Safe Visual Navigation via Deep Learning and Novelty Detection
Open accessThe DOI matches its Crossref record; 10 of the 10 references it cites were found in a registry or an index.
Abstract
Robots that use learned perceptual models in the real world must be able to safely handle cases where they are forced to make decisions in scenarios that are unlike any of their training examples.However, state-of-the-art deep learning methods are known to produce erratic or unsafe predictions when faced with novel inputs.Furthermore, recent ensemble, bootstrap and dropout methods for quantifying neural network uncertainty may not efficiently provide accurate uncertainty estimates when queried with inputs that are very different from their training data.Rather than unconditionally trusting the predictions of a neural network for unpredictable real-world data, we use an autoencoder to recognize when a query is novel, and revert to a safe prior behavior.With this capability, we can deploy an autonomous deep learning system in arbitrary environments, without concern for whether it has received the appropriate training.We demonstrate our method with a vision-guided robot that can leverage its deep neural network to navigate 50% faster than a safe baseline policy in familiar types of environments, while reverting to the prior behavior in novel environments so that it can safely collect additional training data and continually improve.A video illustrating our approach is available at: http://groups.csail.mit.edu/rrg/videos/safe visual navigation.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Dropout as a bayesian approximation: Representing model uncertainty in deep learning
Gal, Y., & Ghahramani, Z. (2015) · arXiv · doi:10.48550/arxiv.1506.02142 · Cited by 6,471
Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
Lakshminarayanan, B., Pritzel, A., & Blundell, C. (2016) · arXiv · doi:10.48550/arxiv.1612.01474 · Cited by 3,666 · cited 1×
“Recognizing this limitation, some work has attempted to quantify uncertainty with Bayesian neural networks [20, 5], by training ensembles of networks on re-sampled versions of the training dataset [26, 17, 8, 18], or by using dropout to train an implicit family of models [9].”
BackgroundEnd to End Learning for Self-Driving Cars
Bojarski, M., Del Testa, D., Dworakowski, D., Firner, B., Flepp, B., Goyal, P., Jackel, L. D., Monfort, M., Muller, U., Zhang, J., et al. (2016) · arXiv · doi:10.48550/arxiv.1604.07316 · Cited by 3,198
Stanley: The robot that won the DARPA Grand Challenge
Thrun, S., Montemerlo, M., Dahlkamp, H., Stavens, D., Aron, A., Diebel, J., Fong, P., Gale, J., Halpenny, M., Hoffmann, G., et al. (2006) · Journal of Field Robotics · doi:10.1002/rob.20147 · Cited by 2,237 · cited 1×
“image-based guidance can significantly extend the effective perceptual range and improve navigation performance [13, 34].”
BackgroundWeight Uncertainty in Neural Networks
Blundell, C., Cornebise, J., Kavukcuoglu, K., & Wierstra, D. (2015) · arXiv · doi:10.48550/arxiv.1505.05424 · Cited by 1,342
ALVINN: An autonomous land vehicle in a neural network
Pomerleau, D. A. (2018) · Carnegie Mellon University · doi:10.1184/r1/6603146.v1 · Cited by 1,340
Mackay, D. J. C. (1995) · Network Computation in Neural Systems · doi:10.1088/0954-898x/6/3/011 · Cited by 915
A Machine Learning Approach to Visual Perception of Forest Trails for Mobile Robots
Giusti, A., Guzzi, J., Ciresan, D. C., He, F.-L., Rodriguez, J. P., Fontana, F., Faessler, M., Forster, C., Schmidhuber, J., Caro, G. D., et al. (2016) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2015.2509024 · Cited by 665
CAD2RL: Real Single-Image Flight Without a Single Real Image
Sadeghi, F., & Levine, S. (2017) · Robotics: Science and Systems XIII · doi:10.15607/rss.2017.xiii.034 · Cited by 618
Cited by
A survey of uncertainty in deep neural networks
Gawlikowski, J., Tassi, C. R. N., Ali, M., Lee, J., Humt, M., Feng, J., Kruspe, A., Triebel, R., Jung, P., Roscher, R., et al. (2023) · Artificial Intelligence Review · doi:10.1007/s10462-023-10562-9 · Cited by 1,406
Planning and Decision-Making for Autonomous Vehicles
Schwarting, W., Alonso-Mora, J., & Rus, D. (2018) · Annual Review of Control, Robotics, and Autonomous Systems · doi:10.1146/annurev-control-060117-105157 · Cited by 972
A Unifying Review of Deep and Shallow Anomaly Detection
Ruff, L., Kauffmann, J. R., Vandermeulen, R. A., Montavon, G., Samek, W., Kloft, M., Dietterich, T. G., & Muller, K.-R. (2021) · Proceedings of the IEEE · doi:10.1109/jproc.2021.3052449 · Cited by 914
DroNet: Learning to Fly by Driving
Loquercio, A., Maqueda, A. I., del-Blanco, C. R., & Scaramuzza, D. (2018) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2018.2795643 · Cited by 514
A Survey on Active Simultaneous Localization and Mapping: State of the Art and New Frontiers
Placed, J. A., Strader, J., Carrillo, H., Atanasov, N., Indelman, V., Carlone, L., & Castellanos, J. A. (2023) · IEEE Transactions on Robotics · doi:10.1109/tro.2023.3248510 · Cited by 332
BADGR: An Autonomous Self-Supervised Learning-Based Navigation System
Kahn, G., Abbeel, P., & Levine, S. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3057023 · Cited by 284
Deep Drone Racing: From Simulation to Reality With Domain Randomization
Loquercio, A., Kaufmann, E., Ranftl, R., Dosovitskiy, A., Koltun, V., & Scaramuzza, D. (2020) · IEEE Transactions on Robotics · doi:10.1109/tro.2019.2942989 · Cited by 272
Motion planning and control for mobile robot navigation using machine learning: a survey
Xiao, X., Liu, B., Warnell, G., & Stone, P. (2022) · Autonomous Robots · doi:10.1007/s10514-022-10039-8 · Cited by 254
Are We Ready for Autonomous Drone Racing? The UZH-FPV Drone Racing Dataset
Delmerico, J., Cieslewski, T., Rebecq, H., Faessler, M., & Scaramuzza, D. (2019) · 2019 International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra.2019.8793887 · Cited by 235
“More generally, high-speed robot navigation in cluttered, unknown environments is currently a very active research area [7], [8], [9], [10], [11], [12], [13] and funding of over 50 million US dollars has been made available through the DARPA”
BackgroundRoboTHOR: An Open Simulation-to-Real Embodied AI Platform
Deitke, M., Han, W., Herrasti, A., Kembhavi, A., Kolve, E., Mottaghi, R., Salvador, J., Schwenk, D., VanderBilt, E., Wallingford, M., et al. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00323 · Cited by 220
“Spatial navigation approaches [62, 17, 64, 23, 25, 56, 79, 49, 30, 14, 10] typically address navigating towards a pre-specified coordinate or a frame of a scene and they focus on understanding the geometry of the scene and learning better exploration strategies.”
BackgroundSemantic anomaly detection with large language models
Elhafsi, A., Sinha, R., Agia, C., Schmerling, E., Nesnas, I. A. D., & Pavone, M. (2023) · Autonomous Robots · doi:10.1007/s10514-023-10132-6 · Cited by 101
“While these existing methods for OOD/anomaly detection haveprovedusefulinincreasingthereliabilityofdownstream decision making (McAllister et al., 2019; Richter & Roy, 2017), their application comes with two limitations.”
MethodModeling Quality of IoT Experience in Autonomous Vehicles
Minovski, D., Ahlund, C., & Mitra, K. (2020) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.2975418 · Cited by 59
“D lidar data [95], and visual navigation using deep learning with autoencoding [96].”
MethodTowards Target-Driven Visual Navigation in Indoor Scenes via Generative Imitation Learning
Wu, Q., Gong, X., Xu, K., Manocha, D., Dong, J., & Wang, J. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3036597 · Cited by 47
A Smooth Representation of Belief over SO(3) for Deep Rotation Learning with Uncertainty
Peretroukhin, V., Giamou, M., Nicholas Greene, W., Rosen, D., Kelly, J., & Roy, N. (2020) · Robotics: Science and Systems XVI · doi:10.15607/rss.2020.xvi.007 · Cited by 41
Self-citationDetect, Reject, Correct: Crossmodal Compensation of Corrupted Sensors
Lee, M. A., Tan, M., Zhu, Y., & Bohg, J. (2021) · 2021 IEEE International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra48506.2021.9561847 · Cited by 24
“Similar to [31], we use reconstruction error (L2 distance between input and its reconstruction) to detect input that is out-of-distribution and therefore deemed to come from a corrupted sensor.”
BackgroundMethodInfluentialError-Aware Imitation Learning from Teleoperation Data for Mobile Manipulation
Wong, J., Tung, A., Kurenkov, A., Mandlekar, A., Fei-Fei, L., Savarese, S., & Martín-Martín, R. (2021) · arXiv · doi:10.48550/arxiv.2112.05251 · Cited by 19
“Most relevant to our work, various approaches have been proposed for detecting OOD states using deep neural network-based architectures, such as direct training of an error prediction mode [49], uncertainty estimation for the policy [50, 50, 51, 52], or computing a reconstruction error for the input [53, 54, 55, 56].”
Evaluating the Use of Reconstruction Error for Novelty Localization
Feeney, P., & Hughes, M. C. (2021) · arXiv · doi:10.48550/arxiv.2107.13379 · Cited by 1
“This threshold may be set using any method for hyperparameter selection, but thresholding at a high percentile of the empirical cumulative distribution function for the reconstruction error on the training data provides a low-cost solution (Richter & Roy, 2017).”
MethodInfluential
Metadata
- Authors
- Richter, C., & Roy, N.
- Venue
- Robotics: Science and Systems XIII
- Published
- 2017-07-12
- OpenAlex
- W2734960757
- Cited by
- 157
- References
- 10
- FWCI
- 9.53
- Type
- conference paper
- Language
- en
- Open access
- gold
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.