Arash Asgharivaskasi
Papers
11
Total Citations
149
H-Index
6
About
Arash Asgharivaskasi is a robotics researcher specializing in autonomous exploration, semantic mapping, and information-theoretic planning for mobile robot systems. His work addresses the fundamental challenge of enabling robots to efficiently perceive, map, and navigate unknown environments using streaming sensor data. Asgharivaskasi's most impactful contributions center on information-based exploration techniques, particularly extending mutual information frameworks to semantic and multi-class settings. His seminal work on semantic OcTree mapping and Shannon mutual information computation (54 citations) and active Bayesian multi-class mapping (33 citations) demonstrate how robots can fuse range and visual observations to build richer, semantically meaningful maps. He has further developed iterative Covariance Regulation (iCR) for continuous trajectory optimization over SE(3) manifolds and pioneered Riemannian optimization approaches for multi-robot active mapping teams. More recently, Asgharivaskasi has bridged classical robotic mapping with modern AI by integrating large language model guidance into scene graph planning (17 citations), allowing robots to interpret and execute natural language tasks within hierarchical metric-semantic models. His research on learning continuous control policies for active perception further connects information-theoretic principles with modern machine learning. With over 140 cumulative citations, his body of work represents meaningful advances in making autonomous robots more capable, adaptive, and semantically aware.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Optimal Scene Graph Planning with Large Language Model Guidance17 citations · 2024
- 4
- 5
- 6Riemannian Optimization for Active Mapping With Robot Teams7 citations · 2025
- 7
- 8Optimal Scene Graph Planning with Large Language Model Guidance4 citations · 2023
- 9
- 10