Associate Professor Tobias Fischer
Brain-inspired, energy-efficient robots that know where they are: from neuromorphic place recognition to field robots on the Great Barrier Reef, and the open-source software that powers them.
Robots that work in the real world
I am an Associate Professor and ARC DECRA Fellow at the Queensland University of Technology, and have been awarded a $1.1M ARC Future Fellowship commencing in November 2027. My research focuses on robot localisation and underwater perception, blending neuroscience and robotics to push the boundaries of intelligent systems operating under resource constraints.
I serve as Faculty Lead of the $1.3M Translation to Deployment (T2D) project in the Reef Restoration and Adaptation Program, and Chief Investigator of the $740k Queensland Quantum and Advanced Technologies Talent Building Program, with previous grants from Intel Labs and Amazon. I am also a core contributor to open-source robotics software used around the world, including RoboStack and Pixi.
After obtaining my PhD from Imperial College London (UK Best PhD in Robotics Award), I have published more than 70 peer-reviewed papers in venues including Science Robotics, PAMI, TRO, CVPR, ECCV, ICCV, IJCAI, ICRA and IROS, and my work has been recognised with several best paper and best poster awards. I serve as Area Chair and Associate Editor for ICRA, IROS, RSS and RA-L, co-chair the IEEE RAS Women in Engineering Committee, and my PhD students have gone on to careers as Assistant Professors, Research Scientists and Research Fellows.
Neuromorphic Computing
Event cameras and spiking
networks for positioning
Robot Localisation
Place recognition across
seasons, light and viewpoint
Underwater Perception
AI for monitoring and
restoring coral reefs
Open-source Robotics
RoboStack, Pixi, ROS2WASM
and Nav2
News
- PreprintNew preprint led by postdoc Somayeh Hussaini: Removing the Compute Bottleneck for Event-based Visual Place Recognition with Spiking Neural Networks, a sparse spiking network that resolves place recognition queries in tens of microseconds on a standard CPU.
- PreprintNew preprint led by PhD student Chelim Lim: Image-Space Refraction Correction for Underwater 3D Reconstruction, which removes flat-port refraction so standard SfM and SLAM produce undistorted underwater reconstructions.
- TalkInvited talks at IROS 2026: a keynote on ground truth and benchmarks at the Data in Field Robotics workshop, and a talk on underwater robotic vision for reef restoration at the Perception and Navigation for Marine Robotics workshop (on behalf of Scarlett Raine).
- PaperThree papers at IROS 2026: AI-driven dispensing of coral reseeding devices for the Great Barrier Reef, event-based visual teach-and-repeat, and quantile transfer for operating point selection in visual place recognition.
- PreprintNew preprint: CGRAS, a robotic system that monitors tens of thousands of coral recruits in aquaculture, cutting monitoring time and labour by a factor of 9.6.
Show more
- AwardAwarded an ARC Future Fellowship ($1.1M AUD, 2027–2031), which will follow on from my current ARC DECRA Fellowship.
- PreprintTwo new event-camera preprints led by Adam Hines: the open-source EventCV library for event-based robotic vision, and MegaEvent for multi-viewpoint geo-localization.
- PaperAccepted to WACV 2027: Why Domain Matters, domain-aware benchmarking of underwater object detection and annotation quality, led by PhD student Melanie Wille.
- TalkThree talks at the ADSTAR Summit 2026 on assured positioning without GPS: latency-driven visual positioning, neuromorphic positioning for energy-efficient autonomy, and neuromorphic and quantum sensing for underwater positioning.
- AppointmentPromoted to Associate Professor at the Queensland University of Technology, effective 1 August 2026.
- PreprintNew preprint: TACO, a test-and-check framework for robust pose graph optimization, led by postdoc Emilio Olivastri.
- PaperPublished in the International Journal of Hydrogen Energy: time-resolved oxygen bubble dynamics in proton exchange membrane electrolysers.
- PaperPresented at ICUAS 2026: aerial visual place recognition in Antarctica for robust monitoring in extreme environments.
- TalkAt ICRA 2026: a keynote tutorial with Peter Corke, From Research Code to Running Systems, workshop talks on low-latency neuromorphic localization and underwater robotic vision for reef restoration, and three papers on sub-millisecond event-based place recognition, Event-LAB and coral spawn monitoring.
- PreprintNew preprint: DisPlace, discriminative place projections for multi-reference visual place recognition, led by PhD student Dhyey Manish Rajani.
Fellowships
Brain-inspired positioning for fast, energy-efficient, resilient robotics
This project will develop the world’s first sub-millisecond, energy-efficient robot positioning system for reliable positioning in complex and resource-limited environments. It develops a novel brain-inspired approach where sensing and computation occur only when the scene changes, mirroring how neurons fire on meaningful events to maximise efficiency.
About the Future Fellowship
Adaptive and efficient robot positioning through model and task fusion
This project aims to create fit-for-purpose positioning systems that continuously adapt to diverse and changing environments. It addresses two fundamental limitations in the state of the art: continuous adaptation, critically important in safety-critical systems, and energy efficiency, critically important in resource-constrained systems.
About the DECRA FellowshipRobots for Reef Restoration
Efficient, robust perception also takes our robots into the field. As Faculty Lead of a Reef Restoration and Adaptation Program project, my team builds AI that guides where coral seeding devices are dispensed, segments reef imagery from a handful of labels, and re-localises robots in changing underwater scenes.
Underwater researchProjects
A selection of what my team is working on, from neuromorphic localisation to open-source software.
Neuromorphic Place Recognition
Spiking neural networks that recognise places on neuromorphic hardware, using a fraction of the usual energy.
Learn More
Event Cameras for Fast Localisation
Sensors that report only what changes, for sub-millisecond localisation in difficult lighting.
Learn More
Robust Visual Place Recognition
Recognising places across seasons, day and night, and large changes in viewpoint, and knowing when to trust the answer.
Learn More
Introspection for Reliable Localisation
Robots that know when their localisation can be trusted, and adapt before it fails.
Learn More
Reef Guidance System
AI that decides in real time where coral seeding devices should be dispensed on the Great Barrier Reef.
Learn More
RoboStack
The Robot Operating System on conda-forge, for Linux, macOS and Windows alike.
Learn More
Pixi
Unified, reproducible software development and distribution for robotics and AI.
Learn More
Nav2 Smac Planner
Open-source, cost-aware, kinematically feasible path planning for mobile and surface robots.
Learn MoreTalks & Videos
Keynotes, tutorials and public-facing explainers. Organising an event? See Media & Speaking.
ICRA 2026 Keynote Tutorial: From Research Code to Running Systems
Live coding with Pixi and conda-forge to compose ROS and PyTorch into portable, reproducible workflows. Tutorial materials.
ICRA 2025 Keynote: Navigating Without GPS Using Bio-Inspired Approaches
Bio-inspired, low-power sensing for place recognition: event cameras, spiking models for localisation, and coral-reef applications.
How Robots Navigate Without GPS Using Vision
A one-minute explainer on visual navigation indoors, underground and underwater. Read the article.
Robotic Vision That Works in the Real World
MIT Robotics Worldwide Workshop: neuromorphic place recognition, reef perception, and open-source tools such as RoboStack and ROS2WASM.
Diversity & Inclusion
Robotics is better when the people building it reflect the people it serves. I want my group to be a place where everyone is welcome and supported to succeed, whatever their gender, culture, background, sexuality or ability. I co-chair the IEEE RAS Women in Engineering Committee, am part of the QUT Ally network, and particularly encourage applications from women and First Nations peoples.