Neural Topological SLAM for Visual Navigation

Chaplot, Devendra Singh; Salakhutdinov, Ruslan; Gupta, Abhinav; Gupta, Saurabh

doi:10.1109/cvpr42600.2020.01289

Cited by 161 publications

(182 citation statements)

References 35 publications

Supporting

Mentioning

182

Contrasting

Order By: Relevance

“…Oculus Insight, by Facebook, uses visual-inertial SLAM (simultaneous localization and mapping) to generate real-time maps and position tracking. 48 More sophisticated approaches, such as Neural Topological SLAM, leverage semantics and geometric information to improve long-horizon navigation (Chaplot et al 2020). Combining audio and visual sensors can further improve navigation of egocentric observations in complex 3D environments, which can be done through deep reinforcement learning approach (Chen et al 2020).…”

Section: Augmented Virtual and Mixed Reality (Vr Ar Mr)mentioning

confidence: 99%

Artificial intelligence in the creative industries: a review

2021

View full text Add to dashboard Cite

This paper reviews the current state of the art in artificial intelligence (AI) technologies and applications in the context of the creative industries. A brief background of AI, and specifically machine learning (ML) algorithms, is provided including convolutional neural networks (CNNs), generative adversarial networks (GANs), recurrent neural networks (RNNs) and deep Reinforcement Learning (DRL). We categorize creative applications into five groups, related to how AI technologies are used: (i) content creation, (ii) information analysis, (iii) content enhancement and post production workflows, (iv) information extraction and enhancement, and (v) data compression. We critically examine the successes and limitations of this rapidly advancing technology in each of these areas. We further differentiate between the use of AI as a creative tool and its potential as a creator in its own right. We foresee that, in the near future, ML-based AI will be adopted widely as a tool or collaborative assistant for creativity. In contrast, we observe that the successes of ML in domains with fewer constraints, where AI is the ‘creator’, remain modest. The potential of AI (or its developers) to win awards for its original creations in competition with human creatives is also limited, based on contemporary technologies. We therefore conclude that, in the context of creative industries, maximum benefit from AI will be derived where its focus is human-centric—where it is designed to augment, rather than replace, human creativity.

show abstract

Section: Augmented Virtual and Mixed Reality (Vr Ar Mr)mentioning

confidence: 99%

Artificial intelligence in the creative industries: a review

2021

View full text Add to dashboard Cite

show abstract

“…However, local matching localization using image level is still not sufficiently robust when the environment changes. Moreover, due to the widespread successful application of deep learning, the learning-based approach has also attracted a lot of interest [12] [13]. It relies on a large number of labeled datasets and cannot be applied well to an unfamiliar environment.…”

Section: Related Workmentioning

confidence: 99%

Lightweight Object-level Topological Semantic Mapping and Long-term Global Localization based on Graph Matching

Wang¹,

Zhang²,

Tang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Mapping and localization are two essential tasks for mobile robots in real-world applications. However, largescale and dynamic scenes challenge the accuracy and robustness of most current mature solutions. This situation becomes even worse when computational resources are limited. In this paper, we present a novel lightweight object-level mapping and localization method with high accuracy and robustness. Different from previous methods, our method does not need a prior constructed precise geometric map, which greatly releases the storage burden, especially for large-scale navigation. We use object-level features with both semantic and geometric information to model landmarks in the environment. Particularly, a learning topological primitive is first proposed to efficiently obtain and organize the object-level landmarks. On the basis of this, we use a robot-centric mapping framework to represent the environment as a semantic topology graph and relax the burden of maintaining global consistency at the same time. Besides, a hierarchical memory management mechanism is introduced to improve the efficiency of online mapping with limited computational resources. Based on the proposed map, the robust localization is achieved by constructing a novel local semantic scene graph descriptor, and performing multi-constraint graph matching to compare scene similarity. Finally, we test our method on a low-cost embedded platform to demonstrate its advantages. Experimental results on a large scale and multi-session real-world environment show that the proposed method outperforms the state of arts in terms of lightweight and robustness.

show abstract

“…Planning and navigation in unstructured environments can be greatly improved by learning environment traversability using prior experience, but previous approaches to explicitly representing environment traversability require expensive human supervision or traversability heuristics [22][23][24][25]. Recent progress in using topological graphs to implicitly reason about traversability [18,[26][27][28] gives a promising way to learn from prior experience but has not been demonstrated for long-range navigation in complex, unstructured environments. In contrast, our hybrid approach employs a form of self-supervised learningbased explicit trajectory traversability estimation in tandem with geometry-based positional traversability estimation, which enables our approach to more robustly deal with previouslyunseen complex obstacles and terrains.…”

Section: Instantiating the Systemmentioning

confidence: 99%

Hybrid Imitative Planning with Geometric and Predictive Costs in Off-road Environments

Dashora,

Shin,

Shah

et al. 2021

Preprint

View full text Add to dashboard Cite

Fig. 1: Our approach combines a learning-based method and a geometry-based planner to plan obstacle-free trajectories for a Husky robot in a deployment area with obstacles (trees, grass, bushes) as seen in (a). An example navigation task is shown in (b), where the robot (marked in red) must navigate to the goal (marked in green) while avoiding obstacles on its path. On its path to the goal, the robot often encounters a scenes like (c) where there are non-traversable elements like shrubs (highlighted in cyan) and novel obstacles like a wall. Learning-based methods successfully avoid previously seen obstacles but can struggle with novel obstacles, leading to a collision with the wall (d). While a geometric approach can avoid this by predicting the collision, it is unable to reason about the traversability of large shrubs that are hard to identify (e). Note that the cyan markers are shown for illustration only and are not available to the robot. Our approach, HIP, combines attributes from both these methods and successfully plans a collision-free trajectory (f). Videos of our results are hosted at sites.google.com/view/hybrid-imitative-planning/.

show abstract

Neural Topological SLAM for Visual Navigation

Cited by 161 publications

References 35 publications

Artificial intelligence in the creative industries: a review

Artificial intelligence in the creative industries: a review

Lightweight Object-level Topological Semantic Mapping and Long-term Global Localization based on Graph Matching

Hybrid Imitative Planning with Geometric and Predictive Costs in Off-road Environments

Contact Info

Product

Resources

About