Probabilistic robotics

By Sebastian Thrun. 2002

7,585 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Artificial Intelligence, Computer Vision and Pattern Recognition, Reinforcement Learning

How researchers see this paper

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

In Aerospace Engineering

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Reinforcement Learning

Citations in the corpus

Memory-based control with recurrent neural networks (2015, background): “A large amount of research has focused on how to efficiently integrate noisy sensory information over multiple timesteps in order to derive accurate estimates of the system state, or to estimate derivatives of important properties of the system [27].”
Embodied Question Answering (2017, background): “input that will maximize answer conﬁdence. Visual Navigation: Vision + Action. The problem of navigating in an environment based on visual perception has long been studied in vision and robotics (see [16] for an extensive survey). Classical techniques divide navigation into two distinct phases – map”
A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation (2016, neutral): “Zhang et al. (2017b) proposed Neural SLAM , based on the Neural Map proposed by Parisotto and Salakhutdinov (2017), where the Neural Turing Machine (NTM) is deployed for the DRL agent to interact with.”
Filtering with State-Observation Examples via Kernel Monte Carlo Filter (2013, methodology): “training data, since any parametric model for the problem may cause signiﬁcant bias errors. On the other hand, simulat ion with the statetransition model is possible, as this represents robot motion (Thrun et al., 2005). The part-nonparametric approach is also effective for the signal-strength local”
The limits and potentials of deep learning for robotics (2018, methodology): “This would allow robots to treat a deep neural network like any other sensor, and use the established Bayesian techniques [53], [58], [117] to fuse the network’s predictions with prior knowledge or other sensor measurements, or to accumulate information over time.”
Self-Supervised Deep Reinforcement Learning with Generalized Computation Graphs for Robot Navigation (2017, background): “d vehicles [1], [2]. Many of these prior works operate under the paradigm of maintaining a map while simultaneously localizing and planning actions. While simultaneous localization and mapping (SLAM) [11] and planning approaches have had difﬁculties with the size, weight, and power (SWaP) constraint”
BADGR: An Autonomous Self-Supervised Learning-Based Navigation System (2020, background): “this map [1].”
Learning to Drive in a Day (2018, background): “ control algorithms [11], [12], [13]. These systems are traditionally composed of many speciﬁc independently engineered components, such as perception, state estimation, mapping, planning and control [14]. However, because each component needs to be individually speciﬁed and tuned, this can be difﬁc”
Cognitive Mapping and Planning for Visual Navigation (2017, methodology): “A comprehensive survey of classical approaches for mapping and planning can be found in [57].”
Building Generalizable Agents with a Realistic and Rich 3D Environment (2018, methodology): “ navigation in real 3D scenes (Leonard & Durrant-Whyte,1992). Classical approaches decompose the task into two subtasks by building a 3D map of the scene using SLAM and then planning in this map (Fox et al., 2005). More recently, end-to-end learning methods were introduced to predict robotic act”
Semi-parametric Topological Memory for Navigation (2018, background): “In contrast, navigation systems developed in robotics are typically based on metric maps, constructed using the available sensory information – sonar, LIDAR, RGB-D, or RGB input (Elfes, 1987; Thrun et al., 2005; Durrant-Whyte & Bailey, 2006).”
Robots that can adapt like animals (2014, background): “Store the performance of c (cid:48) in the behavior-performance map according to its behavioral descriptor x (cid:48) .”
Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016, neutral): “hallenges connected 2 to efﬁciency and robust data association. Two other excellent references describing the three main SLAM formulations of the classical age are the book of Thrun, Burgard, and Fox [240] andthe chapter of Stachniss et al. [234, Ch. 46]. The subsequent period is what we call the al”
Deep reinforcement learning with successor features for navigation across similar environments (2016, background): “h planning and control, all of which often depend on additional information given to the agent. Although individually the problems of SLAM, localization, path planning and control are well understood [1], [2], [3], they have mainly been treated as separable problems within robotics and often require”
Venture: a higher-order probabilistic programming platform with programmable inference (2014, background): “, 1995) to robotics (Thrun et al., 2005), artificial intelligence (Russell and Norvig, 2002), and cognitive science (Tenenbaum et al.”
ALLEGRO: Belief-Based Programming in Stochastic Dynamical Domains (2015, methodology): “There are many interesting avenues for future work, such as relating belief-based programs to decision-theoretic highlevel programming [Boutilier et al., 2000], particle filters [Thrun et al., 2005], and probabilistic programming [Goodman et al., 2008], among others.”
DART: Dense Articulated Real-Time Tracking (2014, background): “The models we track are defined as a set of rigid bodies connected to each other in a kinematic tree.”