
  pages = {1-7},
  address = {Waikoloa, Hawaii},
  year = {2005},
  doi = {10.1109/icsmc.2005.1571113},
  abstract = {The Mars exploration rovers (MER) flight software (FSW) is possibly the most complex software implementation to be deployed on another planet. The requirements dictated a software system that addressed four distinct mission phases (cruise, landing, egress, and surface) and the mission demanded a system with significant autonomy. The structure of the MER flight software reflects its object-oriented beginnings and the overall function reflects the requirements of the MER mission and spacecraft. This paper provides an overview of the function and structure of the MER flight software. The MER mission and spacecraft are briefly discussed to provide context for the flight software decomposition and the discussion of the software execution model.}
}
  8. Reeves, G. E. and Neilson, T. A. (2005). The Mars Rover Spirit FLASH Anomaly . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{reeves2005mars,
  title = {The Mars Rover Spirit FLASH Anomaly},
  author = {Reeves, Glenn E. and Neilson, Tracy A.},
  booktitle = {IEEE Aerospace Conference},
  pages = {4186-4199},
  address = {Big Sky, Montana},
  year = {2005},
  doi = {10.1109/aero.2005.1559723},
  abstract = {The Mars Exploration Rover "Spirit" suffered a debilitating anomaly that prevented communication with Earth for several anxious days. With the eyes of the world upon us, the anomaly team used each scrap of information, our knowledge of the system, and sheer determination to analyze and fix the problem, then return the vehicle to normal operation. This paper will discuss the Spirit FLASH anomaly, including the drama of the investigation, the root cause and the lessons learned from the experience}
}
  9. Golombek, M. P., Anderson, R. C., Barnes, J. R., Bell, J. F., Bridges, N. T., Britt, D. T., Brückner, J., Cook, R. A., Crisp, D., Crisp, J. A., Economou, T., Folkner, W. M., Greeley, R., Haberle, R. M., Hargraves, R. B., Harris, J. A., Haldemann, A. F. C., Herkenhoff, K. E., Hviid, S. F., Jaumann, R., Johnson, J. R., Kallemeyn, P. H., Keller, H. U., Kirk, R. L., Knudsen, J. M., Larsen, S., Lemmon, M. T., Madsen, M. B., Magalhães, J. A., Maki, J. N., Malin, M. C., Manning, R. M., Matijevic, J., McSween, H. Y., Moore, H. J., Murchie, S. L., Murphy, J. R., Parker, T. J., Rieder, R., Rivellini, T. P., Schofield, J. T., Seiff, A., Singer, R. B., Smith, P. H., Soderblom, L. A., Spencer, D. A., Stoker, C. R., Sullivan, R., Thomas, N., Thurman, S. W., Tomasko, M. G., Vaughan, R. M., Wänke, H., Ward, A. W. and Wilson, G. R. (1999). Overview of the Mars Pathfinder Mission: Launch through landing, surface operations, data sets, and science results . Space Science Reviews, E4. Source
    BibTeX
    @article{golombek1999overview,
  title = {Overview of the Mars Pathfinder Mission: Launch through landing, surface operations, data sets, and science results},
  author = {Golombek, Matthew P. and Anderson, R. C. and Barnes, J. R. and Bell, J. F. and Bridges, Nathan T. and Britt, Daniel T. and Brückner, J. and Cook, Richard A. and Crisp, D. and Crisp, Joy A. and Economou, T. and Folkner, W. M. and Greeley, R. and Haberle, R. M. and Hargraves, R. B. and Harris, J. A. and Haldemann, Albert F. C. and Herkenhoff, K. E. and Hviid, S. F. and Jaumann, Ralf and Johnson, Jeffrey R. and Kallemeyn, P. H. and Keller, Horst Uwe and Kirk, Randolph L. and Knudsen, J. M. and Larsen, S. and Lemmon, Mark T. and Madsen, M. B. and Magalhães, J. A. and Maki, Justin N. and Malin, Michal C. and Manning, Robert M. and Matijevic, J. and McSween, H. Y. and Moore, Henry J. and Murchie, Scott L. and Murphy, J. R. and Parker, Timothy J. and Rieder, R. and Rivellini, T. P. and Schofield, John T. and Seiff, A. and Singer, R. B. and Smith, Peter H. and Soderblom, L. A. and Spencer, David A. and Stoker, Carol R. and Sullivan, R. and Thomas, N. and Thurman, Sam W. and Tomasko, M. G. and Vaughan, R. M. and Wänke, H. and Ward, A. W. and Wilson, G. R.},
  journal = {Space Science Reviews},
  volume = {104},
  number = {E4},
  pages = {8523-8553},
  publisher = {American Geophysical Union (AGU)},
  year = {1999},
  doi = {10.1029/98je02554},
  abstract = {Mars Pathfinder successfully landed at Ares Vallis on July 4, 1997, deployed and navigated a small rover about 100 m clockwise around the lander, and collected data from three science instruments and ten technology experiments. The mission operated for three months and returned 2.3 Gbits of data, including over 16,500 lander and 550 rover images, 16 chemical analyses of rocks and soil, and 8.5 million individual temperature, pressure and wind measurements. Path‐finder is the best known location on Mars, having been clearly identified with respect to other features on the surface by correlating five prominent horizon features and two small craters in lander images with those in high‐resolution orbiter images and in inertial space from two‐way ranging and Doppler tracking. Tracking of the lander has fixed the spin pole of Mars, determined the precession rate since Viking 20 years ago, and indicates a polar moment of inertia, which constrains a central metallic core to be between 1300 and ∼2000 km in radius. Dark rocks appear to be high in silica and geochemically similar to anorogenic andesites; lighter rocks are richer in sulfur and lower in silica, consistent with being coated with various amounts of dust. Rover and lander images show rocks with a variety of morphologies, fabrics and textures, suggesting a variety of rock types are present. Rounded pebbles and cobbles on the surface as well as rounded bumps and pits on some rocks indicate these rocks may be conglomerates (although other explanations are also possible), which almost definitely require liquid water to form and a warmer and wetter past. Air‐borne dust is composed of composite silicate particles with a small fraction of a highly magnetic mineral, interpreted to be most likely maghemite; explanations suggest iron was dissolved from crustal materials during an active hydrologic cycle with maghemite freeze dried onto silicate dust grains. Remote sensing data at a scale of a kilometer or greater and an Earth analog correctly predicted a rocky plain safe for landing and roving with a variety of rocks deposited by catstrophic floods, which are relatively dust free. The surface appears to have changed little since it formed billions of years ago, with the exception that eolian activity may have deflated the surface by ∼3–7 cm, sculpted wind tails, collected sand into dunes, and eroded ventifacts (fluted and grooved rocks). Pathfinder found a dusty lower atmosphere, early morning water ice clouds, and morning near‐surface air temperatures that changed abruptly with time and height. Small scale vortices, interpreted to be dust devils, were observed repeatedly in the afternoon by the meteorology instruments and have been imaged.}
}
  10. Malin, M. C., Ravine, M. A., Caplinger, M. A., Ghaemi, F. T., Schaffner, J. A., Maki, J. N., Bell III, J. F., Cameron, J. F., Dietrich, W. E., Edgett, K. S., Edwards, L., Garvin, J. B., Hallet, B., Herkenhoff, K. E., Heydari, E., Kah, L. C., Lemmon, M. T., Minitti, M. E., Olson, T., Parker, T. J., Rowland, S. K., Schieber, J., Sletten, R., Sullivan, R., Sumner, D. Y., Yingst, R. A., Duston, B., McNair, S. and Jensen, E. (2017). The Mars Science Laboratory (MSL) Mast cameras and Descent imager: I. Investigation and instrument descriptions . Earth and Space Science, 8. Source
    BibTeX
    @article{malin2017mars,
  title = {The Mars Science Laboratory (MSL) Mast cameras and Descent imager: I. Investigation and instrument descriptions},
  author = {Malin, Michal C. and Ravine, Michael A. and Caplinger, Michael A. and Ghaemi, F. Tony and Schaffner, Jacob A. and Maki, Justin N. and Bell III, James F. and Cameron, James F. and Dietrich, William E. and Edgett, Kenneth S. and Edwards, Laurence and Garvin, James B. and Hallet, B. and Herkenhoff, K. E. and Heydari, Ezat and Kah, Linda C. and Lemmon, Mark T. and Minitti, Michelle E. and Olson, T. and Parker, Timothy J. and Rowland, Scott K. and Schieber, Juergen and Sletten, Ron and Sullivan, R. and Sumner, Dawn Y. and Yingst, R. Aileen and Duston, Brian and McNair, Sean and Jensen, Elsa},
  journal = {Earth and Space Science},
  volume = {4},
  number = {8},
  pages = {506-539},
  publisher = {American Geophysical Union (AGU)},
  year = {2017},
  doi = {10.1002/2016ea000252},
  abstract = {Abstract The Mars Science Laboratory Mast camera and Descent Imager investigations were designed, built, and operated by Malin Space Science Systems of San Diego, CA. They share common electronics and focal plane designs but have different optics. There are two Mastcams of dissimilar focal length. The Mastcam‐34 has an f /8, 34 mm focal length lens, and the M‐100 an f /10, 100 mm focal length lens. The M‐34 field of view is about 20° × 15° with an instantaneous field of view (IFOV) of 218 μrad; the M‐100 field of view (FOV) is 6.8° × 5.1° with an IFOV of 74 μrad. The M‐34 can focus from 0.5 m to infinity, and the M‐100 from ~1.6 m to infinity. All three cameras can acquire color images through a Bayer color filter array, and the Mastcams can also acquire images through seven science filters. Images are ≤1600 pixels wide by 1200 pixels tall. The Mastcams, mounted on the ~2 m tall Remote Sensing Mast, have a 360° azimuth and ~180° elevation field of regard. Mars Descent Imager is fixed‐mounted to the bottom left front side of the rover at ~66 cm above the surface. Its fixed focus lens is in focus from ~2 m to infinity, but out of focus at 66 cm. The f /3 lens has a FOV of ~70° by 52° across and along the direction of motion, with an IFOV of 0.76 mrad. All cameras can acquire video at 4 frames/second for full frames or 720p HD at 6 fps. Images can be processed using lossy Joint Photographic Experts Group and predictive lossless compression.}
}
  11. Girerd, A. R., Kuhn, S., Roth, B., Gaines, D., Scandore, S., Cummings, D., Mendoza, R., Lefland, M., Bareh, M., Siegfriedt, R., Lenda, M., Reich, K., Shah, B. and Bohannon, E. (2022). Cross-Cutting Flight Infrastructure Improvements on M2020 . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{girerd2022cross,
  title = {Cross-Cutting Flight Infrastructure Improvements on M2020},
  author = {Girerd, Andre R and Kuhn, Stephen and Roth, Brian and Gaines, Dan and Scandore, Steve and Cummings, David and Mendoza, Ricardo and Lefland, Mallory and Bareh, Magdy and Siegfriedt, Rebekah and Lenda, Matthew and Reich, Kevin and Shah, Biren and Bohannon, Emily},
  booktitle = {IEEE Aerospace Conference},
  pages = {1-15},
  publisher = {IEEE},
  year = {2022},
  doi = {10.1109/aero53065.2022.9843523},
  abstract = {Mars2020 (M2020) was formulated as a mission that leveraged as much Mars Science Laboratory (MSL) heritage as possible, while focusing major new development efforts on the original and unique elements needed to accomplish the different mission objectives. Well publicized examples of high profile new developments include precision landing, the sampling and caching system, the specific instrument suite, improved mobility via Autonomous Navigation, and later the addition of the Ingenuity helicopter. Less well known are the refinements to the core flight infrastructure, primarily in the cross-cutting functions of Telecom, Avionics, Data Management, Communications Behaviors, and Parameter Management. These enhancements are introduced predominately via flight software, and represent increases in capability that justified their inclusion in an otherwise heritage-focused project environment. Perseverance's cross-cutting flight infrastructure improvements fall into and across the following five categories. First is a trimming of the software footprint of infrastructure modules, in order to make room for memory demands elsewhere in the system. Second is the minimization of data volume to be downlinked, through various methods such as the incorporation of new compression options. Third is the maximization of the available downlink bandwidth for data, by curtailing content-less data (fill) and introducing an improved UHF proximity link protocol. Fourth is a reduction in vulnerabilities, through increased file system redundancy, robustness, and software process monitoring. Fifth is an increase in operations efficiency by lowering file system mount times, improving parallelism between simultaneous events, minimizing the time to recover from file system errors, streamlining the purging of obsolete data, and reducing the number of commands to service parameters by a factor of 100. Individually, none of the cross-cutting infrastructure improvements are likely to garner headlines, but collectively they appreciably improve the safety and operability of Perseverance over its predecessor. This paper will describe the improvements, their promise, and where applicable, their actual impact in operations.}
}
  12. Snyder, J. F. and Smyth, D. E. (2004). Data management for Mars Exploration Rovers . Aerodynamic Decelerators Conference. Source
    BibTeX
    @inproceedings{snyder2004data,
  title = {Data management for Mars Exploration Rovers},
  author = {Snyder, Joseph F. and Smyth, David E.},
  booktitle = {Aerodynamic Decelerators Conference},
  publisher = {JPL Open Repository},
  year = {2004},
  url = {https://dataverse.jpl.nasa.gov/dataset.xhtml?persistentId=hdl:2014/40490}
}
  13. Judd, M., Mazzonie, D., DeCosta, D., Gaines, D., Estlin, T., Anderson, R. C. and Castano, R. (2002). Mars rover image data prioritization for increases mission science return . AGU Fall Meeting. Source
    BibTeX
    @inproceedings{judd2002mars,
  title = {Mars rover image data prioritization for increases mission science return},
  author = {Judd, M. and Mazzonie, D. and DeCosta, D. and Gaines, D. and Estlin, T. and Anderson, R. C. and Castano, R.},
  booktitle = {AGU Fall Meeting},
  year = {2002},
  url = {https://ntrs.nasa.gov/citations/20060042536}
}
  14. Castano, R., Anderson, R. C., Estlin, T., DeCoste, D., Fisher, F., Gaines, D., Mazzoni, D. and Judd, M. (2003). Rover traverse science for increased mission science return . SPIE Photonics West. Source
    BibTeX
    @inproceedings{castano2003rover,
  title = {Rover traverse science for increased mission science return},
  author = {Castano, Rebecca and Anderson, Robert C. and Estlin, Tara and DeCoste, Dennis and Fisher, Forest and Gaines, Daniel and Mazzoni, Dominic and Judd, Michele},
  booktitle = {SPIE Photonics West},
  publisher = {JPL Open Repository},
  year = {2003},
  url = {https://dataverse.jpl.nasa.gov/dataset.xhtml?persistentId=hdl:2014/39345}
}
  15. Castano, R., Judd, M., Estlin, T., Anderson, R. C., Scharenbroich, L., Song, L., Gaines, D., Fisher, F., Mazzoni, D. and Castano, A. (2004). Autonomous onboard traverse science system . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{castano2004autonomous,
  title = {Autonomous onboard traverse science system},
  author = {Castano, Rebecca and Judd, M. and Estlin, T. and Anderson, R. C. and Scharenbroich, L. and Song, L. and Gaines, D. and Fisher, F. and Mazzoni, Dominic and Castano, Andres},
  booktitle = {IEEE Aerospace Conference},
  pages = {159-167},
  publisher = {IEEE},
  year = {2004},
  doi = {10.1109/aero.2004.1367601},
  abstract = {The Onboard Autonomous Science Investigation System (OASIS) is a technology for increasing science return during rover traverses by prioritizing data onboard, and identifying and reacting to unanticipated science opportunities. Rovers of the future will have the capacity to collect more data than can be downlinked back to Earth. OASIS can increase mission science return by carefully selecting the data with the highest science interest for downlink. These rovers may also be required to traverse long distances with little to no interaction with the science team on Earth. OASIS can act as a geologist's assistant and can autonomously direct the rover to take additional measurements of "interesting" rocks. The importance of characterizing the terrain along these traverses, a study that is now becoming known as traverse science, increases with the distances the rover must travel. This paper provides a brief overview of the entire OASIS system and how it analyzes one type of data - grayscale images taken by the rover for engineering and hazard avoidance purposes. Although the OASIS system can apply the same type of analysis to different data types, such as color images, hyperspectral images or point spectrometer data, we will only focus on grayscale images here. The paper also describes the latest advances in two key aspects of the system: image prioritization and the science alert. In image prioritization, we combine the results from three distinct prioritization methods to arrive at an overall downlink ranking of the images collected during a traverse. The science alert is a capability that enables the rover to identify and react to a pre-specified, and scientifically significant, signature. Once this signature has been detected via the onboard science analysis component, the planning and scheduling module updates the rover command sequence to stop the traverse and signal Earth of the find. If there is sufficient time and onboard resources before the next downlink opportunity, additional data samples of the target may be autonomously collected.}
}
  16. Ono, H. (2020). SCOTI: Science Captioning of Terrain Images for Data Prioritization and Local Image Search . Planetary and Space Science. Source
    BibTeX
    @article{ono2020scoti,
  title = {SCOTI: Science Captioning of Terrain Images for Data Prioritization and Local Image Search},
  author = {Ono, Hiro},
  journal = {Planetary and Space Science},
  volume = {188},
  pages = {104943},
  year = {2020},
  doi = {10.1016/j.pss.2020.104943},
  abstract = {Planetary exploration is full of challenges. Data bandwidth is very limited between planetary rovers and ground-based data system. What’s worse, even though NASA has accumulated over 34 million images from various missions, it requires significant effort and is hardly possible for any scientist to go through all of them. In order to improve the degree of automation and the efficiency of these processes, we propose a system leveraging machine learning for planetary rovers to actively look for scientifically interesting and valuable features according to text instructions from scientists and prioritize the images captured onboard with those features for downlink. Such an image prioritization mechanism can also be naturally applied to content-based image search through text description in any local planetary image data server, allowing scientists to search for images with desired features without going through them one by one. Besides theoretical and engineering details of our proposed approach, we also present both quantitative and qualitative evaluation of the system along with some concrete examples.}
}
  17. Wronkiewicz, M., Lee, J., Lightholder, J., Doran, G., Mauceri, S., Schibler, T., Moorjani, E., Nadeau, J., Lindensmith, C. and Mandrake, L. (2023). Onboard Science Instrument Autonomy for the Detection of Microscopy Biosignatures on the Ocean Worlds Life Surveyor . AIAA Journal of Spacecraft and Rockets: MAPLD. Source
    BibTeX
    @article{wronkiewicz2023onboard,
  title = {Onboard Science Instrument Autonomy for the Detection of Microscopy Biosignatures on the Ocean Worlds Life Surveyor},
  author = {Wronkiewicz, Mark and Lee, Jake and Lightholder, Jack and Doran, Gary and Mauceri, Steffen and Schibler, Thomas and Moorjani, Eshaan and Nadeau, Jay and Lindensmith, Chris and Mandrake, Lukas},
  journal = {AIAA Journal of Spacecraft and Rockets: MAPLD},
  publisher = {JPL Open Repository},
  year = {2023},
  doi = {10.48577/jpl.mflyra}
}
  18. Wronkiewicz, M., Doran, G., Lightholder, J., Hasnain, Z., Stockham, R., Brown, R. J., Zemerick, S. A. and Mandrake, L. (2026). A Flight-Ready Software Platform for Deploying Science Autonomy Onboard Spacecraft . IEEE Transactions on Aerospace and Electronic Systems. Source
    BibTeX
    @article{wronkiewicz2026flight,
  title = {A Flight-Ready Software Platform for Deploying Science Autonomy Onboard Spacecraft},
  author = {Wronkiewicz, Mark and Doran, Gary and Lightholder, Jack and Hasnain, Zaki and Stockham, Rex and Brown, Robert J. and Zemerick, Scott A. and Mandrake, Lukas},
  journal = {IEEE Transactions on Aerospace and Electronic Systems},
  volume = {62},
  pages = {3879--3890},
  year = {2026},
  doi = {10.1109/taes.2025.3650728},
  abstract = {Spacecraft exploring Earth and other bodies in the solar system are producing increasingly large and complex volumes of scientific data. Recently, the field of science autonomy has begun to develop data analysis algorithms to select, prioritize, and characterize science instrument data directly onboard the spacecraft. These data-derived insights can then inform downstream autonomous decisions like data prioritization to mitigate data bandwidth constraints. However, science autonomy development has largely focused on bespoke, instrument-specific algorithms; there are no generalized platforms for sharing reusable utilities or centralizing the operation of multiple science autonomy algorithms. To accelerate future development of flight-ready science autonomy, we developed a software package called Science Yield Improvement via Onboard Prioritization and Summary of Information System (SYNOPSIS). SYNOPSIS manages the execution of one or more science autonomy algorithms, stores their outputs in a centralized database, and prioritizes data even across multiple instruments. We simulated the use of SYNOPSIS with science autonomy algorithms to analyze and prioritize data collected by the Curiosity rover on Mars and integrated it into two popular flight software packages — core Flight System (cFS) and F-Prime — to ease infusion of future science autonomy algorithms. As opportunities for onboard data analysis grow with the volume and complexity of data, we expect that platform technologies like SYNOPSIS will facilitate future exploration of our solar system.}
}
  19. Edgett, K. S., Yingst, R. A., Ravine, M. A., Caplinger, M. A., Maki, J. N., Ghaemi, F. T., Schaffner, J. A., Bell, I. J. F., Edwards, L. J., Herkenhoff, K. E., Heydari, E., Kah, L. C., Lemmon, M. T., Minitti, M. E., Olson, T. S., Parker, T. J., Rowland, S. K., Schieber, J., Sullivan, R. J., Sumner, D. Y., Thomas, P. C., Jensen, E. H., Simmonds, J. J., Sengstacken, A. J., Willson, R. G. and Goetz, W. (2012). Curiosity's Mars Hand Lens Imager (MAHLI) Investigation . Space Science Reviews. Source
    BibTeX
    @article{edgett2012curiosity,
  title = {Curiosity's Mars Hand Lens Imager (MAHLI) Investigation},
  author = {Edgett, Kenneth S. and Yingst, R. Aileen and Ravine, Michael A. and Caplinger, Michael A. and Maki, Justin N. and Ghaemi, F. Tony and Schaffner, Jacob A. and Bell, III, James F. and Edwards, Laurence J. and Herkenhoff, Kenneth E. and Heydari, Ezat and Kah, Linda C. and Lemmon, Mark T. and Minitti, Michelle E. and Olson, Timothy S. and Parker, Timothy J. and Rowland, Scott K. and Schieber, Juergen and Sullivan, Robert J. and Sumner, Dawn Y. and Thomas, Peter C. and Jensen, Elsa H. and Simmonds, John J. and Sengstacken, Aaron J. and Willson, Reg G. and Goetz, Walter},
  journal = {Space Science Reviews},
  volume = {170},
  pages = {259--317},
  year = {2012},
  doi = {10.1007/978-1-4614-6339-9_10},
  abstract = {The Mars Science Laboratory (MSL) Mars Hand Lens Imager (MAHLI) investigation will use a 2-megapixel color camera with a focusable macro lens aboard the rover, Curiosity, to investigate the stratigraphy and grain-scale texture, structure, mineralogy, and morphology of geologic materials in northwestern Gale crater. Of particular interest is the stratigraphic record of a ∼5 km thick layered rock sequence exposed on the slopes of Aeolis Mons (also known as Mount Sharp). The instrument consists of three parts, a camera head mounted on the turret at the end of a robotic arm, an electronics and data storage assembly located inside the rover body, and a calibration target mounted on the robotic arm shoulder azimuth actuator housing. MAHLI can acquire in-focus images at working distances from ∼2.1 cm to infinity. At the minimum working distance, image pixel scale is ∼14 μm per pixel and very coarse silt grains can be resolved. At the working distance of the Mars Exploration Rover Microscopic Imager cameras aboard Spirit and Opportunity, MAHLI’s resolution is comparable at ∼30 μm per pixel. Onboard capabilities include autofocus, auto-exposure, sub-framing, video imaging, Bayer pattern color interpolation, lossy and lossless compression, focus merging of up to 8 focus stack images, white light and longwave ultraviolet (365 nm) illumination of nearby subjects, and 8 gigabytes of non-volatile memory data storage.}
}
  20. Herkenhoff, K. E., Squyres, S. W., Bell, J. F., Maki, J. N., Arneson, H. M., Bertelsen, P., Brown, D. I., Collins, S. A., Dingizian, A., Elliott, S. T., Goetz, W. and Hagerott, E. C. (2003). Athena Microscopic Imager Investigation . Journal of Geophysical Research: Planets, E12. Source
    BibTeX
    @article{herkenhoff2003athena,
  title = {{Athena} Microscopic Imager Investigation},
  author = {Herkenhoff, K. E. and Squyres, Steven W. and Bell, J. F. and Maki, Justin N. and Arneson, H. M. and Bertelsen, P. and Brown, D. I. and Collins, S. A. and Dingizian, A. and Elliott, S. T. and Goetz, Walter and Hagerott, E. C.},
  journal = {Journal of Geophysical Research: Planets},
  volume = {108},
  number = {E12},
  pages = {8065},
  year = {2003},
  doi = {10.1029/2003je002076},
  abstract = {The Athena science payload on the Mars Exploration Rovers (MER) includes the Microscopic Imager (MI). The MI is a fixed‐focus camera mounted on the end of an extendable instrument arm, the Instrument Deployment Device (IDD). The MI was designed to acquire images at a spatial resolution of 30 microns/pixel over a broad spectral range (400–700 nm). The MI uses the same electronics design as the other MER cameras but has optics that yield a field of view of 31 × 31 mm across a 1024 × 1024 pixel CCD image. The MI acquires images using only solar or skylight illumination of the target surface. A contact sensor is used to place the MI slightly closer to the target surface than its best focus distance (about 66 mm), allowing concave surfaces to be imaged in good focus. Coarse focusing (∼2 mm precision) is achieved by moving the IDD away from a rock target after the contact sensor has been activated. The MI optics are protected from the Martian environment by a retractable dust cover. The dust cover includes a Kapton window that is tinted orange to restrict the spectral bandpass to 500–700 nm, allowing color information to be obtained by taking images with the dust cover open and closed. MI data will be used to place other MER instrument data in context and to aid in petrologic and geologic interpretations of rocks and soils on Mars.}
}
  21. Maki, J. N., Bell III, J. F., Herkenhoff, K. E., Squyres, S. W., Kiely, A., Klimesh, M., Schwochert, M., Litwin, T., Willson, R., Johnson, A., Maimone, M., Baumgartner, E., Collins, A., Wadsworth, M., Elliot, S. T., Dingizian, A., Brown, D., Hagerott, E. C., Scherr, L., Deen, R., Alexander, D. and Lorre, J. (2003). Mars Exploration Rover Engineering Cameras . Journal of Geophysical Research: Planets, E12. Source
    BibTeX
    @article{maki2003mars,
  title = {Mars Exploration Rover Engineering Cameras},
  author = {Maki, Justin N. and Bell III, J. F. and Herkenhoff, K. E. and Squyres, Steven W. and Kiely, A. and Klimesh, M. and Schwochert, Mark and Litwin, T. and Willson, R. and Johnson, A. and Maimone, M. and Baumgartner, E. and Collins, A. and Wadsworth, M. and Elliot, S. T. and Dingizian, A. and Brown, David and Hagerott, E. C. and Scherr, L. and Deen, R. and Alexander, Doug and Lorre, J.},
  journal = {Journal of Geophysical Research: Planets},
  volume = {108},
  number = {E12},
  pages = {8071},
  year = {2003},
  doi = {10.1029/2003je002077},
  abstract = {NASA's Mars Exploration Rover (MER) Mission will place a total of 20 cameras (10 per rover) onto the surface of Mars in early 2004. Fourteen of the 20 cameras are designated as engineering cameras and will support the operation of the vehicles on the Martian surface. Images returned from the engineering cameras will also be of significant importance to the scientific community for investigative studies of rock and soil morphology. The Navigation cameras (Navcams, two per rover) are a mast‐mounted stereo pair each with a 45° square field of view (FOV) and an angular resolution of 0.82 milliradians per pixel (mrad/pixel). The Hazard Avoidance cameras (Hazcams, four per rover) are a body‐mounted, front‐ and rear‐facing set of stereo pairs, each with a 124° square FOV and an angular resolution of 2.1 mrad/pixel. The Descent camera (one per rover), mounted to the lander, has a 45° square FOV and will return images with spatial resolutions of ∼4 m/pixel. All of the engineering cameras utilize broadband visible filters and 1024 × 1024 pixel detectors.}
}
  22. Pell, B., Sawyer, S. R., Muscettola, N., Smith, B. and Bernard, D. E. (1998). Mission Operations with an Autonomous Agent . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{pell1998mission,
  title = {Mission Operations with an Autonomous Agent},
  author = {Pell, Barney and Sawyer, Scott R. and Muscettola, Nicola and Smith, Benjamin and Bernard, Douglas E.},
  booktitle = {IEEE Aerospace Conference},
  volume = {2},
  pages = {289-313},
  year = {1998},
  doi = {10.1109/aero.1998.687917},
  abstract = {The Remote Agent (RA) is an Artificial Intelligence (AI) system which automates some of the tasks normally reserved for human mission operators and performs these tasks autonomously on-board the spacecraft. These tasks include activity generation, sequencing, spacecraft analysis, and failure recovery. The RA will be demonstrated as a flight experiment on Deep Space One (DS1), the first deep space mission of the NASA's New Millennium Program (NMP). As we moved from prototyping into actual flight code development and teamed with ground operators, we made several major extensions to the RA architecture to address the broader operational context in which RA would be used. These extensions support ground operators and the RA sharing a long-range mission profile with facilities for asynchronous ground updates; support ground operators monitoring and commanding the spacecraft at multiple levels of detail simultaneously; and enable ground operators to provide additional knowledge to the RA, such as parameter updates, model updates, and diagnostic information, without interfering with the activities of the RA or leaving the system in an inconsistent state. The resulting architecture supports incremental autonomy, in which a basic agent can be delivered early and then used in an increasingly autonomous manner over the lifetime of the mission. It also supports variable autonomy, as it enables ground operators to benefit from autonomy when they want it, but does not inhibit them from obtaining a detailed understanding and exercising tighter control when necessary. These issues are critical to the successful development and operation of autonomous spacecraft.}
}
  23. Francis, R., Estlin, T., Doran, G., Johnstone, S., Gaines, D., Verma, V., Burl, M., Frydenvang, J., Montano, S., Wiens, R. C., Schaffer, S., Gasnault, O., DeFlores, L., Blaney, D. and Bornstein, B. (2017). AEGIS autonomous targeting for ChemCam on Mars Science Laboratory: Deployment and results of initial science team use . Science Robotics, 7. Source
    BibTeX
    @article{francis2017aegis,
  title = {AEGIS autonomous targeting for ChemCam on Mars Science Laboratory: Deployment and results of initial science team use},
  author = {Francis, Raymond and Estlin, T. and Doran, Gary and Johnstone, Stephen and Gaines, D. and Verma, V. and Burl, Michael and Frydenvang, Jens and Montano, Suzanne and Wiens, Roger C. and Schaffer, S. and Gasnault, Olivier and DeFlores, Lauren and Blaney, Diana and Bornstein, B.},
  journal = {Science Robotics},
  volume = {2},
  number = {7},
  pages = {eaan4582},
  year = {2017},
  doi = {10.1126/scirobotics.aan4582},
  abstract = {Autonomous target selection and pointing refinement steer a remote geochemical LIBS instrument on Mars, increasing data return.}
}
  24. Chien, S., Castano, R., Bornstein, B., Fukunaga, A., Castano, A., Greeley, R., Whelley, P., Biesiadecki, J. and Lemmon, M. (2008). Results from Automated cloud and dust devil detection onboard the MER . International Symposium on Artificial Intelligence, Robotics and Automation in Space. Source
    BibTeX
    @inproceedings{chien2008results,
  title = {Results from Automated cloud and dust devil detection onboard the MER},
  author = {Chien, Steve and Castano, Rebecca and Bornstein, Benjamin and Fukunaga, Alex and Castano, Andres and Greeley, Ron and Whelley, Patrick and Biesiadecki, Jeffrey and Lemmon, Mark},
  booktitle = {International Symposium on Artificial Intelligence, Robotics and Automation in Space},
  publisher = {JPL Open Repository},
  year = {2008},
  url = {https://dataverse.jpl.nasa.gov/dataset.xhtml?persistentId=hdl:2014/41748}
}
  25. Jones, R. M. (2009). Disruption Tolerant Network Technology Flight Validation Report: DINET . Jet Propulsion Laboratory, California Institute of Technology, JPL Publication 09-2. Source
    BibTeX
    @techreport{jones2009distribution,
  title = {Disruption Tolerant Network Technology Flight Validation Report: DINET},
  author = {Jones, Ross M.},
  number = {JPL Publication 09-2},
  institution = {Jet Propulsion Laboratory, California Institute of Technology},
  year = {2009},
  url = {https://dataverse.jpl.nasa.gov/dataset.xhtml?persistentId=hdl:2014/41477}
}
  26. CCSDS. (2020). CCSDS File Delivery Protocol (CFDP) . Consultative Committee for Space Data Systems, CCSDS 727.0-B-5, Blue Book. Source
    BibTeX
    @techreport{ccsds2020cfdp,
  title = {CCSDS File Delivery Protocol (CFDP)},
  author = {{CCSDS}},
  number = {CCSDS 727.0-B-5, Blue Book},
  institution = {Consultative Committee for Space Data Systems},
  year = {2020},
  url = {https://ccsds.org/Pubs/727x0b5e1.pdf}
}
  27. Makovsky, A., Ilott, P. and Taylor, J. (2009). Mars Science Laboratory Telecommunications System Design . Jet Propulsion Laboratory, California Institute of Technology, Article 14. Source
    BibTeX
    @techreport{makovsky2009mars,
  title = {Mars Science Laboratory Telecommunications System Design},
  author = {Makovsky, Andre and Ilott, Peter and Taylor, Jim},
  series = {DESCANSO Design and Performance Summary Series},
  number = {Article 14},
  institution = {Jet Propulsion Laboratory, California Institute of Technology},
  year = {2009},
  url = {https://descanso.jpl.nasa.gov/DPSummary/Descanso14_MSL_Telecom.pdf}
}
  28. Rosemurgy, P., Gao, J., DeBaun, S., Starch, M., Levison, J. and Castano, R. (2024). Enabling DTN in Spaceflight Systems: Integration of ION with the F Prime Flight Software Framework . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{rosemurgy2024enabling,
  title = {Enabling DTN in Spaceflight Systems: Integration of ION with the F Prime Flight Software Framework},
  author = {Rosemurgy, Paul and Gao, Jay and DeBaun, Sky and Starch, Michael and Levison, Jeffrey and Castano, Rebecca},
  booktitle = {IEEE Aerospace Conference},
  publisher = {JPL Open Repository},
  year = {2024},
  doi = {10.48577/jpl.idifr0}
}
  29. Schlesinger, A., Willman, B. M., Pitts, L., Davidson, S. R. and Pohlchuck, W. A. (2016). Delay/Disruption Tolerant Networking for the International Space Station (ISS) . SpaceOps. Source
    BibTeX
    @inproceedings{schlesinger2016delay,
  title = {Delay/Disruption Tolerant Networking for the International Space Station (ISS)},
  author = {Schlesinger, Adam and Willman, Brett M. and Pitts, Lee and Davidson, Suzanne R. and Pohlchuck, William A.},
  booktitle = {SpaceOps},
  pages = {1-14},
  year = {2016},
  doi = {10.1109/aero.2017.7943857},
  abstract = {Disruption Tolerant Networking (DTN) is an emerging data networking technology designed to abstract the hardware communication layer from the spacecraft/payload computing resources. DTN is specifically designed to operate in environments where link delays and disruptions are common (e.g., space-based networks). The National Aeronautics and Space Administration (NASA) has demonstrated DTN on several missions, such as the Deep Impact Networking (DINET) experiment, the Earth Observing Mission 1 (EO-1) and the Lunar Laser Communication Demonstration (LLCD). To further the maturation of DTN, NASA is implementing DTN protocols on the International Space Station (ISS). This paper explains the architecture of the ISS DTN network, the operational support for the system, the results from integrated ground testing, and the future work for DTN expansion.}
}
  30. CCSDS. (2015). CCSDS Bundle Protocol Specification . Consultative Committee for Space Data Systems, CCSDS 734.2-B-1, Blue Book. Source
    BibTeX
    @techreport{ccsds2015bundle,
  title = {CCSDS Bundle Protocol Specification},
  author = {{CCSDS}},
  number = {CCSDS 734.2-B-1, Blue Book},
  institution = {Consultative Committee for Space Data Systems},
  year = {2015},
  url = {https://ccsds.org/Pubs/734x2b1.pdf}
}
  31. Young, E., Yang, G., Wagner, T., Ridenhour, F., Lawler, C. and Cox, N. (2023). Relay Planning in the Perseverance Rover's First 600 Solar Days on Mars . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{young2023relay,
  title = {Relay Planning in the Perseverance Rover's First 600 Solar Days on Mars},
  author = {Young, Emma and Yang, Genevie and Wagner, Travis and Ridenhour, Flora and Lawler, Christopher and Cox, Nagin},
  booktitle = {IEEE Aerospace Conference},
  publisher = {JPL Open Repository},
  year = {2023},
  doi = {10.48577/jpl.f9zcgz},
  abstract = {Since landing in Jezero Crater on Mars on February 18, 2021, the Mars 2020 mission’s Perseverance rover has been performing daily operations on the Martian surface and has been collecting samples that may one day be returned to Earth. The majority of science and engineering data from the Perseverance rover is returned through the Mars orbiters operated by the National Aeronautics and Space Administration (NASA) and the European Space Agency (ESA) that make up the Mars Relay Network: Mars Reconnaissance Orbiter (MRO), Mars Odyssey (ODY), Mars Atmosphere and Volatile Evolution (MAVEN), and Trace Gas Orbiter (TGO). Prior to the rover’s landing, the Mars 2020 team joined the Mars Relay Network to begin planning relay through the coordinated, multi-mission process that is the cornerstone of relay planning. The Perseverance rover has now been on the Martian surface for more than 600 Martian solar days (“sols”) with several UHF relay sessions planned and executed per sol. The Mars 2020 relay planning team has established and improved upon the recurring process and tool suite to enable both data return and forward link of rover uplink products, and continues to coordinate and negotiate relay asset usages and constraints with the Mars Science Laboratory and Insight relay planning teams. Among many accomplishments, the relay planning team has supported checkouts and commissioning of Low-Density Parity-Check (LDPC) relay link configurations, the mission’s first solar conjunction period, and several rover flight software transitions.The Mars 2020 team has also been performing a checkout and commissioning campaign for the use of bitstream, or “unreliable”, relay sessions. Nominal use of bitstream relay is a new operational capability intended for the Perseverance rover that will allow specific science or engineering activities to run in parallel with the relay session, rather than pausing all other activities during relay, enabling additional and more timely data return. Pending the completion of the checkout and commissioning campaign, the operations team plans to approve and begin regular scheduling and use of bitstream UHF relay sessions beginning in 2023. This paper describes the Mars 2020 relay planning processes and tool architecture, key accomplishments (including progress for the bitstream checkout and commissioning campaign), and lessons learned and ongoing challenges during the first 600 sols of the Mars 2020 surface mission.}
}
  32. Collins, R., Pashai, P., Young, E., Bennett, C., Laubach, S. and Thomas, S. (2019). MSL relay coordination and tactical planning in the era of InSight, MAVEN, and TGO . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{collins2019msl,
  title = {MSL relay coordination and tactical planning in the era of InSight, MAVEN, and TGO},
  author = {Collins, Rachael and Pashai, Pegah and Young, Emma and Bennett, Chris and Laubach, Sharon and Thomas, Steven},
  booktitle = {IEEE Aerospace Conference},
  pages = {1-12},
  publisher = {IEEE},
  year = {2019},
  doi = {10.1109/aero.2019.8741564}
}
  33. Rink, K., Boehmer, R., Kaplan, K., Larsen, R., Clark, J. and Neilson, T. (2023). Martian Mayday: The Evolution of Curiosity's Safe Mode Communication Over Ten Years . IEEE Aerospace Conference. Source
    BibTeX
    @inproceedings{rink2023martian,
  title = {Martian Mayday: The Evolution of Curiosity's Safe Mode Communication Over Ten Years},
  author = {Rink, Kimberly and Boehmer, Rudy and Kaplan, Kyle and Larsen, Reidar and Clark, Jessica and Neilson, Tracy},
  booktitle = {IEEE Aerospace Conference},
  pages = {1-12},
  address = {Big Sky, Montana},
  year = {2023},
  doi = {10.1109/aero58975.2024.10521326},
  abstract = {When the Curiosity Mars rover encounters a serious fault, it enters a configuration called safe mode. Curiosity’s autonomous safe mode behavior must keep the rover communicative and stable until the operations team can triage the issue and send recovery commands. In safe mode, Curiosity follows a preset schedule of windows that specify the antenna, telemetry for transmission, and solar time of each communication opportunity. The timing of these windows, called system fault protection (SFP) windows, is crucial for Curiosity to maintain both communication and power-positivity, or generating more energy that consumed. Not only do SFP windows dictate when Curiosity attempts to communicate over its ultra-high frequency and X-band radios, they also largely control the rover’s power consumption by dictating when its primary computer powers on and off, known as awake and sleep states. Because Curiosity is power-negative when awake, sleep time is necessary to charge its batteries to a higher state of charge than the previous sol, or Martian day.During Curiosity’s prime mission, fault protection engineers regularly updated SFP windows to accommodate changing Mars relay orbits and relative geometry of Earth and Mars. The initial SFP windows at landing directed Curiosity to stay awake for nine Martian hours of communication each sol. This design allowed enough sleep time for Curiosity to increase the state of charge of its batteries by 22% over a single sol in safe mode. But as the mission progressed, updates to SFP windows were additionally driven by Curiosity’s aging power source, which continuously requires increased sleep time to maintain power-positivity. As of 2023, 6.6 hours of safe mode communication per sol will only increase the battery state of charge by 5.5%. Secondary factors that were not present during Curiosity’s prime mission—including changing uplink strategies and Mars relay infrastructure—have introduced new challenges to updating SFP windows. The process of selecting SFP window times has evolved into an iterative analysis that must consider several dynamic operational factors to achieve a balance of communication time and power consumption. By using this analysis to converge on a design that satisfies energy requirements and desired communication opportunities, fault protection engineers ensure Curiosity’s safe mode configuration in its extended mission remains truly safe. This paper discusses how Curiosity uses SFP windows to communicate in safe mode, the factors that influence how SFP windows are designed, and how both the design process and the windows themselves have changed over ten years of surface operations.}
}
  34. Dhamani, N., Johnston, M. D. and Lucena, G. (2022). Scheduling Within a Demand Access Paradigm for NASA's Deep Space Network . International Conference on Automated Planning and Scheduling / SPARK Workshop. Source
    BibTeX
    @inproceedings{dhamani2022scheduling,
  title = {Scheduling Within a Demand Access Paradigm for NASA's Deep Space Network},
  author = {Dhamani, Nihal and Johnston, Mark D. and Lucena, Girly},
  booktitle = {International Conference on Automated Planning and Scheduling / SPARK Workshop},
  publisher = {JPL Open Repository},
  year = {2022},
  doi = {10.48577/jpl.6vt78p}
}
  35. Johnston, M. D. (2020). Scheduling NASA's Deep Space Network: Priorities, Preferences, and Optimization . Journal of Field Robotics. Source
    BibTeX
    @article{johnston2020scheduling,
  title = {Scheduling NASA's Deep Space Network: Priorities, Preferences, and Optimization},
  author = {Johnston, Mark D},
  journal = {Journal of Field Robotics},
  publisher = {JPL Open Repository},
  year = {2020},
  url = {https://dataverse.jpl.nasa.gov/dataset.xhtml?persistentId=hdl:2014/53269}
}
  36. Chien, S., Mclaren, D., Rabideau, G., Mandl, D. and Hengemihle, J. (2011). Scheduling onboard processing for the proposed HyspIRI Mission . Journal of Aerospace Information Systems. Source
    BibTeX
    @article{chien2011scheduling,
  title = {Scheduling onboard processing for the proposed HyspIRI Mission},
  author = {Chien, Steve and Mclaren, David and Rabideau, Gregg and Mandl, Daniel and Hengemihle, Jerry},
  journal = {Journal of Aerospace Information Systems},
  publisher = {JPL Open Repository},
  year = {2011},
  url = {https://dataverse.jpl.nasa.gov/dataset.xhtml?persistentId=hdl:2014/43466}
}
  37. Denniston, C. E., Chang, Y., Reinke, A., Ebadi, K., Sukhatme, G. S., Carlone, L., Morrell, B. and Agha-mohammadi, A.-A. (2022). Loop Closure Prioritization for Efficient and Scalable Multi-Robot SLAM . IEEE International Conference on Robotics and Automation. Source
    BibTeX
    @inproceedings{denniston2022loop,
  title = {Loop Closure Prioritization for Efficient and Scalable Multi-Robot SLAM},
  author = {Denniston, Christopher E. and Chang, Yun and Reinke, Andrzej and Ebadi, Kamak and Sukhatme, Gaurav S. and Carlone, Luca and Morrell, Benjamin and Agha-mohammadi, Ali-akbar},
  booktitle = {IEEE International Conference on Robotics and Automation},
  year = {2022},
  doi = {10.48577/jpl.9emoms},
  abstract = {Multi-robot SLAM systems in GPS-denied environments require loop closures to maintain a drift-free centralized map. With increasing number of robots and size of the environment, checking and computing the transformation for all the loop closure candidates becomes computationally infeasible. In this work, we describe a loop closure module that is able to prioritize which loop closures to compute based on the underlying pose graph, the proximity to known beacons, and the characteristics of the point clouds. We validate this system in the context of the DARPA Subterranean Challenge and on numerous challenging underground datasets and demonstrate the ability of this system to generate and maintain a map with low error. We find that our proposed techniques are able to select effective loop closures which results in 51% mean reduction in median error when compared to an odometric solution and 75% mean reduction in median error when compared to a baseline version of this system with no prioritization.}
}