Sven Koenig

About

Sven Koenig is a pioneering researcher in artificial intelligence, robotics, and autonomous systems, whose work has fundamentally shaped how robots navigate and coordinate in complex, uncertain environments. Best known for his contributions to **heuristic search and replanning**, Koenig developed D* Lite (394 citations) and related fast replanning algorithms (679 and 296 citations), enabling mobile robots to efficiently replan paths as they discover unknown terrain — a cornerstone capability in modern autonomous navigation. His early work on probabilistic robot navigation using partially observable Markov models (488 citations) established rigorous probabilistic foundations for robot localization that influenced an entire generation of researchers. Koenig has also been instrumental in advancing **multi-agent pathfinding (MAPF)**, contributing landmark algorithms including PRIMAL (399 citations), EECBS (186 citations), and foundational benchmark frameworks (276 citations) that have become standard references in the field. His Theta* algorithm (212 citations) introduced elegant any-angle path planning widely adopted in robotics and game AI. Beyond algorithms, his theoretical work on auction-based multi-robot routing (294 citations) provided rigorous performance guarantees for decentralized coordination. With thousands of cumulative citations across diverse topics, Koenig stands as one of the most influential figures in autonomous robot planning research.

Research Focus

Key Achievements

Top Papers

1. 1
   Fast replanning for navigation in unknown terrain

   679 citations · 2005
2. 2
   Probabilistic robot navigation in partially observable environments

   488 citations · 1995
3. 3
   PRIMAL: Pathfinding via Reinforcement and Imitation Multi-Agent Learning

   399 citations · 2019
4. 4
   D*lite

   394 citations · 2002
5. 5
   Improved fast replanning for robot navigation in unknown terrain

   296 citations · 2003
6. 6
   Auction-Based Multi-Robot Routing

   294 citations · 2005
7. 7
   Multi-Agent Pathfinding: Definitions, Variants, and Benchmarks

   276 citations · 2021
8. 8
   Theta*: any-angle path planning on grids

   212 citations · 2007
9. 9
   A layered architecture for office delivery robots

   196 citations · 1997
10. 10
    EECBS: A Bounded-Suboptimal Search for Multi-Agent Path Finding

    186 citations · 2021

Key Collaborators

Contact & Links