Probabilistic Active Loop Closure for Exploration

Uncertainty-aware autonomous exploration to improve mapping accuracy.

This project models loop closure as a stochastic event and uses that model to guide exploration decisions.

Main ideas:

  • Formulated active loop closure as an uncertainty reduction objective over pose graphs.
  • Designed exploration policies that trade off frontier coverage and loop-closure probability.
  • Reduced accumulated localization uncertainty and improved mapping quality.

Related publication:

  • Probabilistic Active Loop Closure for Autonomous Exploration (ICRA 2024)