Stephen Marsland
Universities UK, University of Manchester, Massey University, University of Edinburgh
Papers
15
Total Citations
1,050
H-Index
13
About
Stephen Marsland is a computational researcher whose work spans machine learning, autonomous robotics, and biological pattern recognition. He is perhaps best known for his pioneering contributions to novelty detection in mobile robots — developing algorithms that enable robots to identify unexpected or unusual features in their environments by learning internal models of "normality." This thread of research, represented by multiple influential papers from 2000 to 2005 and accumulating nearly 200 citations across several works, drew on habituation-inspired mechanisms and self-organising neural networks to give robots a form of perceptual awareness. Marsland also made significant contributions to autonomous map-building, co-authoring papers on globally consistent mapping techniques that helped robots overcome the persistent challenge of odometry drift — work that together attracted over 250 citations. His 2002 paper introducing a self-organising network that grows dynamically as needed remains his most cited work, with 387 citations, reflecting its broad applicability across machine learning communities. Beyond robotics, Marsland extended his pattern recognition expertise into biology, contributing to automated pollen classification systems for palynology laboratories, earning nearly 100 citations and demonstrating the real-world versatility of his computational methods. His body of work reflects a sustained commitment to intelligent, adaptive systems.
Research Focus
Key Achievements
Top Papers
- 1A self-organising network that grows when required387 citations · 2002
- 2Fast, On-Line Learning of Globally Consistent Maps126 citations · 2002
- 3Learning globally consistent maps by relaxation122 citations · 2002
- 4
- 5On-line novelty detection for autonomous mobile robots89 citations · 2005
- 6A Real-Time Novelty Detector for a Mobile Robot48 citations · 2000
- 7Detecting Novel Features of an Environment Using Habituation36 citations · 2000
- 8Learning to select distinctive landmarks for mobile robot navigation30 citations · 2001
- 9A tale of two filters-on-line novelty detection26 citations · 2003
- 10Environment-Specific Novelty Detection21 citations · 2002