Automating UAV for Inspecting Bridge Elements

-1-

Automating UAV for Inspecting Bridge Elements

Nie-Jia YAU1, Hsien-Ke LIAO2, Ming-Kuan TSAI3, Cheng-Wei SU4, Jyun-Hao HUANG5, Ming-Yi JIANG6, and Po-Yuan CHEN7

1Professor, Graduate Institute of Construction Engineering and Management (CEM),

National Central University (NCU), No.300, Zhung-Da Road, Zhungli District, Taoyuan 32001, Taiwan

2Post-Doctoral Researcher, Graduate Institute of CEM, NCU, No.300, Zhung-Da Road, Zhungli District, Taoyuan 32001, Taiwan

3Associate Research Fellow, Research Center for Hazard Mitigation and Prevention, NCU, No.300, Zhung-Da Road, Zhungli District, Taoyuan 32001, Taiwan

4Division Head, Transportation Planning Division, Institute of Transportation (IOT), Ministry of Transportation and Communications (MOTC), No.240, Dunhua North

Road, Taipei 10548, Taiwan 5Research Fellow, Transportation Planning Division, IOT, MOTC, No.240, Dunhua

North Road, Taipei 10548, Taiwan 6Associate Research Fellow, Transportation Planning Division, IOT, MOTC, No.240,

Dunhua North Road, Taipei 10548, Taiwan 7Technical Specialist, Department of Railways and Highways, MOTC, No.50, Section

1, Ren-Ai Road, Taipei 10052, Taiwan E-mails: [email protected]; [email protected]; [email protected];

[email protected]; [email protected]; [email protected]; [email protected]

ABSTRACT Currently there are around 28,000 bridges in Taiwan that are managed and

maintained by Taiwan Railways Administration, Taiwan Area National Freeway Bureau, Directorate General of Highways, and city and county governments. According to bridge maintenance manuals and codes, these bridges are inspected visually and assisted by devices if necessary. Normally, the visual inspection is performed by inspectors who approach bridge components by walking or climbing. For bridges having high piers or crossing deep rivers, inspection trucks or boats are utilized to perform such inspection. However, for bridges having large amount of traffic volume or being located in rural areas, those trucks are too bulky to be used. There is much room for improvement for the bridge inspection task.

This study attempts to apply unmanned aerial vehicles (UAV) to bridge inspection to improve safety, efficiency, and results of bridge inspection. A low-cost UAV system is automated by preplanned flying routes to capture images of bridge components using its loaded cameras. In order to capture images under bridge deck, such as girder and deck, the UAV is designed to fly beneath the bridge even when GPS signals are minimum. After several test flights of the UAV system, it shows the system fulfills the aforementioned tasks of bridge inspection. There are three restrictions when applying this UAV system: (1) the wind speed should be under 8 meters per second; (2) the clearance under bridge girder should be over 6 meters; and (3) the width of bridge should be less than 60 meters. In the future, this system is expected to integrate with the second-generation of Taiwan Bridge Management System (TBMS2) to record the bridge inspection results.

1. INTRODUCTION

Bridges are common infrastructures in many countries. In any country, since bridges connect various areas, the result is a rapidly developing economy and culture. Exposed to both natural and human-made effects, bridge elements may be damaged; for example, bridge foundations can be displaced by sudden earthquakes, bridge piers compromised by recurrent floods, and bridge pavement damaged by heavy vehicles. For bridges built several decades ago, to confirm whether their structural safety corresponds to the original design, inspection has become imperative. If a bridge is found to have a safety problem, it should be repaired as soon as possible.

Taiwan has more than 28,000 bridges. To assist bridge inspectors in performing bridge inspections, many researchers have proposed various approaches (Clark et al., 2003; Sung and Wang, 2013). These approaches can be classified into destructive testing (e.g., core drilling, loading capacity) and non-destructive testing (e.g., r, laser, ultrasonic, infrared, and thermal detection). Compared to destructive testing, non-destructive testing does not severely damage bridge structures, so this type of testing is more frequently used. Among the approaches in non-destructive testing, because of the limited resources (e.g., personnel, funding, equipment), visual inspection is the most common (Liao and Yau, 2011). To assist inspectors in getting close to bridge elements, tools such as bridge inspection vehicles, high-resolution cameras, and telescopes are available.

However, after inspectors finish on-site bridge inspection via visual observation in Taiwan, a main problem (i.e., inaccurate reports regarding bridge inspection) emerges. Since many (more than 30%) of the bridges in Taiwan are either stream-crossing or mountain-surrounding, the inspectors have difficulty in observing bridge elements and looking for defects and damage (Phares et al., 2000; Liao and Yau, 2011). For example, the inspectors must tie ropes to bridge railings so as to rappel to the bridge piers that stand in rivers. In such an inconvenient and dangerous working environment, the inspectors cannot perform bridge inspection completely and correctly. In other words, the efficiency of current bridge inspection via visual observation needs improvement.

Thus, this study proposes an unmanned aerial vehicle (UAV)-based approach. With this approach, inspectors can rapidly and safely obtain the imagery of bridge elements although they do not get close to these elements. The accuracy of their reports regarding bridge inspection would be enhanced. This study will serve as a useful reference for similar applications in bridge management. 2. BRIDGE INSPECTION

Bridge inspection in Taiwan is divided into three types: daily patrol, regular inspection every two years, and damage inspection after sudden events (e.g., earthquakes, rainstorms). The emphases of the three types are to identify the use patterns for bridges, examine the status of bridge elements, and understand whether sudden events affect the structural safety of bridges, respectively. 2.1 Daily Patrol

Daily patrols are meant to inspect deterioration that may cause discomfort or danger to the users, and the patrols are performed by two inspectors by driving over the bridge to see if anything is unusual.

-3-

2.2 Damage Inspection

There are many typhoons and earthquakes that hit Taiwan every year. Strong flood currents and earthquakes may cause various degrees of damages to bridges. Thus, a special inspection; i.e., damage inspection, is required only after extremely heavy rains or earthquakes exceeding a particular intensity. The inspection form is simpler and different from that of regulation inspection. Once a bridge is seriously damaged due to these natural disasters, the bridge needs to be closed and maintenance activities are immediately activated by a relevant maintenance agency. 2.3 Regular Inspection

Regular inspections are the most important type of inspection for a bridge, and each bridge needs to be inspected once every two years according to the current maintenance regulations. Some agencies may increase the frequency of regular inspections to once per year due to flood threats.

Such inspections are performed directly by trained inspectors with handheld tools. For either stream-crossing or mountain-surrounding bridges, inspectors adopt specific tools at the bridge sites. These include bridge inspection trucks, cherry pickers, and inflatable boats. Instead of manual data collection through bridge inspection vehicles and boats, 3D LiDAR illuminates bridge elements with a laser and analyses the reflected light (Arayici, 2007). Based on optical theories and image recognition, the inspectors can define any defect in bridge elements. 3. PROBLEM STATEMENT

Based on observation of bridge inspectors executing regular bridge inspections, this study determined that the adopted tools were frequently unavailable because of several on-site limitations. Due to the huge volume of bridge inspection trucks and cheery pickers, these vehicles cannot arrive at some bridge sites, and the inflatables boat cannot work in rapids. When 3D LiDAR is configured inappropriately on-site, no data for specific bridge elements (e.g., the bottom of bridge slabs) can be retrieved. With these scenarios, many bridge inspectors simplify various processes of bridge inspection, especially for stream-crossing and mountain-surrounding bridges. A common situation is that the inspectors document inspection results of bridge elements by referring to previous data. As a result, the Taiwanese government receives inaccurate reports regarding bridge inspection. To improve the accuracy of inspection reports, inspectors need an alternative solution to overcome the aforementioned limitations in visual inspection. 4. APPROACH

Visual observation with imagery is common for data analysis in various fields, including military activities, disaster management, land management, and earth science (Pratt and Murphy, 2006; Shiau and Ma, 2015). To collect the imagery of inaccessible sites, UAVs have become popular in recent years because they can act as camera carriers and image transmitters.

For bridge inspection via visual observation, it is necessary for bridge inspectors to note the appearance of bridge elements to define the condition of these elements. The use of UAVs seems to fulfill such requirement (Murphy et al., 2011). To explore whether UAVs offer more advantages for bridge inspection than other tools (i.e., bridge

inspection vehicles, inflatable boats, 3D LiDAR), this study performs an evaluation. Table 1 shows the results. In contrast to other tools, UAVs are better; they are cheaper to purchase and maintain, involve fewer personnel for manual operation, need less time for imagery collection, result in lower interference for traffic, and have better carry at bridge sites. More importantly, UAVs can capture the images of bridge elements once and easily repeat identical activities.

Table 1: Comparing various tools for bridge inspection Tools UAVs Bridge inspection

vehicle Inflatable

boat 3D LiDAR

Cost (USD) 1,000-10,000 200,000-400,000 1,000-5,000 70,000-180,000 Working location Sky Ground River Ground

Portable Easy Difficult Hard Easy Operators 1-2 3-4 2-3 1-2

Affecting traffic Low High Low Low Capturing

imagery once Yes No No No

Speed for imagery

collection Fast Slow Slow Medium

Repeating activities Easy Difficult Medium Easy

Some researchers had attempted to apply UAVs to bridge inspection. For

example, Whang et al. (2007) developed a UAV with coaxial rotors to fly and navigate the lower surface of bridges. Metni and Hamel (2007) presented a control law based on computer vision for quasi-stationary flights above a planar target. Ceng (2011) adopted a multi-rotor copter to manually capture images of bridges and analysed the imagery to identify any defect. Guerrero and Bestaoui (2013) proposed an approach for bridge inspection in windy environments so as to reach time-optimal UAV path planning. Also, several commercial platforms (e.g., Aibotix Inc., 2015) were available. These solutions ensured that bridge inspectors saved time on building their hardware and software regarding UAVs.

However, the aforementioned studies neither described how to help bridge inspectors who are new to UAVs command their UAVs during bridge inspection nor illustrated whether their UAVs autonomously worked under bridges while obstacles existed. Therefore, this study focuses on providing an effective approach for inspectors. In this approach, the research hypotheses include:

1. Simple Operation: After inspectors plan the pathways regarding bridge inspection, UAVs can reach autonomous flight based on these pathways instead of manual control.

2. Transparent Information: When UAVs work around bridges (i.e., top, bottom, left, and right sides), inspectors can receive and save the data transferred from the UAVs.

3. Safe Inspection: If UAVs arrive at the specified bridge elements, inspectors can capture the images of these elements through the cameras embedded in the UAVs. Meantime, UAVs could reach obstacle avoidance.

-5-

4. Reliable Reports: Since inspectors access the images of the bridge elements in real time, they can detail inspection results at bridge sites.

5. IMPLEMENTATION

UAVs roughly include two main categories (i.e., planes and copters). Unlike planes with fixed wings, copters carry multi-rotors. In contrast to planes, copters more easily fly based on vertical directions. In the meantime, copters can hover over assigned spots for a period of time. Thus, during bridge inspection, it is more likely that copters rather than planes can capture the images of specific bridge elements from various viewpoints (e.g., from top to bottom of a bridge pier). This study implemented this approach by using copters.

A copter-based UAV includes hardware and software development. For hardware, this study assembles a quad-rotor copter consisting of a 55-centimeter (CM) frame and numerous necessary components. These components included a 6200 milli-ampere-hour (mAH) battery, a global positioning system (GPS) sensor, a radio transmitter with 433 mega-hertz (MHz) frequency, a video transmitter with 5.8 giga-hertz (GHz) frequency, a radio receiver with 2.4 GHz frequency, a camera, two 10-inch propeller sets, and four 880 kilo-voltage (kV) rotors with four individual electronic speed controllers. Considering that many bridge inspectors may have little experience regarding UAV operation, this study adopted fully automatic operations (i.e., UAVs are completely controlled by users from take-off through landing). However, UAVs are easily crashed by a number of obstacles that exist at bridge sites (e.g., trees and electric cables). To avoid such condition, this study embedded six sonars in the used UAVs. When the UAVs detected any obstacles, the UAVs would intelligently fly away.

For software, this study constructed a ground control station (GCS). Through this GCS, the inspectors can receive the status of UAVs in real time. To enable simple GCS use by inspectors at bridge sites, this study developed the GCS in Google Android-based tablets. The tablets can work for several hours continuously and are lightweight. Redundant devices can be reduced to increase on-site mobility. The inspectors can interact with the GCS through gestures. In the meantime, the GCS can inform the inspectors of the messages received from the UAVs via text-to-speech. Thus, the inspectors do not need to focus on the tablet’s screen.

Figure 1 shows a flow chart for when bridge inspectors apply UAVs and the GCS in bridge inspection. The GCS includes four modules (i.e., flight plans, flight missions, imagery control, and inspection results). Flight plans guide the UAVs to collect the anticipated images of bridge elements. The inspectors have two options (i.e., on-site production and historical records) to create their plans. Regarding on-site production, the inspectors draft waypoints that are near bridge elements in a Google Maps-based geographic information system (GIS; Figure 2) to form flight pathways. If the inspectors want to ask UAVs to hover over specific waypoints, they should describe the flight altitude (e.g., flying above or below a bridge element) and stay time.

However, the representation of the GIS may differ from the bridge sites. For example, the inspectors may have difficulty in identifying piers of a bridge. To resolve this problem, they can execute on-site surveying. Because of the GPS chips embedded in the tablets, when inspectors walk along the bridges, the GCS records their tracks. For this reason, the correct position of any bridge element can be determined. Unlike in on-site production, by importing historical records, the inspectors can immediately finish the flight plans if they have ever used GCS and UAVs to inspect

identical bridges.

Figure 1: Flowchart for the approach

Figure 2: Drafting flight pathways

-7-

A flight plan is a flight mission. When receiving the flight plans from the

inspectors, UAVs are ready to execute the flight missions. During the flight missions, the GCS provides three functions (i.e., flight status, flight traces, flight imagery). Based on the flight status, the inspectors understand the flight speed and altitude, the usage of battery, and the conditions of GPS. Since the GCS detects the just-in-time locations of UAVs, the flight traces help inspectors confirm whether the true flight pathways correspond to their flight plans.

Figure 3: Accessing videos to fill in inspection results

Figure 4: Accessing photos to fill in inspection results

After UAVs finish their flight missions, by accessing the videos and photos that

are stored in the cameras and GCS, the inspectors can fill in the inspection results regarding the bridge elements. For videos, if the inspectors find any defect in bridge elements, they can take a snapshot of the frame of the videos and explain such defect in the inspection report (Figure 3). Regarding photos, the inspectors have difficulty in verifying the relationship between the quantity of photos and the bridge elements. Therefore, the GCS combines these photos and the flight plans in the GIS. Figure 4

shows that 25 photos are captured when a UAV flies under the bottom of the slab of a bridge. By clicking the presented icons, the inspectors can fill in the inspection results. Based on the aforementioned four modules, bridge inspectors can benefit from the approach when dealing with bridge inspections via vision observation. 6. TEST

To understand whether the approach supported the research hypotheses, this study performed several tests with various tools at bridge sites. Table 2 details the bridges. Bridge 1 is a stream-crossing concrete girder bridge with five spans. Some public pipes and cables are also attached to this bridge. Bridge 2 is both a stream-crossing and a mountain-surrounding steel arch bridge with one span. For the two bridges, the lengths are 64 and 220 meters, the widths of the slabs are 8 and 20 meters, and the heights from the slabs to the rivers are 3.5 and 150 meters, respectively. The width of Bridge 1 is small. The bridge inspection vehicles cannot work at the bridge site. Since the water areas of the river are narrow, this study failed when operating with inflatable boats. Although the traffic on this bridge was heavy, the 3D LiDAR and UAVs were smoothly applied in bridge inspection. Table 2 also shows the test results. To capture the imagery of bridge elements, the 3D LiDAR took 70 minutes and the UAVs needed only 10. When filling in the inspection reports, this study spent more than 15 minutes waiting for data generation with the 3D LiDAR (Figure 5). In contrast, the UAVs offered the true imagery of the bridge elements. This study could easily identify the defects of these elements, including the cracks in the wearing surface of the bridge slab, the efflorescence for the concrete beams, and the rusty public pipes (Figure 6).

Table 2: Tested bridges Bridge 1 Bridge 2

Age More than 15 years 10 years Type Girder bridge Steel arch bridge Span 5 1 Lane 2 2

Length 64 m 220 m Width 8 m 20 m Height 3.5 m 150 m Tool UAV 3D LiDAR UAV 3D LiDAR

Operators 1 2 1 2 Using time 10 minutes 70 minutes 24 minutes 70 minutes

Applied device Quadcopter Trimble TX5 Quadcopter Trimble TX5 Cost of device USD 5,000 USD 60,000 USD 5,000 USD 60,000

Affecting traffic No No No No

The test results at Bridge 2 are shown in Table 2 also. The bridge inspection vehicles cannot arrive at the bridge site since such vehicles failed in passing through the mountain roads. Two conditions meant that the inflatable boats were unavailable. One was that the inflatable boats cannot be operated in some of the water areas of the river. The other was that the distance from the river to the bridge was too far, so visual observation of bridge elements failed. Although the 3D LiDAR worked (needing 60 minutes), the captured imagery of the bridge elements was incomplete

-9-

because the 3D LiDAR cannot be configured for observations under the bridge. The status regarding the bottom of the bridge slab was still unknown. The UAVs took 24 minutes to capture the imagery of all bridge elements. For the inspection reports, this study was able to illustrate several defective bridge elements, such as rusty slabs (Figure 7) and efflorescent concrete.

Figure 5: Results of 3D LiDAR at Bridge 1

Figure 6: Results of UAV at Bridge 1

Cracks

Figure 7: Results of UAV at Bridge 2

Moreover, this study compared the test results of two bridges and the regular

inspection reports. Although the defects in the bridge elements defined through the approach did not immediately affect the structural safety of the two bridges, these defects were not recorded in the regular inspection reports. Therefore, the approach indeed fulfilled the research hypotheses. When bridge inspectors simply established on-site flight plans, UAVs captured the transparent imagery of bridge elements. Associated with the increased safety for bridge inspection, the results for these elements were still reliable. 6. DISCUSSION

Although the approach succeeded at many bridge sites, some limitations may affect the efficiency of bridge inspection: 1. The wind speed: In order to reduce the cost of UAVs, in this study, a quad-rotor

copter consisting of a 55-centimeter frame was applied. For safety, the UAVs can perform when the wind speed under 8 meters per second.

2. The specifications of bridge type: UAVs will not work for all of these bridges because the GPS embedded in the UAVs may be interrupted when they fly under the bottom of bridge slabs. To avoid having UAVs collide with bridge elements, for automatic flight missions, the height from the bottom of the slabs to the surface of the rivers should be more than 6 meters. The width of bridge should less than 60 meters. To resolve this limitation, integrating various positioning and collision-avoidance techniques with the approach is necessary.

3. The length of a flight pathway: Due to the limited capacity of batteries, UAVs neither fly for a long distance nor hover for a long time. With a 6,200 mAH battery, the UAVs that were adopted in this study can work for 14 minutes continuously. When the UAVs executed automatic flight missions, the flight

Rustines

-11-

speed was 3 meters per second. Thus, the total length of the flight pathway was 2,520 meters. Based on the tests, because of this limitation, bridge inspectors had difficulty collecting the imagery of all bridge elements for long-spin bridges. Although the inspectors could use batteries with higher capacity (e.g., 10,000 or 15,000 mAH), the UAVs still have the limitation of flight time. Therefore, flight missions should be divided into several sub-missions.

7. CONCLUSION

In a number of countries, many people have been killed by suddenly damaged bridges. Such scenario highlights that bridge inspection is the foundation of ensuring public safety. For Taiwan, according to the regulation of regular inspection, bridge inspectors should inspect more than 14,000 bridges every year. Although several tools (i.e., bridge inspection vehicles, inflatable boats, 3D LiDAR) are beneficial for visual inspection, the inspectors still have difficulty in using these tools at bridge sites, especially for stream-crossing and mountain-surrounding bridges. Since the reports regarding bridge inspection were inaccurate, the Taiwanese government could not comprehend the current status of the unsafe bridges and schedule maintenance for them. As a result, with natural disasters and human-made events, the unsafe bridges were damaged and caused accidents.

To improve the aforementioned problem, this study proposed a UAV-based approach. During bridge inspection, bridge inspectors simply established their flight plans for the UAVs with the developed mobile device-based GCS. According to the received plans, since the UAVs automatically captured the images of bridge elements, the inspectors did not have to move close to the bridge elements. In the meantime, the information exchange between the UAVs and the inspectors was in real time. Finally, by browsing the obtained videos and photos, the inspectors rapidly finished their inspection reports. In contrast to bridge inspection vehicles, inflatable boats, and 3D LiDAR, the test results showed that the UAV-based approach was more appropriately applied at bridge sites. In sum, this study serves as a useful reference for similar applications in bridge management. 8. ACKNOWDEMENTS

The authors appreciate the financial support provided by the Institute of Transportation, Ministry of Transportation and Communications (MOTC), Taiwan, R.O.C. under project number MOTC-IOT-103-PEB017, and MOTC-IOT-104-PEB019. 9. REFERENCES Aibotix Inc. (2015), The UAV Talent for All Sensors, Kassel, Germany:

https://www.aibotix.com/en/overview-aibot-uav.html. Arayici, Y. (2007), “An Approach for Real World Data Modelling with the 3D

Terrestrial Laser Scanner for Built Environment,” Automation in Construction, Vol. 16, No. 6, pp. 816-829.

Ceng, Z.-J. (2011), The Application of Multi-Axis Aircraft in Bridge Inspection, National Kaohsiung University of Applied Sciences, Kaohsiung, Taiwan.

Clark, M.R., McCann, D.M., and Forde, M.C. (2003), “Application of Infrared Thermography to the Non-Destructive Testing of Concrete and Masonry Bridges,” NDT & E International, Vol. 36, No. 4, pp. 265-275.

Guerrero, J.A. and Bestaoui, Y. (2013), “UAV Path Planning for Structure Inspection in

Windy Environments,” Journal of Intelligent and Robotic Systems, Vol. 69, pp. 297-311.

Liao, H.-K. and Yau, N.-J. (2011), “Development of Various Bridge Condition Indices for Taiwan Bridge Management System,” Proceedings of the 28th International Symposium on Automation and Robotics in Construction (ISARC), Korea.

Metni, N. and Hamel, T. (2007), “A UAV for Bridge Inspection: Visual Serving Control Law with Orientation Limits,” Automation in Construction, Vol. 17, No. 1, pp. 3-10.

Murphy, R.R., Steimle, E., Hall, M., Lindemuth, M., Trejo, D., Hurlebaus, S., Zenon, M.C., and Slocum, D. (2011), “Robot-Assisted Bridge Inspection,” Journal of Intelligent and Robotic Systems, Vol. 64, No. 1, pp. 77-95.

Phares, B.M., Rolander, D.D., Graybeal, B.A., and Washer, G.A. (2000), “Studying the Reliability of Bridge Inspection,” Public Roads, Vol. 64, No. 3, pp. 15-19.

Pratt, K. and Murphy, R.R. (2006), Requirements for Semi-Autonomous Flight in Miniature UAVs for Structural Inspection, U.S.A.: Association for Unmanned Vehicle Systems.

Shiau, J.-K. and Ma, D.-M. (2015), “Development of an Experimental Solar-Powered Unmanned Aerial Vehicle,” Journal of the Chinese Institute of Engineers, Vol. 38, No. 6, pp. 701-713.

Sung, Y.-C. and Wang, C.-Y. (2013), “A Study on Damage Assessment of the Scoured Bridges,” Journal of the Chinese Institute of Engineers, Vol. 36, No. 8, pp. 994-1007.

Whang, S.-H., Kim, D.-H., Kang, M.-S., Cho, K., Park, S., and Son, W.-H. (2007), “Development of a Flying Robot System for Visual Inspection of Bridges,” Proceedings of the 3rd International Conference on Structural Health Monitoring of Intelligent Infrastructure, British Columbia, Canada.

Yau, N.-J., Ye, C.-G., and Chen, M.-C. (2010), Establishment of Bridge Visual Inspection and Evaluation Manual (Draft), Report to Institute of Transportation, MOTC, Taipei, Taiwan.

Yau, N.-J., Tsai, M.-K., and Liao, H.-K. (2013), Development of Taiwan Bridge Management System for the Second Generation, Report to Institute of Transportation, MOTC, Taipei, Taiwan.

-13-

Hsien-Ke LIAO, Ph.D. Post-Doctoral Researcher, Graduate Institute of Construction Engineering and Management, National Central University Mailing Address: No.300, Zhung-Da Road,

Zhungli District, Taoyuan 32001, Taiwan Phone: +886-919-972-281 Fax: +886-3-425-7092 E-mail: [email protected] RESEARCH AND ACADEMIC INTERESTS Bridge Management System.

3D bridge inspection application software (APP).

EDUCATION M.Sc. in Construction Engineering and Management, National Central University, Taoyuan,

Taiwan, 2005. Ph.D. in Civil Engineering, National Central University, Taoyuan, Taiwan, 2015.

CURRENT POSITION Post-Doctoral Researcher, Graduate Institute of Construction Engineering and Management,

National Central University, 2015.12~Present. Exploring the development of bridge inspection methodologies in Taiwan.

WORK EXPERIENCES AND ACHIEVEMENTS Research Assistant, Graduate Institute of Construction Engineering and Management, National

Central University, 2013.04~2015.12. Establishing the second generation of Taiwan Bridge Management System (TBMS2).

SELECTED PUBLICATIONS Yen, C.-I., Liao, H.-K., Chai, J.-J., and Yau, N.-J. (2010), “A Heuristic Approach for

Optimizing Bridge Inspection Route,” The 5th International Conference on Bridge Maintenance, Safety and Management (IABMAS2010), Philadelphia, Pennsylvania, U.S.A.

Liao, H.-K. and Yau, N.-J. (2011), “Development of Various Bridge Condition Indices for Taiwan Bridge Management System,” The 28th International Symposium on Automation and Robotics in Construction (ISARC2011), Seoul, Korea.

Yau, N.-J., Tsai, M.-K., Liao, H.-K., Su, C.-W., Huang, J.-H., Jiang, M.-Y., and Chen, P.-Y. (2015), “Innovative 3-Dimensional Bridge Modeling for Bridge Management in Taiwan,” International Bridge Conference 2015 (IBC 2015), Pittsburgh, Pennsylvania, U.S.A.

Yau, N.-J. and Liao, H.-K. (2015), “Indices for Fast Assessment of Bridge Condition in Taiwan,” Applied Mechanic and Materials, Vols. 752-753, pp. 689-697.

Yau, N.-J. and Liao, H.-K. (2015), “Establishing a Decision Support Module for Bridge Maintenance in Taiwan,” accepted by Advanced Materials and Engineering Structural Technology.

Proceedings of the 11th US-Taiwan Bridge Engineering Workshop Taipei, Taiwan, October 20~21, 2016