NeuralSymbolic SLAM Cognitive Architecture Autonomous Navigation Complex Environments

This dataset is derived from the research of ' NeuralSymbolic SLAM Cognitive Architecture Autonomous Navigation Complex Environments '. It contains 1200 sets of real indoor and simulated environment samples, covering four typical navigation environments : hospital corridors, intelligent warehouses, outdoor scenes and large shopping malls. The data is synchronously collected by RGB-D camera and 2D laser radar, and provides pixel-level 50-class target semantic annotation, logical relationship label and corresponding navigation decision record, which is used to support the perception, reasoning and decision-making algorithm verification of neural-symbol fusion SLAM system in dynamic semantic scene.