Skip to main content

Autonomous Air-ground Cooperative Tunnel Inspector

Hong Kong Productivity Council, Civil Engineering and Development Department and Hyder-Meinhardt Joint Venture have co-developed the world’s first autonomous air-ground cooperative tunnel inspector. With the synergy of Artificial Intelligence (AI), robotics and total-station positioning techniques, the air-ground cooperated robotic system will be employed in the Trunk Road T2 and Cha Kwo Ling Tunnel project to carry out all necessary tunnel inspections, covering tunnel linings and various tunnel internal structures - bringing intelligence, automation and digital advancement to tunnel inspection in Hong Kong.

Three state-of-the-art techniques are applied in this system, including

  1. edge-computing technology to achieve on-site AI inspection;
  2. heterogeneous multi-robot cooperation based on Unmanned Ground Vehicle (UGV) and Unmanned Aerial Vehicle (UAV) to achieve full inspection coverage in tunnels;
  3. integration with automated total stations to achieve global localisation in a non-GPS area.


In addition, by providing the HK local coordinates of the captured images and digitalising the tunnel inspection records, the system provides an automatic and intelligent upgrade to future tunnel maintenance for smart asset management.

Service Details:
  • Minimise the deficiencies and drawbacks inherited from traditional manual inspection of tunnel structures.
  • Automatically collect high-resolution images without assigning additional workers and substantive working platforms.
  • Achieve full autonomy in detecting different defects on the internal surface of the tunnels instantly on-site, such as concrete cracks, concrete spalling and water leakage on tunnel linings, thermal barriers, vitreous enamel panel claddings and carriageway paving.

Autonomous Air-ground Cooperative Tunnel Inspector
Autonomous Air-ground Cooperative Tunnel Inspector
Autonomous Air-ground Cooperative Tunnel Inspector
Autonomous Air-ground Cooperative Tunnel Inspector