Efficient and autonomous processing and classification of images on small spacecraft

Gillette, Antony; Wilson, C. W.; George, Alan

doi:10.1109/naecon.2017.8268758

Cited by 6 publications

(4 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, data compression is already deployed in many missions (e.g., in [22] a 2:1 compression is employed) because it mitigates the bottleneck of the downlink. The efficiency of the downlink can be increased even further, removing useless data instead of sending it to the ground (i.e., data removal [23]). For instance, in the Landsat datasets [24], the average cloud cover in an archived scene is 34%, with 38% of the scenes containing less than 10% cloud cover.…”

Section: Benefits Of Data Removal and Compressionmentioning

confidence: 99%

On-Board Decision Making in Space with Deep Neural Networks and RISC-V Vector Processors

Mascio¹,

Menicucci²,

Gill³

et al. 2021

Journal of Aerospace Information Systems

View full text Add to dashboard Cite

The use of deep neural networks (DNNs) in terrestrial applications went from niche to widespread in a few years, thanks to relatively inexpensive hardware for both training and inference, and large datasets available. The applicability of this paradigm to space systems, where both large datasets and inexpensive hardware are not readily available, is more difficult and thus still rare. This paper analyzes the impact of DNNs on the system-level capabilities of space systems in terms of on-board decision making (OBDM) and identifies the specific criticalities of deploying DNNs on satellites. The workload of DNNs for on-board image and telemetry analysis is analyzed, and the results are used to drive the preliminary design of a RISC-V vector processor to be employed as a generic platform to enable energy-efficient OBDM for both payload and platform applications. The design of the memory subsystem is carried out in detail to allow full exploitation of the computational resources in typically resource-constrained space systems.

show abstract

Section: Benefits Of Data Removal and Compressionmentioning

confidence: 99%

On-Board Decision Making in Space with Deep Neural Networks and RISC-V Vector Processors

Mascio¹,

Menicucci²,

Gill³

et al. 2021

Journal of Aerospace Information Systems

View full text Add to dashboard Cite

show abstract

“…On-board decision making helps overcome typical bottlenecks of space systems like the downlink bandwidth. Link bandwidth is a problem especially for severely resource-limited spacecraft like smallsats (e.g., it can take hours to days to downlink an image from a CubeSat to the ground) [54]. The capability of filtering out images that do not meet certain criteria (data reduction) can greatly increase the potentiality of CubeSats, as improvements in sensors generate increasingly larger data volumes and storage space on CubeSats is typically very limited [54].…”

Section: Processors To Enable On-board Decision Makingmentioning

confidence: 99%

“…Link bandwidth is a problem especially for severely resource-limited spacecraft like smallsats (e.g., it can take hours to days to downlink an image from a CubeSat to the ground) [54]. The capability of filtering out images that do not meet certain criteria (data reduction) can greatly increase the potentiality of CubeSats, as improvements in sensors generate increasingly larger data volumes and storage space on CubeSats is typically very limited [54]. Reference [55] describes the validation on a CubeSat of intelligent algorithms to discard images with no information and other autonomous capabilities (e.g., computer vision).…”

Section: Processors To Enable On-board Decision Makingmentioning

confidence: 99%

Leveraging the Openness and Modularity of RISC-V in Space

Mascio¹,

Menicucci²,

Gill³

et al. 2019

Journal of Aerospace Information Systems

View full text Add to dashboard Cite

This paper proposes a roadmap to address present and future needs in space systems with RISC-V processors. RISC-V is an open and modular instruction set architecture, which is rapidly growing in popularity in terrestrial applications. To satisfy different applications with contrasting requirements in satellite data systems, four different types of processors are identified: 1) low-area/low-power microcontrollers, 2) on-board computers, 3) general-purpose processors for payloads, and 4) enhanced payload processors for artificial intelligence. Several solutions based on RISC-V are proposed for each of these types of processors and compared with proprietary commercial-off-the-shelf and spacegrade solutions. An extensive analysis of the results available from literature is conducted to show that RISC-V has the potential to solve such a wide range of needs. This paper will also show the unprecedented number of open-source implementations and models that were developed in a relative short time on a single instruction set architecture. Future space systems could benefit from many of those developments, and this work identifies and highlights what is still missing to satisfy the specific needs of processors for space, especially in terms of fault tolerance and technology readiness level.

show abstract

“…Using SHREC's image processing library [19], images can be classified and compressed before downlink which can greatly improve the expected science return of the mission. For example, images can be classified by the Color Classifier app prior to downlink which uses RGB thresholds to estimate a percentage of clouds, water, land, and dark in a terrestrial-scene image, and images with a high land percentage can be prioritized in the downlink queue.…”

Section: Mission Management On Stp-h6 (Ssivp)mentioning

confidence: 99%

Spacecraft mission agent for autonomous robust task execution

Gillette

O’Connor

Wilson

et al. 2018

2018 IEEE Aerospace Conference

Self Cite

View full text Add to dashboard Cite

Autonomy in space systems can drastically reduce the workload of ground crews for satellite missions, especially for clusters of satellites. Additionally, autonomy can increase the efficiency of missions by maximizing the utilization of resources and rapidly handling any issues that arise without having to wait for instructions from the ground. This research presents an agentbased, task-execution approach to onboard spacecraft autonomy. Instead of the traditional approach requiring onboard planning and scheduling, this method uses a combination of constraint and priority parameters associated with every task to ensure robust task execution with behavior as intended. Using this method, tasks will only run under safe conditions (e.g. no conflict with any running tasks), which enables conflicting tasks to be scheduled closer together or even overlapping for lower-priority tasks. This approach manages the execution of tasks on the timescale of seconds, enabling conflicting tasks to run sequentially, thereby increasing productivity if earlier tasks finish ahead of schedule. This framework leverages the NASA-developed, open-source projects cFE and PLEXIL and was tested on development boards comparable to flight hardware.

show abstract

Efficient and autonomous processing and classification of images on small spacecraft

Cited by 6 publications

References 5 publications

On-Board Decision Making in Space with Deep Neural Networks and RISC-V Vector Processors

On-Board Decision Making in Space with Deep Neural Networks and RISC-V Vector Processors

Leveraging the Openness and Modularity of RISC-V in Space

Spacecraft mission agent for autonomous robust task execution

Contact Info

Product

Resources

About