Infrastructure robotics is being explored as a way to inspect, maintain and repair essential systems with less disruption to everyday life. Research at the University of Leeds has focused on two challenging environments: roads, where defects such as cracks need to be found and repaired, and buried pipes, where inspection robots must operate in confined spaces. The work combines robotics with computer vision, civil and mechanical engineering, materials science and wireless power research.
These systems are research demonstrators rather than fully deployed maintenance services. Their value lies in showing how autonomous machines might perform specific tasks and in identifying the technical integration challenges that must be addressed before wider use.
Infrastructure robotics for responsive city maintenance
The Self-Repairing Cities project considers how robotic systems could help maintain urban infrastructure while supporting the health and wellbeing of residents and the surrounding environment. Its themes include maintaining lighting columns, finding and patching road defects, and deploying robots inside utility pipes.
These tasks place different demands on automation. A road repair system must detect a defect and position its repair equipment accurately on an exposed surface. A pipe robot must navigate a narrow, enclosed route and gather information or perform work without relying on a conventional tether. The projects therefore address both the task itself and the supporting technologies, including sensing, control, mobility and power.
A road repair drone that detects and patches cracks
The Self-Repairing Cities team developed an asphalt 3D-printing drone as a demonstrator for the “Perceive and Patch” concept. The system is designed to identify cracks in a road surface during flight, land near a detected defect and deposit asphalt into it using an extruder.
The demonstrator brings together a DJI hexacopter, a CoLiDo delta frame, custom tracks for ground mobility and an asphalt extruder. An onboard computer vision system processes images of the road and identifies points corresponding to the crack. The resulting X-Y coordinates are passed to a control process, which positions the extruder and guides it along the identified path as asphalt is deposited.
This workflow highlights a practical systems-integration challenge: detection is only useful if the output can guide the repair mechanism. In this case, computer vision, software and mechanical control must work together to translate an irregular crack in an image into movement by the extrusion system. The demonstrator showed a narrow, irregular crack being filled. However, testing of the repair’s quality and longevity was still ongoing, so the work should not be interpreted as evidence of long-term road performance.
Pipe inspection robots and the power challenge
The Pipebots project focuses on autonomous robots for buried water, drainage and sewerage pipes. Its broader ambition is to develop a group of robots able to operate within pipes for extended periods while carrying out tasks such as inspection, repair, metering and reporting. For sustained operation, a key engineering question is how to recharge a robot without a physical power tether.
Conventional wireless power methods involve trade-offs. Near-field approaches, including inductive and capacitive coupling, are generally limited to short distances. Far-field transmission can operate over greater distances, but atmospheric losses and limits on permitted transmission power constrain potential applications.
A metal pipe offers a different possibility. Under suitable conditions, it can act as a circular waveguide, guiding electromagnetic energy along its interior. This creates a potential route for wireless power transfer to a robot inside the pipe. The method depends on factors such as pipe dimensions, operating frequency and the condition of the inner surface. Corrosion and material buildup can increase propagation loss, making the actual pipe environment important to measure rather than assume.
Testing wireless power in a metal pipe
Researchers tested a small inspection robot designed for a 25.4 mm-diameter gas pipe. The experiment used a two-metre decommissioned cast-iron pipe, with radio-frequency energy coupled into the pipe using a horn antenna. The robot included a controller, a brushed DC motor, backup batteries, a night-vision camera and a wireless power receiver.
The receiver used an antenna and an eight-stage voltage-doubler rectifier built with Schottky diodes. Its purpose was to convert received radio-frequency energy into direct-current power. The reported best result was 23% rectification efficiency, producing 18 mW of DC power at 12 GHz with 19 dBm of RF power at the rectifier input. These measurements demonstrate power conversion in the test setup; they do not establish that the system can fully power or indefinitely recharge a deployed robot. A voltage regulator and battery-charging circuit could be connected to the rectifier, although the available space inside a small robot is a design constraint.
What these demonstrations mean for automation teams
For industrial automation engineers and infrastructure operators, the research illustrates several considerations for robotic maintenance projects:
- Design around the operating environment. Road surfaces and buried pipes present different requirements for mobility, sensing and equipment protection.
- Connect perception to action. A defect-detection system must provide information in a form that the repair mechanism can use reliably.
- Validate the complete system. Robot, control software, sensing, repair tools and power supply must function as an integrated system.
- Measure real-world constraints. Pipe material and surface condition affect electromagnetic propagation, while space limits the size of onboard electronics and charging components.
Infrastructure robotics is not yet a replacement for established maintenance methods on the basis of these demonstrations. It is a developing approach that shows how targeted automation could address difficult inspection and repair tasks. The road drone links computer vision with asphalt deposition, while the pipe robot research investigates whether a metal conduit can support wireless power delivery. Together, these projects underline the importance of testing performance, reliability and integration under realistic operating conditions before moving from research prototypes to routine infrastructure maintenance.