Steen Savstrup Kristensen

About

Steen Savstrup Kristensen is a robotics researcher whose work spans mobile robot localization, sensor planning, and human-robot interaction for manufacturing applications. He is perhaps best known for developing the Multi-Hypothesis Localization (MHL) method, a probabilistic approach to global robot localization that combines multi-hypothesis Kalman filtering with Bayesian pose tracking. This foundational contribution, first presented in 1999 and expanded in a 2001 publication that has since accumulated over 316 citations, addressed the critical challenge of enabling mobile robots to determine their position reliably within incomplete topological world models. Beyond localization, Kristensen made early contributions to model-driven vision for indoor navigation and advanced sensor planning using Bayesian decision theory, demonstrating a consistent commitment to principled probabilistic methods in robotics. His later research shifted toward interactive and cooperative robotics, exploring how manufacturing assistant robots could be taught new tasks through human instruction — work that bridged autonomous operation with practical human collaboration on the factory floor. A 2004 comparative study further cemented his methodological rigor by benchmarking MHL against six competing localization approaches using standardized datasets. Across his career, Kristensen's research has garnered over 550 citations, reflecting lasting influence on the fields of robot localization and intelligent manufacturing systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Active global localization for a mobile robot using multiple hypothesis tracking

   316 citations · 2001
2. 2
   Active global localisation for a mobile robot using multiple hypothesis tracking

   55 citations · 1999
3. 3
   Model-driven vision for in-door navigation

   33 citations · 1994
4. 4
   Sensor planning with Bayesian decision theory

   29 citations · 1997
5. 5
   Toward interactive learning for manufacturing assistants

   27 citations · 2003
6. 6
   Bin-picking with a solid state range camera

   26 citations · 2001
7. 7
   An experimental comparison of localisation methods, the mhl sessions

   20 citations · 2004
8. 8
   Towards interactive learning for manufacturing assistants

   17 citations · 2002
9. 9
   Human-friendly interaction for learning and cooperation

   15 citations · 2002
10. 10
    Sensor planning with bayesian decision theory

    15 citations · 1996

Key Collaborators

Contact & Links