Research on Technology for Three-Dimensional Shape Reconstruction of Offshore Structures
Precision Navigation is the Key to 3D Shape Reconstruction
With recent advances in robotics and autonomous navigation technology, vigorous research is being carried out globally on the use of unmanned surface vessels (USVs), not only in dangerous maritime duties that would previously have required manned vessels, but also in various other tasks such as long-term maritime survey and reconnaissance and surveillance. In particular, USVs can be utilized for surveying damage in costal marine environments in the aftermath of massive disasters such as typhoons or tidal waves, and for performing physical safety inspections of offshore structures such as bridges, port facilities, and offshore oil rig platforms. For successful 3D shape reconstruction of a surrounding environment, it is essential to identify the exact location of a USV equipped with an instrumentation device, and this requires a high level of precision in navigation.
Limitations of Existing Navigation Equipment and Shape Reconstruction
A GPS/INS system, in which GPS is integrated with an inertial navigation system (INS), is the most commonly used navigation system for USVs today. Sometimes, however, GPS signals are not received successfully, not only in indoor environments but also in areas where high-rise buildings are densely located, and its use is substantially limited in and around large-scale offshore structures such as bridges and floating offshore plants. Precision navigation of USVs in an environment where GPS-based navigation is limited requires a method of obtaining separate information for location correction. In principle, it is possible to extract and use relative position information from the location and shape information of structural components that exist in the surrounding environment such as bridge piers or hull structures of a floating offshore platform. To obtain shape information of the surrounding environment, it is common to use a light detection and ranging (lidar) system capable of providing the relative distance between the sensor and the surrounding environment, as well as their bearing information. It is possible to obtain point cloud information and to reconstruct a 3D shape by scanning various parts of the surrounding environment. However, most research on 3D shape reconstruction is premised on the use of a ground vehicle that is in a stationary state or is able to move after stopping with position information from GPS or external beacons. As such, this may present much greater challenges to reconstructing 3D shapes with precision in a marine environment where influences of relatively bigger disturbances such as tidal currents and waves occur continuously with no external position fixes from GPS.
3D Shape Reconstruction through Relative Navigation and Sensor Fusion
A research team led by Professor Kim Jinwhan of the Department of Mechanical Engineering at KAIST is garnering much attention, as they have successfully developed a technology to perform precision 3D shape reconstruction in a marine environment using a newly invented navigation algorithm and sensor fusion techniques. The research team configured a relative navigation algorithm using the relative location information of offshore structures along with inertial sensors and Doppler velocity log (DVL) information to estimate the location of a USV around an offshore structure such as a bridge. In addition, the team proposed an algorithm and procedure that can be used to reconstruct 3D shapes simultaneously above and beneath the surface by combining the information from lidar and sonar systems installed on a vessel. Furthermore, the researchers validated the usefulness of the proposed algorithm by successfully performing outdoor experiments using the team’s own USV in an actual bridge and semi-submerged offshore structure environment. As mentioned above, this technology can be utilized in performing safety inspections and external inspections for the structural stability of large offshore structures such as bridges and floating structures. This is anticipated to be especially useful in natural disaster-prone areas or regions experiencing leakage of dangerous material, where it is difficult for people to perform such tasks. The research outcome earned high recognition by winning third place at the Student Poster Competition, which was held as part of the “2016 MTS/IEEE Oceans” conference, the biggest academic symposium in the field of marine IT, co-sponsored by the Marine Technology Society (MTS) and the Institute of Electrical and Electronics Engineers (IEEE).
Prof. Kim, Jin Whan
2016 Annual Report