NB Embedded Pty Ltd Logo
NB Embedded Pty Ltd

Presenting at MRS 2025 — Resilient Multi-Robot Coordination Under Unreliable Networks

Presenting at MRS 2025 — Resilient Multi-Robot Coordination Under Unreliable Networks

I’m excited to share that I’ll be travelling to Singapore next month to attend the IEEE International Symposium on Multi-Robot & Multi-Agent Systems (IEEE MRS 2025), where I’ll be presenting a poster on my PhD work: a protocol for resilient multi-robot coordination under highly unreliable and bandwidth-constrained networks.

This is my first time attending MRS, and I’m looking forward to meeting researchers and practitioners pushing the limits of autonomy, coordination, and field robotics in demanding real-world environments.

Why This Research Matters

Most multi-robot systems research assumes reasonably stable, high-throughput networks. Real robots rarely get that luxury.

Even with high-quality radios and modern 4G/5G, Wi-Fi 6/7, and MANET systems, field network behaviour is shaped by physics, terrain, interference, load, and mobility. As a result, network performance is often:

This isn’t a failure of networking hardware—it’s the nature of challenging environments.

When coordinating multiple autonomous robots under these conditions, traditional assumptions can break down. Shared maps diverge, synchronisation lags, and systems become overly conservative or inefficient.

My work explores how robots can coordinate effectively even when the network is behaving exactly as expected in the real world, using approximate and value-aware information sharing instead of continuous full-state exchange.

A Lightweight Protocol for Real-World Constraints

The protocol I’m presenting is designed for:

Rather than transmitting full maps or trajectories, robots exchange compressed, prioritised, and predictive updates that are:

In simulation, this leads to:

It’s part of a broader theme in my PhD at UTS:

How can we design autonomy systems that remain robust across the entire spectrum of real network behaviour?

Robots should be network-aware, taking advantage of good connectivity while continuing to function effectively during degradation.

Where This Fits in the Larger Picture

In industry—across defence, mining, agriculture, and mobile robotics—connectivity is rarely predictable. Vendors are doing outstanding work pushing the limits of modern radios, but autonomous systems place a unique combination of constraints on networks that were originally designed for humans, smartphones, or static IoT deployments [1].

Robots operate as teams, move through rapidly changing RF environments, and depend on timely, mission-critical information. This creates a growing gap between what general-purpose communication systems were built for and what autonomous robots actually need.

The path forward isn’t to expect radios to solve autonomy problems, or for autonomy algorithms to assume perfect networks. The most robust systems will come from co-designing both layers together—communication hardware that understands mission context and autonomy algorithms that remain stable across the full spectrum of network behaviour.

That’s the motivation behind this research and why I’m so interested in the emerging opportunity for robotics-focused communication platforms built specifically for real-world autonomy.

If You’ll Be at MRS 2025

I’ll be at the full event and would love to talk about:

If you’re attending and want to meet up, feel free to reach out:

📩 hello@nbembedded.com

Acknowledgements

Thank you to my advisors and the Robotics Institute at UTS for their support.

More Soon

After the conference, I’ll be posting:

Looking forward to seeing everyone in Singapore.

References

[1] J. Gielis, A. Shankar, and A. Prorok, “A Critical Review of Communications in Multi-robot Systems,” Curr Robot Rep, vol. 3, no. 4, pp. 213–225, Dec. 2022, doi: 10.1007/s43154-022-00090-9. https://arxiv.org/abs/2206.09484

NB Embedded is a consulting practice focused on embedded Linux, wireless networking, and robotics autonomy infrastructure, helping teams build reliable systems for real-world environments.