Spatial Relation Graph and Graph Convolutional Network for Object Goal Navigation
This addresses the problem of efficient object search for robots in unknown environments, but it appears incremental as it builds on existing graph-based and learning techniques.
The paper tackles the object-goal navigation task by developing a framework that uses a Spatial Relational Graph and Graph Convolutional Network to learn embeddings for region proximity and object occurrence, enabling a robot to locate target objects through Bayesian inference and region ranking. The result is a method that improves navigation efficiency, though no concrete numbers are provided in the abstract.
This paper describes a framework for the object-goal navigation task, which requires a robot to find and move to the closest instance of a target object class from a random starting position. The framework uses a history of robot trajectories to learn a Spatial Relational Graph (SRG) and Graph Convolutional Network (GCN)-based embeddings for the likelihood of proximity of different semantically-labeled regions and the occurrence of different object classes in these regions. To locate a target object instance during evaluation, the robot uses Bayesian inference and the SRG to estimate the visible regions, and uses the learned GCN embeddings to rank visible regions and select the region to explore next.