Topological mapping using spectral clustering and classification

In this work we present an online method for generating topological maps from raw sensor information. We first describe an algorithm to automatically decompose a map into submap segments using a graph partitioning technique known as spectral clustering. We then describe how to train a classifier to recognize graph submaps from laser signatures using the AdaBoost machine learning algorithm. We demonstrate that the we can perform topological mapping by incrementally segmenting the world as the robot moves through its environment, and we can close the loop when the learned classifier recognizes that the robot has returned to a previously visited location.

Authors

• Emma Brunskill
• Thomas Kollar
• Nicholas Roy

Keywords

• Clustering algorithms
• Robotics and automation
• Machine learning algorithms
• Partitioning algorithms
• Robot sensing systems
• Simultaneous localization and mapping
• Machine learning
• Connectors
• Orbital robotics
• Robustness
• Spectral Clustering
• Topological Map
• AdaBoost
• Graph Partitioning
• Training Data
• Laser Scanning
• Training Dataset
• Precision And Recall
• Similarity Matrix
• Voronoi Diagram
• Map Representation
• Grid Map
• Web Map
• Nonexpansive Mapping
• Creation Of Maps
• Loop Closure
• Robot Trajectory
• Training Data Points
• Robot Pose
• High Recall Rate
• Spectral Clustering Algorithm
• Magnitude Of Eigenvalues
• Fraction Of Points
• Scale-invariant
• Clustering Techniques
• Machine Learning Techniques
• Training Examples
• Mapping Algorithm
• Precision Rate