TacEdge 2026: International Workshop on Tactical Edge Technologies

A one-day workshop on sensing, communications and autonomy in contested environments, with invited talks and hands-on tutorials. TacEdge 2026 is a one-day, in-person workshop on sensing, communications and autonomy in contested environments, run by the School of Engineering and Technology at UNSW Canberra. The subjects are RF and optical surveillance, tactical telecommunications, and applied AI in the field. The workshop is for people who are not specialists in these fields. It suits Defence and defence-industry staff, researchers and students who want to understand current methods and their limits. The morning is free and open to everyone. Doors open at 08:30 and the first talk starts at 09:00. There are two keynotes and four short talks before lunch. Session times include introductions and questions. Registration is still required. The afternoon tutorials are in addition to the talks. A paid ticket covers the whole day, the morning talks and the tutorials, and is limited to 80 participants. Group registration is possible. Four hands-on tutorials run in parallel, one in each room, and each runs twice. Every participant attends two of the four, and gives a first and second choice when registering. Each room takes up to 20 people per session. No prior experience with radio, radar or machine learning is assumed. Bring a laptop. The first tutorial is an introduction to software-defined radio and spectrum monitoring. Participants capture a live band with an SDR receiver and learn to read a waterfall display like the one above: which signals are present, when they transmit and how much bandwidth they occupy. The second, led by MAJ Matthew Hume and Dr Lily Qiao, shows how to use open satellite imagery as an intelligence source. Participants process Sentinel-1 radar data (calibration, speckle reduction and terrain correction) and compare the result with optical imagery of the same area. The picture at the top of this page is a Sentinel-1 radar image of the Strait of Gibraltar. The bright points on the water are ships. In the third, led by Muhammad Saeed and Drew Stevenson, participants train YOLO and similar detectors on real imagery and examine the confidence scores they report, including cases where a confident detection is wrong. The session then shows how an ensemble of detectors reduces misclassification and the number of detections an operator has to check by hand. In the picture above, a detector has boxed the ships in a radar image. The fourth, led by Dr Josh Andres, is a scenario-based workshop on human-AI interaction at the tactical edge. Working in small groups, participants use a structured Prepare-Make-Reflect process to build scenarios involving people, AI systems, autonomous platforms and their operating environment, then examine human oversight, trust, responsibility, operational dependencies and unintended consequences. A robot capture-the-flag competition on the Plato Pod arena, shown below, is planned for 16:00 in the lecture hall and is still to be confirmed. Coffee and lunch are included for everyone attending, morning or afternoon, sponsored by the School of Engineering and Technology at UNSW Canberra. The full-day ticket costs $150 for domestic students, $300 for all other participants. Prices are in Australian dollars and include GST. Payment is by Eventbrite only, and invoices and bank transfers are not accepted. Refunds can be requested within 7 days of purchase. A small number of tutorial places may be offered free of charge to selected participants. TacEdge runs the day after the Australian Space Domain Awareness and Space Traffic Management Conference (10-12 November), at the same venue. The workshop is organised with the UNSW Defence Research Institute, UNSW Canberra Space, the Korean Institute of Communications and Information Sciences, the National Taiwan University of Science and Technology, Philippine Coding Camp and ASR Defence. The programme is provisional, and sessions and timings may change. Full details are at tacedge.org. Questions go to Artem Lensky, [email protected].