Papers

5

Total Citations

120

H-Index

5

About

Lachlan Nicholson is a roboticist whose research advances the frontiers of semantic perception and reproducible benchmarking in robotics. He is best known for pioneering **QuadricSLAM**, a landmark framework that integrates object-level detections—specifically constrained dual quadrics—into the Simultaneous Localization And Mapping (SLAM) pipeline. This work enables robots to build richer, semantic world models, moving beyond sparse point clouds to meaningful object representations that support higher-level reasoning and interaction. His contributions in this area have been cited over 25 times, establishing a foundation for subsequent semantic SLAM research. Nicholson also made a significant impact on reproducible robotics through **The ACRV Picking Benchmark (APB)**, a standardized shelf-picking challenge designed to foster fair, comparable evaluation of robotic manipulation systems. With over 80 citations, this benchmark has been adopted by labs worldwide to assess and improve robotic picking performance. Additionally, his investigation of **Dropout Sampling for robust object detection** under open-set conditions has informed Bayesian deep learning approaches in perception. Nicholson’s work bridges the gap between robust perception and practical, reproducible experimentation, making him a key figure in the push toward more intelligent, deployable robotic systems.

Research Focus

Key Achievements

5
H-Index
5
Papers
120
Total Citations
24
Avg Citations/Paper
🏆 Most Cited Paper
The ACRV picking benchmark: A robotic shelf picking benchmark to foster reproducible research
81 citations · 2017
📈 Most Prolific Year: 2018 (3 Papers)
🤝 Key Collaborators: 22
🏛 Institutions: Queensland University of Technology, Australian Centre for Robotic Vision

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago