Completed Engineering Computing & AI

1901330122 hyperSwarm

Funder UKRI
Recipient organisation Hypertunnel Limited
Funding £1.6M
Period Nov 2022 — Jul 2024

In plain English

AI plain-English summary

A swarm of thousands of robots will 3D-print a tunnel from the inside out, then dig the surrounding soil away afterwards. Conventional tunnelling requires digging a hole first and then reinforcing it, a slow process that risks collapse in unknown geology. This project flips that sequence. Robots first drill and line a set of small pilot bores along the planned tunnel route. They scan the surrounding ground with ground-penetrating radar and take core samples, feeding the data into a digital twin—a virtual model of the entire tunnel and its geology. Artificial intelligence then designs the optimal structure. A swarm of thousands of autonomous bots, coordinated like a termite colony, enters the lined bores and injects chemicals into the ground at precise locations, 3D-printing the tunnel walls in place. Survey bots then inspect the result. The final tunnel is a smart structure that can be monitored for its entire lifespan. If successful, the technology could enable up to 200 metres of tunnel construction per day—far faster than current methods. The team has already built a 6-metre-long, 2.5-metre-wide demonstration tunnel and is currently testing the system with Network Rail. The impact would be on underground infrastructure: faster, cheaper, and safer rail, road, and utility tunnels, built with less surface disruption.

View original technical description
hyperTunnel will revolutionise underground construction by applying digital twins, 3D printing and autonomy. Where tunnellers previously dug a hole and built a tunnel, we will build a tunnel, then dig the hole. We drill and line pilot bores and send robots inside to inspect the geology taking core samples and scanning using ground penetrating radar, for a near perfect understanding of the entire tunnel length’s geology ? Using this data, we develop a virtual model of the tunnel structure; the digital twin. With AI and machine learning we design the optimum solution to create a sound structure in the geology. ? Once the structure profile is defined, we send a swarm of bots into these lined bores to visit planned locations in order to drill and deploy chemistry according to the AI generated design. ? Thousands of robots will be used, all controlled using swarm technology to 3D print the tunnel in the same way that bees build a hive or termites build a mound. ?? We then inspect again with our survey bots to ensure the chemical has spread evenly and matches the digital twin design. ? The tunnel walls are prepared for final use, leaving a smart structure that can be monitored and maintained throughout its life. We have successfully built a 6m long and 2.5m wide tunnel. We are currently demonstrating our technology for the UK rail infrastructure operator, Network Rail. Our aim in hyperSwarm is to develop robotics and autonomous technology allowing us to scan and construct using a common platform, with multiple compatible toolsets. Success will allow up to 200m of construction per day.

View the original record at the funder ↗

Related Research

Grants with similar aims, by meaning.

Global Centre of Rail Excellence: Underpass
GCRE: Railway Construction Innovation Phase 2, (Project Nr: 10060478)
COnstruction-phase diGItal Twin mOdel
Scalable and Modular robotic tools for pipeline inspection and repair
Digitally Assured Tunnel Assets Installation System (DATA-IS)

Original classification

EU-Funded

Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.