Active Neural Localization

By Devendra Singh Chaplot, Emilio Parisotto, Ruslan Salakhutdinov. 2018

90 citations recorded by the source corpus.

Localization is the problem of estimating the location of an autonomous agent from an observation and a map of the environment. Traditional methods of localization, which filter the belief based on the observations, are sub-optimal in the number of steps required, as they do not decide the actions taken by the agent. We propose "Active Neural Localizer", a fully differentiable neural network that learns to localize accurately and efficiently. The proposed model incorporates ideas of traditional filtering-based localization methods, by using a structured belief of the state with multiplicative interactions to propagate belief, and combines it with a policy model to localize accurately while minimizing the number of steps required for localization. Active Neural Localizer is trained end-to-end with reinforcement learning. We use a variety of simulation environments for our experiments which include random 2D mazes, random mazes in the Doom game engine and a photo-realistic environment in the Unreal game engine. The results on the 2D environments show the effectiveness of the learned policy in an idealistic setting while results on the 3D environments demonstrate the model's capability of learning the policy and perceptual model jointly from raw-pixel based RGB observations. We also show that a model trained on random textures in the Doom environment generalizes well to a photo-realistic office space environment in the Unreal engine.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation (2016, neutral): “Chaplot et al. (2018) introduce a network structure mimicking the Bayesian ﬁltering process with a specially designed perception model.”
Cognitive Mapping and Planning for Visual Navigation (2017, background): “to the real world [12,64], and how the domain gap between simulation and the real world may be reduced [76]. A number of works have studied related navigation problems in such simulation environments [6,15,36]. Researchers have also gone beyond specifying goals as a desired location in space and ﬁnd”
Scene Memory Transformer for Embodied Agents in Long-Horizon Tasks (2019, background): “They are applied to visual navigation [15, 16, 32, 47], interactive question answering [13], and localization [7, 18, 20].”
Neural Topological SLAM for Visual Navigation (2020, background): “Researchers have also tackled the problem of passive and active localization [7, 26], in order to aid building such consistent metric representations.”
Semantic Curiosity for Active Visual Learning (2020, unknown): “on the task by gaining more information about the environment. This has been instantiated in the context of object detection [1], amodal object detection [56], scene completion [29], and localization [11,22]. We consider the problem in a dierent setting and study how to eciently move around to best ”
Talk the Walk: Navigating New York City through Grounded Dialogue (2018, background): “…AI, many of which involve agents following natural language instructions with respect to an environment(Artzi & Zettlemoyer, 2013; Yu et al., 2017; Hermann et al., 2017; Mei et al., 2016; Chaplot et al., 2018b;a), following-up on early work in this area (MacMahon et al., 2006; Chen & Mooney, 2011).”
Benchmarking Classic and Learned Navigation in Complex 3D Environments (2019, methodology): “…have been largely incompatible in Task No GT map No GT pose Continuous state Locomotion DPF, PF-net [22, 24] Localization no yes yes no ANL [10] Localization no yes no yes MapNet [18] Mapping yes yes no no EgoSM [46] Mapping yes no yes no VIN, GPPN [39, 29] Planning no no no yes QMDP-Net,…”
Global Pose Estimation with an Attention-Based Recurrent Network (2018, methodology): “They have been successfully applied to several related multiview vision problems, including learning optical flow [11], depth [28], homography between frame pairs [9], and localization [5] and re-localization problems.”
Learning to Visually Navigate in Photorealistic Environments Without any Supervision (2020, unknown): “For our CNN module, we use a network with 3 convolutional layers with kernels of size [9 , 7 , 5], strides of size [5 , 4 , 3] and number of channels [32 , 64 , 128].”
Object Goal Navigation using Goal-Oriented Semantic Exploration (2020, methodology): “These include methods which use recurrent neural networks [30, 25, 6, 37, 21, 7, 38, 42], structured spatial representations [16, 33, 8, 19, 15] and topological representations [35, 36].”
Learning to Explore using Active Neural SLAM (2020, background): “Some works limit agent motion on a grid with 90 degree rotations (Zhu et al., 2017; Gupta et al., 2017; Chaplot et al., 2018).”
Learning to Navigate in Cities Without a Map (2018, methodology): “Deep RL is used for active localisation in (Chaplot et al., 2018).”
A Survey on Active Simultaneous Localization and Mapping: State of the Art and New Frontiers (2022, background): “This strategy is typical in deep reinforcement learning (DRL) approaches [121], [122], [123], for which local optimality is alleviated by network memorization.”