Value Iteration Networks on Multiple Levels of Abstraction

Abstract

Learning-based methods are promising to plan robot motion without performing extensive search, which is needed by many non-learning approaches. Recently, Value Iteration Networks (VINs) received much interest since-in contrast to standard CNN-based architectures-they learn goal-directed behaviors which generalize well to unseen domains. However, VINs are restricted to small and low-dimensional domains, limiting their applicability to real-world planning problems.To address this issue, we propose to extend VINs to representations with multiple levels of abstraction. While the vicinity of the robot is represented in sufficient detail, the representation gets spatially coarser with increasing distance from the robot. The information loss caused by the decreasing resolution is compensated by increasing the number of features representing a cell. We show that our approach is capable of solving significantly larger 2D grid world planning tasks than the original VIN implementation. In contrast to a multiresolution coarse-to-fine VIN implementation which does not employ additional descriptive features, our approach is capable of solving challenging environments, which demonstrates that the proposed method learns to encode useful information in the additional features. As an application for solving real-world planning tasks, we successfully employ our method to plan omnidirectional driving for a search-and-rescue robot in cluttered terrain.

3D Dynamic Scene Graphs: Actionable Spatial Perception with Places, Objects, and Humans

“…The multiple levels of abstraction afforded by a DSG have the potential to enable hierarchical and multi-resolution planning approaches [52,97], where a robot can plan at different levels of abstraction to save computational resources.…”

A survey of learning‐based robot motion planning

“…To address the aforementioned challenges and improve generalisation ability, different methods [66][67][68][69][70] have been raised under the framework of VIN. Sufeng Niu et al [68], Daniel Schleich et al [69], and Lisa Lee et al [70] improved the performance of VIN by changing the network architecture with LSTM module, graph representation, multiscale inputs etc. [66,67] extended the model to Semi-Markov Decision Process (SMDP) and Partially Observable Markov Decision Process (POMDP).…”

“…In recent years, deep neural networks are introduced in traditional value-based RL algorithms. In [65][66][67][68][69][70], value-based RL method serves as an end-to-end solution of the motionplanning problem. Aviv Tamar et al [65] proposed a novel Value Iteration Network (VIN) that is able to directly map the image observation of the environment to the planning computation and generate action predictions.…”

Kimera: From SLAM to spatial perception with 3D dynamic scene graphs

“…The multiple levels of abstraction afforded by a DSG have the potential to enable hierarchical and multi-resolution planning approaches (Larsson et al, 2020; Schleich et al, 2019), where a robot can plan at different levels of abstraction to save computational resources.…”