Neural SLAM

By Jingwei Zhang, Lei Tai, Joschka Boedecker, Wolfram Burgard. 2017

68 citations recorded by the source corpus.

We present an approach for agents to learn representations of a global map from sensor data, to aid their exploration in new environments. To achieve this, we embed procedures mimicking that of traditional Simultaneous Localization and Mapping (SLAM) into the soft attention based addressing of external memory architectures, in which the external memory acts as an internal representation of the environment. This structure encourages the evolution of SLAM-like behaviors inside a completely differentiable deep neural network. We show that this approach can help reinforcement learning agents to successfully explore new environments where long-term memory is essential. We validate our approach in both challenging grid-world environments and preliminary Gazebo experiments. A video of our experiments can be found at: this https URL.

Subfields: Aerospace Engineering, Artificial Intelligence

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

Citations in the corpus

A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation (2016, positive): “…with only pixel inputs (Mnih et al., 2015, 2016), to learning policies capable of driving real robotic systems in navigation (Zhu et al., 2017b; Zhang et al., 2017a; Tai et al., 2017) and manipulation (Levine et al., 2016; Yu et al., 2018) tasks, the learning power of deep networks drives the…”
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].”
Learning to plan with uncertain topological maps (2020, unknown): “Research in learning to use a metric map is extensive and includes spatially structured memory [40], Neural SLAM based approaches [61] and approaches incorporating projective geometry and neural memory [21,8], these techniques are combined, extended and evaluated in [6].”
Learning Exploration Policies for Navigation (2019, methodology): “Zhang et al. (2017) uses a differentiable map structure to mimic the SLAM techniques.”
MapNet: An Allocentric Spatial Memory for Mapping Environments (2018, background): “These methods fill in only part of the role of SLAM, since they do not attempt to build a map of the environment, and make necessarily incremental predictions.”
Training recurrent networks to generate hypotheses about how the brain solves hard navigation problems (2016, background): “(Also see [26], a new preprint posted while this paper was under review, reporting on a SLAM implementation with recurrent neural network components.”
Unifying Map and Landmark Based Representations for Visual Navigation (2017, result): “In contrast to the above methods, techniques based on learned map construction [12,17,25,35] and value iteration networks [31] have sought to directly incorporate elements of mapping and planning into end-to-end trained navigational policies.”
Global Pose Estimation with an Attention-Based Recurrent Network (2018, methodology): “Relaxing this assumption and resembling traditional occupancy grid SLAM, Neural SLAM [49] uses an occupancy-grid-like memory map, assuming only an initial pose is provided, and updates the pose beliefs and grid map using end-to-end DRL.”
VR-Goggles for Robots: Real-to-Sim Domain Adaptation for Visual Control (2018, background): “For example, 2D Lidar [12, 13, 14] and depth images [15, 16] are sometimes chosen as the sensor modality for transferring the navigation policies to the real world, since the discrepancies between the simulated domain and the real-world domain for them are smaller than those for color images.”
Socially Compliant Navigation Through Raw Depth Inputs with Generative Adversarial Imitation Learning (2017, background): “from raw sensory inputs [5], [6], [7], [8].”