Lunar Analogue Dataset for Traversability Assessment and Novelty Detection

Stefanuk, Braden W.; Battler, M.; Raimalwala, K.; Deslauriers, A.; Faragalli, M.; Skonieczny, Krzysztof; Smal, E.; Reid, E.

doi:10.2514/6.2021-4210

Cited by 2 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Lunar Analogue dataset developed and presented in (Stefanuk et al, 2021) was also used in this research. In contrast to the Curiosity Mastcam dataset, which is already curated into small selected patches, this Lunar Analogue dataset starts with highresolution images of whole landscapes.…”

Section: Lunar Analoguementioning

confidence: 99%

See 1 more Smart Citation

Novelty detection in rover-based planetary surface images using autoencoders

Stefanuk

Skonieczny²

2022

Front. Robot. AI

View full text Add to dashboard Cite

In the domain of planetary science, novelty detection is gaining attention because of the operational opportunities it offers, including annotated data products and downlink prioritization. Using a variational autoencoder (VAE), this work improves upon state-of-the-art novelty detection performance in the context of Martian exploration by >7% (measured by the area under the receiver operating characteristic curve (ROC AUC)). Autoencoders, especially VAEs, perform well across all classes of novelties defined for Martian exploration. VAEs are shown to have high recall in the Martian context, making them particularly useful for on-ground processing. Convolutional autoencoders (CAEs), on the other hand, demonstrate high precision making them good candidates for onboard downlink prioritization. In our implementation adversarial autoencoders (AAEs) are also shown to perform on par with state-of-the-art. Dimensionality reduction is a key feature of autoencoders for novelty detection. In this study the impact of dimensionality reduction on detection quality is explored, showing that both VAEs and AAEs achieve comparable ROC AUCs to CAEs despite observably poorer (blurred) image reconstructions; this is observed both in Martian data and in lunar analogue data.

show abstract

Section: Lunar Analoguementioning

confidence: 99%

“…There also exist cases of pyroclastic deposits that appear following fountainstyle volcanic eruptions, they have been documented to exhibit green and orange hues NASA (2017). With consultation and validation from lunar geologists (Stefanuk et al, 2021), Table 1 outlines the novelty labels used in the dataset.…”

Section: Lunar Analoguementioning

confidence: 99%

Novelty detection in rover-based planetary surface images using autoencoders

Stefanuk

Skonieczny²

2022

Front. Robot. AI

View full text Add to dashboard Cite

show abstract

“…Non-reinforcement learning cases are numerous, such as anomaly detection to try to detect areas of high interest via unsupervised learning [17], or classifying terrain via supervised learning [13]. AI has also been used to help with slip prediction for avoiding entrapment [9].…”

Section: Space Roboticsmentioning

confidence: 99%

SegVisRL: Visuomotor Development for a Lunar Rover for Hazard Avoidance using Camera Images

Blum,

Paillet,

Masawat

et al. 2021

Preprint

View full text Add to dashboard Cite

The visuomotor system of any animal is critical for its survival, and the development of a complex one within humans is large factor in our success as a species on Earth. This system is an essential part of our ability to adapt to our environment. We use this system continuously throughout the day, when picking something up, or walking around while avoiding bumping into objects. Equipping robots with such capabilities will help produce more intelligent locomotion with the ability to more easily understand their surroundings and to move safely. In particular, such capabilities are desirable for traversing the lunar surface, as it is full of hazardous obstacles, such as rocks. These obstacles need to be identified and avoided in real time. This paper seeks to demonstrate the development of a visuomotor system within a robot for navigation and obstacle avoidance, with complex rock shaped objects representing hazards. Our approach uses deep reinforcement learning with only image data. In this paper, we compare the results from several neural network architectures and a preprocessing methodology which includes producing a segmented image and downsampling.

show abstract

Lunar Analogue Dataset for Traversability Assessment and Novelty Detection

Cited by 2 publications

References 0 publications

Novelty detection in rover-based planetary surface images using autoencoders

Novelty detection in rover-based planetary surface images using autoencoders

SegVisRL: Visuomotor Development for a Lunar Rover for Hazard Avoidance using Camera Images

Contact Info

Product

Resources

About