Infrastructure Robotics E-Book
111,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: IEEE Series on Systems Science and Engineering
- Sprache: Englisch
Infrastructure Robotics Illuminating resource presenting commonly used robotic methodologies and technologies, with recent developments and clear application examples across different project types Infrastructure Robotics presents state-of-the-art research in infrastructure robotics and key methodologies that enable the development of intelligent robots for operation in civil infrastructure environments, describing sensing, perception, localization, map building, environmental and operation awareness, motion and task planning, design methodologies, robot assistance paradigms, and physical human-robot collaboration. The text also presents many case studies of robotic systems developed for real-world applications in maintaining various civil infrastructures, including steel bridges, tunnels, underground water mains, underwater structures, and sewer pipes. In addition, later chapters discuss lessons learned in deployment of intelligent robots in practical applications overall. Infrastructure Robotics provides a timely and thorough treatment of the subject pertaining to recent developments, such as computer vision and machine learning techniques that have been used in inspection and condition assessment of critical civil infrastructures, including bridges, tunnels, and more. Written by highly qualified contributors with significant experience in both academia and industry, Infrastructure Robotics covers topics such as: * Design methods for application of robots in civil infrastructure inspired by biological systems including ants, inchworms, and humans * Fundamental aspects of research on intelligent robotic co-workers for human-robot collaborative operations * The ROBO-SPECT European project and a robotized alternative to manual tunnel structural inspection and assessment * Wider context for the use of additive manufacturing techniques on construction sites Infrastructure Robotics is an essential resource for researchers, engineers, and graduate students in related fields. Professionals in civil engineering, asset management, and project management who wish to be on the cutting edge of the future of their industries will also benefit from the text.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 663
Veröffentlichungsjahr: 2023
Ähnliche
Table of Contents
Cover
Table of Contents
Title Page
Copyright
About the Editors
Preface
Acronyms
Part I: Methodologies
1 Infrastructure Robotics: An Introduction
1.1 Infrastructure Inspection and Maintenance
1.2 Infrastructure Robotics
1.3 Considerations in Infrastructure Robotics Research
1.4 Opportunities and Challenges
1.5 Concluding Remarks
Bibliography
2 Design of Infrastructure Robotic Systems
2.1 Special Features of Infrastructure
2.2 The Design Process
2.3 Types of Robots and Their Design and Operation
2.4 Software System Design
2.5 An Example: Development of the CROC Design Concept
2.6 Some Other Examples
2.7 Actuator Systems
2.8 Concluding Remarks
Bibliography
3 Perception in Complex and Unstructured Infrastructure Environments
3.1 Introduction
3.2 Sensor Description
3.3 Problem Description
3.4 Theoretical Foundations
3.5 Implementation
3.6 Case Studies
3.7 Conclusion and Discussion
Bibliography
4 Machine Learning and Computer Vision Applications in Civil Infrastructure Inspection and Monitoring
4.1 Introduction
4.2 GNN-Based Pipe Failure Prediction
4.3 Computer Vision-Based Signal Aspect Transition Detection
4.4 Conclusion and Discussion
Bibliography
5 Coverage Planning and Motion Planning of Intelligent Robots for Civil Infrastructure Maintenance
5.1 Introduction to Coverage and Motion Planning
5.2 Coverage Planning Algorithms for a Single Robot
5.3 Coverage Planning Algorithms for Multiple Robots
5.4 Conclusion
Bibliography
Note
6 Methodologies in Physical Human–Robot Collaboration for Infrastructure Maintenance
6.1 Introduction
6.2 Autonomy, Tele-Operation, and pHRC
6.3 Control Methods
6.4 Adaptive Assistance Paradigms
6.5 Safety Framework for pHRC
6.6 Performance-Based Role Change
6.7 Case Study
6.8 Discussion
Acknowledgements
Bibliography
Note
Part II: Robotic System Design and Applications
7 Steel Bridge Climbing Robot Design and Development
7.1 Introduction
7.2 Recent Climbing Robot Platforms Developed by the ARA Lab
7.3 Overall Design
7.4 Overall Control Architecture
7.5 Experiment Results
7.6 Conclusion and Future Work
Bibliography
Note
8 Underwater Robots for Cleaning and Inspection of Underwater Structures
8.1 Introduction to Maintenance of Underwater Structures
8.2 Robot System Design
8.3 Sensing and Perception in Underwater Environments
8.4 Software Architecture
8.5 Robot Navigation, Motion Planning and System Integration
8.6 Testing in a Lab Setup and Trials in the Field
8.7 Reflection and Lessons Learned
8.8 Conclusion and Future Work
Acknowledgments
Bibliography
Note
9 Tunnel Structural Inspection and Assessment Using an Autonomous Robotic System
9.1 Introduction
9.2 ROBO-SPECT Project
9.3 Inspection Procedure
9.4 Extended Kalman Filter for Mobile Vehicle Localization
9.5 Mobile Vehicle Navigation
9.6 Field Experimental Results
9.7 Conclusion
Bibliography
Note
10 BADGER: Intelligent Robotic System for Underground Construction
10.1 Introduction
10.2 Boring Systems and Methods
10.3 Main Drawbacks
10.4 BADGER System and Components
10.5 Future Trends
Bibliography
Note
11 Robots for Underground Pipe Condition Assessment
11.1 Introduction to Ferro-Magnetic Pipeline Maintenance
11.2 Inspection Robots
11.3 PEC Sensing for Ferromagnetic Wall Thickness Mapping
11.4 Gaussian Processes for Spatial Regression from Sampled Inspection Data
11.5 Field Robotic CA Inspection Results
11.6 Concluding Remarks
Bibliography
Notes
12 Robotics and Sensing for Condition Assessment of Wastewater Pipes
12.1 Introduction
12.2 Nondestructive Sensing System for Condition Assessment of Sewer Walls
12.3 Robotic Tool for Field Deployment
12.4 Laboratory Evaluation
12.5 Field Deployment and Evaluation
12.6 Lessons Learned and Future Directions
12.7 Concluding Remarks
Bibliography
Note
13 A Climbing Robot for Maintenance Operations in Confined Spaces
13.1 Introduction
13.2 Robot Design
13.3 Methodologies
13.4 Experiments and Results
13.5 Discussion
13.6 Conclusion
Bibliography
Note
14 Multi-UAV Systems for Inspection of Industrial and Public Infrastructures
14.1 Introduction
14.2 Multi-UAV Inspection of Electrical Power Systems
14.3 Inspection Planning
14.4 Onboard Online Semantic Mapping
14.5 Conclusion
Bibliography
Note
15 Robotic Platforms for Inspection of Oil Refineries
15.1 Refining Oil for Fuels and Petrochemical Basics
15.2 The Inspection Process
15.3 Inspection and Mechanical Integrity of Oil Refinery Components
15.4 Plant Operations, Surveillance, Maintenance Activities, and Others
15.5 Robotic Systems for Inspection
15.6 Robotics for Plant Operations, Surveillance, Maintenance, and Other Related Activities
15.7 Conclusion
Note
16 Drone-Based Solar Cell Inspection With Autonomous Deep Learning
16.1 Introduction
16.2 Aerial Robot and Detection Framework
16.3 Learning Framework
16.4 Conclusion
Acknowledgments
Bibliography
Note
17 Aerial Repair and Aerial Additive Manufacturing
17.1 Review of State of the Art in Additive Manufacturing at Architectural Scales
17.2 Review of Demonstrations of Aerial Manufacturing and Repair
17.3 Initial Experimental Evaluations
17.4 Conclusion and Discussion
Bibliography
Note
Index
End User License Agreement
List of Tables
Chapter 4
Table 4.1 Results of failure prediction on water pipe dataset.
Chapter 8
Table 8.1 Sensors installed in the robot.
Chapter 9
Table 9.1 Inspection process timing.
Table 9.2 Robotic System accuracy: average precision.
Table 9.3 Robotic System accuracy: component maximum error.
Table 9.4 Robotic System accuracy: chain maximum error.
Chapter 11
Table 11.1 Core component specifications.
Chapter 13
Table 13.1 Success rate of going through a manhole over 20 trials.
Table 13.2 Percentage of false positives and false negatives over 100 trials...
Table 13.3 Robot movement efficiency.
Chapter 14
Table 14.1 Planning results associated with Figure 14.5.
Chapter 16
Table 16.1 Comparison of EL measurement systems.
Table 16.2 Comparison of measurement techniques.
Table 16.3 Open datasets for PV module inspection.
Table 16.4 Data distrubution in dataset.
Table 16.5 Parameters of the CNN model for defective types classification.
List of Illustrations
Chapter 1
Figure 1.1 (a) Sydney Harbor Bridge, an example of a complex steel bridge st...
Figure 1.2 (a) An example of confined space with partition plates each with ...
Figure 1.3 (a) Bridge piles covered by marine growth; (b) A bridge with a nu...
Figure 1.4 (a) A team of human workers painting a transmission tower using b...
Figure 1.5 A robot for stripping rust and old paint using grit-blasting.
Figure 1.6 A Wall-pushing Autonomous Maintenance roBOT (WAuMBot) inside a st...
Figure 1.7 A climbing robot inspecting the steel condition inside a confined...
Figure 1.8 An underwater robot firmly attached to a bridge pile by grasping ...
Chapter 2
Figure 2.1 The CROC robot in front of a partition plate with a manhole in a ...
Figure 2.2 (a) A legged-climbing robot concept; (b) Inchworm(c) an inchw...
Figure 2.3 The magnet release mechanism contained within each toe of CROC....
Figure 2.4 The geometry involved in CROC's ability to place its feet on eith...
Figure 2.5 Software modules of the bio-inspired climbing robot (CROC).
Figure 2.6 This robot is designed to work in the same environment as CROC, b...
Chapter 3
Figure 3.1 (a) The grit blasting robot. Nozzle attached to the end-effector ...
Figure 3.2 Different environment representations. (a) An example of a 2D fea...
Figure 3.3 (a) The red dot in the sonar sensor model is the position of the ...
Figure 3.4 (a) The robot localization in a 2D feature-based map. The dark gr...
Figure 3.5 (a) Illustration of a SLAM problem. A mobile robot moves in an en...
Chapter 4
Figure 4.1 An overview of the pipe failure prediction framework.
Figure 4.2 An example of IOU-based object detection. (a) Image grid; (b) IOU...
Figure 4.3 An example of semantic segmentation for tracks.
Figure 4.4 An overview of the bilateral segmentation network. (a) Network ar...
Figure 4.5 Sample frames with multiple signals and tracks in the camera simu...
Figure 4.6 Sample of track detection results in different cases. (a) Case 1,...
Figure 4.7 Sample of split lines(skeleton) of tracks.
Figure 4.8 The final outputs of the proposed model.
Chapter 5
Figure 5.1 Two mobile AIRs performing grit-blasting on a complex structure. ...
Figure 5.2 Squircles with spirals within them created using varying values o...
Figure 5.3 (Left) FG-squircular mapping; (Right) Mapping of a spiral path. S...
Figure 5.4 (Left) Areas to be covered; (Right) Coverage using Squircular-CPP...
Figure 5.5 (a) Testing using (SPIR); (b) SPIR in real-world environment; (c)...
Figure 5.6 A grazing path of a prey.
Figure 5.7 Eight different scenarios and a path created for each scenario in...
Figure 5.8 (Left) the trajectory of each obstacle; (Right) an example path w...
Figure 5.9 Optimizing base placement of two AIRs relative to complex structu...
Figure 5.10 All base placements and FBPs of one of the AIRs.
Figure 5.11 (a) AIRs applied for steel bridge maintenance; (b) placement of ...
Figure 5.12 Overlapped and specific areas as well as the final paths associa...
Figure 5.13 (Left) overlapped areas; (Middle and Right) two examples of Voro...
Figure 5.14 Four AIRs with different capabilities to cover a flat surface: (...
Figure 5.15 (Left) the scenario; (Right) the resulting paths.
Chapter 6
Figure 6.1 Examples of robotic systems with different levels of human–robot ...
Figure 6.2 Examples of autonomous robots for bridge maintenance: (a) An auto...
Figure 6.3 Teleoperated pipe inspection robot.
Figure 6.4 Depiction of Physical Human–Robot Collaboration. Unlike autonomou...
Figure 6.5 Basic motion control of a robotic system.
Figure 6.6 An admittance control scheme.
Figure 6.7 Example of an admittance control scheme used in Carmichael and Li...
Figure 6.8 Safety framework for risk reduction in applications involving pHR...
Figure 6.9 Example of an FSPN model for a generic task described in Tran e...
Figure 6.10 The Assistance-As-Needed Robot allows skilled workers to perform...
Figure 6.11 Control system of the ANBOT that utilized an admittance control ...
Chapter 7
Figure 7.1 Overall approach for steel bridge inspection robot.
Figure 7.2 (a) Cylinder steel bridge, (b) Complex I bar steel bridge with nu...
Figure 7.3 Climbing robot platforms developed by the ARA lab, University of ...
Figure 7.4 Overall design of robot.
Figure 7.5 Robot function (a) Mobile mode; (b) Transforming mode, or worm mo...
Figure 7.6 The robots foot with flexible magnet array.
Figure 7.7 (a) The robot on flat surface, (b) The robot passing nuts, (c) Th...
Figure 7.8 The robot body-6 DOF robot arm.
Figure 7.9 Extended statics diagram.
Figure 7.10 Turn-over/adhesion diagram.
Figure 7.11 The proposed control system framework for autonomous navigation....
Figure 7.12 Steel bridge inspection robot in (a) mobile transformation, and ...
Figure 7.13 The control architecture integrated into the ARA robot.
Figure 7.14 Boundary point estimation from 3D point cloud data.
Figure 7.15 Distance control system of magnetic array, (a) 3D model, and (b)...
Figure 7.16 An encoder–decoder-based CNN architecture for steel bridge inspe...
Figure 7.17 Inch-worm jump from one steel surface to another.
Figure 7.18 Inch-worm jump procedure.
Figure 7.19 Aruco Marker and coordinate frames for localization.
Figure 7.20 Planar surface extraction from a 3D point cloud of the steel sur...
Figure 7.21 (a), (c) The Boundary set and (b), (d) Area rectangle set.
Figure 7.22 (a) The selected area rectangle, and (b) Pose estimation.
Figure 7.23 Surface height availability check: (a) Same height, and (b) Diff...
Figure 7.24 The movement of robot in
Mobile transformation
and visual inspec...
Figure 7.25 The motion planning for the ARA robot movement.
Figure 7.26
Worming transformation
: (a) magnetic array of the second foot to...
Figure 7.27 The robot is being tested on a steel structure indoors.
Figure 7.28 Deployment of the robot on a steel structure.
Figure 7.29 Mobile mode example on the bridge.
Chapter 8
Figure 8.1 Manual operation in removing marine growth using high-pressure wa...
Figure 8.2 The underwater robot annotated with the main components of the sy...
Figure 8.3 Computational fluid dynamics for evaluating hull designs.
Figure 8.4 3-DOF arm for removing marine growth using water-jet blasting Le ...
Figure 8.5 Top-down view of different grasping arm configurations: (Left) fo...
Figure 8.6 ICP-based registration of a 3D map onto a ground truth pile: Side...
Figure 8.7 Marine growth identified based on a threshold of point-to-pile di...
Figure 8.8 Marine growth identification results compared against actual pile...
Figure 8.9 Software architecture.
Figure 8.10 SPIR navigation in open water.
Figure 8.11 General view of SPIR system.
Figure 8.12 Lab setup: water tank (left) and the mock bridge pile in the tan...
Figure 8.13 Field trial: the team was putting the robot into water next to a...
Figure 8.14 Flowchart of the operation procedure (diagram
Figure 8.15 Diagram of the robot's spiral trajectory used to cover the pile ...
Figure 8.16 Bridge pile detection using sonar scanning.
Figure 8.17 Navigation toward the target pile.
Figure 8.18 Thick marine growth on the Windang Bridge piles.
Figure 8.19 Deformable spiral path (diagram source Le et al. [2020])/IEEE.
Figure 8.20 A surface before and after cleaning.
Figure 8.21 (Left) circled regions showing large marine growth on bridge pil...
Chapter 9
Figure 9.1 The ROBO-SPECT robotic system is composed of a mobile vehicle, an...
Figure 9.2 The robotic arm with ultrasonic sensor tool and vision system are...
Figure 9.3 US tool positioned on a detected crack to take width and depth me...
Figure 9.4 The IGC communicates with the different subsystems of the robotic...
Figure 9.5 The GCS allows the operator to prepare and monitor the state of t...
Figure 9.6 Crane platform locations to take images of each segment of the Me...
Figure 9.7 General scheme of approaching process of the US Inspection Tool o...
Figure 9.8 The robotic arm positions the ultrasonic sensor tool to perform c...
Figure 9.9 The mobile vehicle is able to estimate its position with the refl...
Figure 9.10 Schematics of the pure pursuit method for an Ackermann vehicle....
Figure 9.11 Autonomous inspection with ongoing traffic in the Metsovo tunnel...
Chapter 10
Figure 10.1 Trenchless methods costs vs. open-cut costs.
Figure 10.2 Comparison of emissions from the equipment used in both open-cut...
Figure 10.3 (a) HDD method, (b) pipe bursting, and (c) pipe ramming.
Figure 10.4 (a) Ground Mole robot.
Figure 10.5 BADGER concept. (a) Overview of the concept and drilling goal, (...
Figure 10.6 BADGER underground system.
Figure 10.7 BADGER surface rover.
Figure 10.8 Underground mapping: (a) area of interest, (b) automatic path pl...
Figure 10.9 BADGER underground robot path planning method. (a) A* planning a...
Figure 10.10 Drilling operation. (a) and (b) Entry points and start drilling...
Chapter 11
Figure 11.1 Example of various configurations of nondestructive testing/eval...
Figure 11.2 An omnidirectional NDT inspection robot. Design with two arms mo...
Figure 11.3 Self-alignment Mecanum wheel layout. Longitudinal wheel spacing ...
Figure 11.4 Stabilizing arm.
Figure 11.5 A cart with umbrella sensor arrangement NDT inspection robot.
Figure 11.6 A multicart with umbrella sensor arrangement NDT inspection robo...
Figure 11.7 Axial and circumferential coordinates of a 2.5D pipe thickness m...
Figure 11.8 Typical PEC sensing setup (a) and signals (b).
Figure 11.9 System block diagram depicting the four onboard major components...
Figure 11.10 Typical PEC scans of pipe sections; dense (a) and sparse (b).
Figure 11.11 A typical inspection plan supplied to the utility partner for t...
Figure 11.12 Various examples of remaining pipe wall thickness maps as measu...
Figure 11.13 Final 2.5D spool thickness maps attained from the field deploym...
Figure 11.14 Sparse and dense remaining pipe wall thickness data maps at two...
Chapter 12
Figure 12.1 The CRAFT robot.
Figure 12.2 3D CAD model showing sensing unit camera distance and angle sepa...
Figure 12.3 Physical prototype of the RGB-D sensing unit.
Figure 12.4 Left infrared image from the RGB-D sensor displays artifacts (wh...
Figure 12.5 Right infrared image from the RGB-D sensor showing reflections (...
Figure 12.6 RGB image from the RGB-D sensor showing pipe walls (gray) and ar...
Figure 12.7 Depth image from the RGB-D sensor shows the pipe wall and defect...
Figure 12.8 RGB image from the RGB-D sensor (located in the center) of the c...
Figure 12.9 CRAFT floatation system.
Figure 12.10 Cable drum.
Figure 12.11 Cable feeding mechanism.
Figure 12.12 Image showing the 3D pipe reconstruction with the 3D model of t...
Figure 12.13 Image showing the 3D reconstruction of the concrete pipe.
Figure 12.14 Remote console.
Figure 12.15 The CRAFT robot is being lowered into a 600 mm diameter manhole...
Figure 12.16 The CRAFT robot enters the nontraversable section of the concre...
Figure 12.17 3D reconstruction of a section of the nontraversable sewer pipe...
Figure 12.18 3D reconstruction of a 100 meter long section of the nontravers...
Chapter 13
Figure 13.1 An example of a confined tunnel space with partition plates, man...
Figure 13.2 (a) The overview of the WAuMBot, and (b) the robot moving in a c...
Figure 13.3 (a) A scissor lift unit design, and (b) free-body diagram of a s...
Figure 13.4 (a) Foot-pad design, and (b) toe designed to avoid rivets.
Figure 13.5 (a) Design of toes, and (b) toe positions on the footpad for riv...
Figure 13.6 (a) 3-DOF body design, and (b) free-body diagram of the 3-DOF me...
Figure 13.7 (a) A 6-DOF arm carrying maintenance tools, and (b) protective e...
Figure 13.8 Software architecture.
Figure 13.9 Process diagram.
Figure 13.10 Plane detection.
Figure 13.11 (a) A manhole in front of the robot, (b) a manhole detected by ...
Figure 13.12 Manhole detection method.
Figure 13.13 Rivet detection method.
Figure 13.14 Template-based rivet detection.
Figure 13.15 Mechanical sag of the robot (solid line: with mechanical sag; d...
Figure 13.16 Two-stage correction process. Solid lines: start pose, dotted l...
Figure 13.17 (a) Initial robot position, (b) Step 1, (c) Step 2, (d) Step 3,...
Figure 13.18 Motion control in areas with rivets: (a) Step 1, (b) Step 2, (c...
Figure 13.19 (a) WAuMBot in the lab test rig, (b) WAuMBot inside the still t...
Figure 13.20 Displacement and angular errors over 10 runs.
Chapter 14
Figure 14.1 Uses cases: Periodic detailed inspection and accurate 3D modelin...
Figure 14.2 Graph-based representation of the inspection problem (left) and ...
Figure 14.3 Power line located in the surroundings of flight test center ATL...
Figure 14.4 Fleet of heterogeneous UAVs selected for power-line inspection: ...
Figure 14.5 Plan computed for the inspection of a power line located in ATLA...
Figure 14.6 View of a transmission tower when the UAVs follow the computed p...
Figure 14.7 Scheme of the proposed LiDAR-based real-time segmentation method...
Figure 14.8 (Left) Original map obtained from two inspection experiments usi...
Figure 14.9 (Left) Resulting segmentation of a flight map classified by
Poin
...
Figure 14.10 Segmented map using the proposed method in an area where two UA...
Chapter 15
Figure 15.1 Offshore floating oil production, storage, and offloading. Sourc...
Figure 15.2 Partial view of an oil refinery. Note the number of vessels and ...
Figure 15.3 Liquid storage tanks.
Figure 15.4 Pressurized process vessel.
Figure 15.5 Process piping in a refinery.
Figure 15.6 Petrochemical facility. General view.
Figure 15.7 Inspection camera accessing a restricted area.
Figure 15.8 Robotic platform for inspection.
Chapter 16
Figure 16.1 Various defect types on the solar cell: seven types of solar cel...
Figure 16.2 The flow chart of the proposed framework: the aerial robot first...
Figure 16.3 NEST model: a research and energy hub part of
Swiss Federal Labo
...
Figure 16.4 Flow chart of the solar panel recognition: because of the shape ...
Figure 16.5 The trajectory of the aerial robot: the system starts from 1 and...
Figure 16.6 The solar panel detection with field data: (a) original image; (...
Figure 16.7 Dataset distribution.
Figure 16.8 Dataset processing.
Figure 16.9 Dataset augmentation: The original image and its augmented versi...
Figure 16.10 CNN architecture. There is an
rectified linear unit
(
ReLU
) laye...
Figure 16.11 Classification results of the trained model. Blue means the cor...
Figure 16.12 Confusion matrix of the evaluation dataset.
Chapter 17
Figure 17.1 Illustrations of the main categories of additive construction pl...
Figure 17.2 The timeline of the technological development of additive manufa...
Figure 17.3 Conceptual representations of aerial repair and aerial additive ...
Figure 17.4 (a) The Red-head Woodpecker and (b) the conceptual representatio...
Figure 17.5 The conceptual visualization of aerial additive manufacturing of...
Figure 17.6 Evolution of aerial platforms for aerial repair and manufacturin...
Guide
Cover
Table of Contents
Title Page
Copyright
About the Editors
Preface
Acronyms
Begin Reading
Index
End User License Agreement
Pages
ii
iii
iv
xv
xvi
xvii
xix
xx
xxi
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
IEEE Press
445 Hoes Lane
Piscataway, NJ 08854
IEEE Press Editorial Board
Sarah Spurgeon, Editor in Chief
Jón Atli Benediktsson
Anjan Bose
James Duncan
Amin Moeness
Desineni Subbaram Naidu
Behzad Razavi
Jim Lyke
Hai Li
Brian Johnson
Jeffrey Reed
Diomidis Spinellis
Adam Drobot
Tom Robertazzi
Ahmet Murat Tekalp
Infrastructure Robotics
Methodologies, Robotic Systems and Applications
Edited by
Dikai Liu
University of Technology SydneySydney, Australia
Carlos Balaguer
Universidad Carlos III de MadridSpain
Gamini Dissanayake
University of Technology SydneySydney, Australia
Mirko Kovac
Imperial College LondonLondon, UK
IEEE Press Series on Systems Science and Engineering
MengChu Zhou, Series Editor
Copyright © 2024 by The Institute of Electrical and Electronics Engineers, Inc.All rights reserved.
Published by John Wiley & Sons, Inc., Hoboken, New Jersey.Published simultaneously in Canada.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission.
Trademarks: Wiley and the Wiley logo are trademarks or registered trademarks of John Wiley & Sons, Inc. and/or its affiliates in the United States and other countries and may not be used without written permission. All other trademarks are the property of their respective owners. John Wiley & Sons, Inc. is not associated with any product or vendor mentioned in this book.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.
Library of Congress Cataloging-in-Publication Data
Names: Liu, Dikai, editor.
Title: Infrastructure robotics : methodologies, robotic systems and applications / edited by Dikai Liu [and three others].
Description: Hoboken, New Jersey : Wiley, [2024] | Includes bibliographical references and index.
Identifiers: LCCN 2023041014 (print) | LCCN 2023041015 (ebook) | ISBN 9781394162840 (hardback) | ISBN 9781394162857 (adobe pdf) | ISBN 9781394162864 (epub)
Subjects: LCSH: Robotics. | Infrastructure (Economics)
Classification: LCC TJ211 .I48144 2024 (print) | LCC TJ211 (ebook) | DDC 629.8/92–dc23/eng/20231101
LC record available at https://lccn.loc.gov/2023041014
LC ebook record available at https://lccn.loc.gov/2023041015
Cover Design: WileyCover Images: Courtesy of Dikai Liu; Courtesy of Prof. Mirko Kovac, Laboratory of Sustainability Robotics, Empa & Aerial Robotics Laboratory, Imperial College London. Video taken by Schwarzpictures.com; Carlos Balaguer
About the Editors
Dikai Liu received his BEng, MEng, and PhD degrees from the Wuhan University of Technology in 1986, 1991, and 1997, respectively. He currently holds the position of distinguished professor at the Robotics Institute of the University of Technology Sydney (UTS), Australia. His primary research interests lie in the field of intelligent robotics, with a specific focus on perception, human–robot collaboration, brain–robot interface, human–robot teaming, robot systems, and design methodology. Besides conducting fundamental robotics research, he has successfully translated his research into practical applications, including infrastructure maintenance, construction automation, manufacturing, and health/aged care. Prof. Liu has led the development of over 10 intelligent robotic systems designed for real-world applications. Examples include autonomous robots for steel bridge maintenance, bio-inspired climbing robots for inspection in confined spaces in steel structures, intelligent robotic co-workers for human–robot collaborative abrasive blasting, smart hoists for patient transfer, and autonomous underwater robots for underwater structure maintenance. Since 2006, his research has received numerous awards, including the 2019 UTS Medal for Research Impact, the 2019 ASME DED Leonardo da Vinci Award, the 2019 BHERT Award for Outstanding Collaboration in Research and Development, and the 2016 Australian Engineering Excellence Awards.
Carlos Balaguer received the BSc, MSc, and PhD degrees from Polytechnic University of Madrid in 1977, 1981, and 1983, respectively. Since 1996, he has been a Full Professor at University Carlos III of Madrid where he is the coordinator of the RoboticsLab, a research group in the field of intelligent robots. His main research topics are intelligent robots design and control, humanoid robots, healthcare and assistive robotics, soft robotics, manipulation, and locomotion planning. He was a member of the Board of Directors of the euRobotics (2015–2021), an association of European robotics with more than 300 affiliated organizations. He was also the President of the International Association for Automation and Robotics in Construction (IAARC) for the period 2001–2004. He participated in 29 competitive European Union projects and being the coordinator of several of them. He organized numerous important scientific events; among them, he was the General Chair of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS'2018), the biggest worldwide scientific conference in the field of Intelligent Robotics with about 5000 attendants. He received several awards, among them for the best book in Robotics by McGraw-Hill (1988); the best paper of the ISARC'2003 in Eindhoven (The Netherlands); IMSERSO's Award 2004 for assistive robots' research; the Industrial Robot journal Innovation Award of the CLAWAR'2005 in London (UK); Tucker-Hasegawa Award 2006 in Tokyo (Japan) for a major contribution in the field of Robotics and Automation in Construction; and FUE's Award 2014 for AIRBUS-UC3M Joint R&D Center.
Gamini Dissanayake is an emeritus professor at the University of Technology Sydney (UTS). He was the James N Kirby Distinguished Professor of Mechanical and Mechatronic Engineering at UTS until his retirement in 2020. He graduated in Mechanical/Production Engineering from the University of Peradeniya, Sri Lanka. He received his MSc in Machine Tool Technology and PhD in Mechanical Engineering (Robotics) from the University of Birmingham, England. He taught at University of Peradeniya, National University of Singapore, and University of Sydney before joining UTS in 2002. At UTS, he founded the UTS Centre for Autonomous Systems that grew to a team of 75 staff and students working in Robotics by 2020. His main contribution to robotics has been in Simultaneous Localization and Mapping (SLAM), which resulted in the most cited journal publication in robotics in the past 20 years. SLAM is the robotic equivalent of a human finding their way around in a city without GPS and maps, thus underpins many robot applications ranging from household vacuum cleaning robots to self-driving cars. He has also been involved in developing robots for a range of industry applications including cargo handling, disaster response, mining, infrastructure maintenance, and aged care.
Mirko Kovac received his BSc and MSc degrees in Mechanical Engineering from the Swiss Federal Institute of Technology in Zurich (ETHZ) in 2005. He obtained his PhD at the Swiss Federal Institute of Technology in Lausanne (EPFL) in 2010 and pursued a Postdoc at Harvard University until 2012. He is now director of the Aerial Robotics Laboratory and full Professor at Imperial College London. He is also heading the Laboratory of Sustainability Robotics at the Swiss Federal Laboratories for Materials Science and Technology (Empa) in Dübendorf in Switzerland. His research group focuses on the development of novel, aerial robots for distributed sensing and autonomous manufacturing in complex natural and man-made environments. Prof. Kovac's specialization is in robot design, hardware development, and multimodal robot mobility. He has received numerous awards, including the ICE Howard Medal in 2021, as well as several fellowships, including the ERC Consolidator Grant in 2022 and the Royal Society Wolfson Fellowship in 2018.
Preface
Maintaining civil infrastructure assets, including bridges, tunnels, water mains, power and telecommunication transmission towers, underwater wharf piles, and sewers, has traditionally been labor intensive and often hazardous for workers. The need for safe, efficient, and effective infrastructure maintenance has led to a desire, across industry sectors, to automate maintenance operations. Intelligent robots that can work either on their own or collaboratively with humans in a structural environment offer a highly promising solution to maintenance operations.
Significant progress has been made in developing robotic systems for civil infrastructure inspection and maintenance. Increasing numbers of students, practitioners, researchers, and professionals are becoming interested in this exciting field. The editors, who have been conducting infrastructure robotics research for over 20 years, think that a book that provides comprehensive, state-of-the-art information about infrastructure robotics will address an important gap in the literature, as there are currently very few books available on this topic. Internationally renowned researchers in infrastructure robotics have authored various chapters of this book.
The book is organized into two parts. Part I discusses the methodologies that enable robots to operate in civil infrastructure environments for inspection and maintenance. It includes an introduction to infrastructure robotics (Chapter 1), the design of infrastructure robotic systems (Chapter 2), robot perception, localization and SLAM (Chapter 3), machine learning and computer vision algorithms (Chapter 4), robotic coverage planning (Chapter 5), and physical human–robot collaboration for infrastructure maintenance (Chapter 6). Part I does not cover the fundamentals of robotics, such as kinematics, dynamics, and control because there are many textbooks that cover these topics
Part II presents 11 case studies that highlight different types of intelligent robotic systems designed for maintaining various civil infrastructures. These include climbing robots, underwater robots, wheeled robots, legged robots, and unmanned aerial vehicles (UAVs). Each case study focuses on a specific type of infrastructure and its corresponding robotic systems. The case studies cover a wide range of civil infrastructures, such as steel bridges (Chapter 7), underwater structures (Chapter 8), tunnels (Chapter 9), underground construction (Chapter 10), underground water mains (Chapter 11), wastewater pipes (Chapter 12), confined spaces in steel structures (Chapter 13), electrical power lines (Chapter 14), oil refineries (Chapter 15), and solar panels (Chapter 16). Additionally, Chapter 17 discusses aerial repair and manufacturing techniques. These case studies provide insights and lessons learned from the development and deployment of robots for practical applications. Readers can gain knowledge and understanding of the challenges, successes, and practical considerations associated with implementing robotic systems for infrastructure maintenance.
The intended readers of this book are researchers, engineers, and graduate students who are interested in design, control, and use of intelligent robotic systems for inspection and maintenance of civil infrastructures. Professionals in civil engineering, asset management, and project management may be also interested in reading this book. We hope that the readers would find this book interesting and useful in their research as well as in practical engineering work.
The editors would sincerely like to express their gratitude to the contributing authors for their professionalism as well as their commitment to the success of the book and to those researchers who contributed to the work included in the book. The editors would like to thank Mr. Dinh Dang Khoa Le for assisting the compilation of the book chapters. The editors would also like to express their appreciation for the support from the IEEE RAS Technical Committee for Robotics Research for Practicality.
Sydney, Australia
Dikai Liu
Carlos Balaguer
Gamini Dissanayake
Mirko Kovac
Acronyms
ASME
AmericanSociety of Mechanical Engineers
ASNT
American Society for Nondestructive Testing
AUT
automated ultrasonics, corrosion mapping or C-scan
API
American Petroleum Institute
CML
corrosion monitoring locations
LASER
Light amplification by stimulated emission of radiation
LIDAR
light detection and ranging
MFL
magnetic flux leakage
NDE
nondestructive examinations, or nondestructive evaluation
NDI
nondestructive inspection
RFID
radio frequency identification
UT
ultrasonic testing
Part IMethodologies
Part I consists of six chapters that discuss the methodologies enabling robots to operate in civil infrastructure environments for inspection and maintenance. Chapter 1 introduces infrastructure robotics and reviews examples of research in several topics. Chapter 2 discusses the design of infrastructure robotic systems, including the design process and examples. Chapter 3 presents the theoretical foundations for perception, localization, mapping, and simultaneous localization and mapping (SLAM). Chapter 4 discusses machine learning, computer vision algorithms and their applications in water utilities and transport. Chapter 5 presents methods for robotic coverage planning in civil infrastructure inspection and maintenance. Chapter 6 presents methodologies for physical human-robot collaboration and their applications in infrastructure maintenance. Part I does not cover the fundamentals of robotics, such as kinematics, dynamics, and control, because there are many textbooks that cover these topics.
1Infrastructure Robotics: An Introduction
Dikai Liu and Gamini Dissanayake
Robotics Institute, University of Technology Sydney, Sydney, NSW, Australia
1.1 Infrastructure Inspection and Maintenance
Civil infrastructure, such as bridges, tunnels, oil refineries, and pipelines, plays a vital role in both the economy and the community. Due to aging, environmental factors, increased loading, damages caused by human and natural factors, and/or inadequate maintenance, civil infrastructure is progressively deteriorating. Appropriate periodic inspection and maintenance are required to ensure that the designed life of service of civil infrastructure can be achieved or extended. The purpose of inspection, e.g., visual examination or testing, is to assess the condition of infrastructure and detect any defects that may affect the infrastructure's functionality. Maintenance involves activities of cleaning, painting, or fixing problems identified during inspection to keep the infrastructure in good condition. Typical civil infrastructure such as bridges and power transmission towers are large and complex structures. Manual inspection and maintenance require that people to work in confined spaces, at heights, near water, or near live and high-voltage power lines, leading to significant health and safety issues.
Bridges are essential in transport infrastructure worldwide. Bridge maintenance or replacement is one of the biggest expenditure items in transport infrastructure development and maintenance. For example, there are approximately 42,000 steel bridges in Europe, and 210,000 and 270,000 steel bridges, respectively, in the United States and in Japan [Balaguer et al., 2005]. Corrosion is the primary cause of failure in steel bridges [Hare, 1987], and is minimized by painting the steel structure. Inadequate maintenance may result in structural failures such as the Mississippi Bridge incident in Minneapolis that led to 13 fatalities, 145 injuries, and a replacement cost of US $234 million.
Truss joints, rivets, and box girders are common in steel bridges. One example is the Sydney Harbor Bridge (Figure 1.1a), which is a very complex structure. Thus, steel bridge inspection requires special equipment such as special lifts and scaffolds. Some areas may not be easily inspected by humans due to safety concerns (Figure 1.1b). A team is needed to support each inspector, which implies high cost and low productivity.
Figure 1.1 (a) Sydney Harbor Bridge, an example of a complex steel bridge structure; (b) A section of a steel bridge structure; (c) Grit-blasting operation.
Steel bridge coating maintenance consists of two procedures: (i) removing rust and old paint, and (ii) repainting. The most effective and efficient method of large-scale paint stripping is grit-blasting (Figure 1.1c), and herein lies the critical problem. Grit-blasting is extremely labor-intensive and hazardous and is arguably the most expensive operation needed in steel bridge maintenance. Workers have to not only spend long periods of time handling forces of 100N or above [Joode, 2004] but also need to take precautionary measures to avoid exposure to dust containing hazardous chemicals. Paints used until the 1980s in most steel bridges in Australia and much of the industrialized world contain red lead. As the long-term health damage due to exposure to lead is now obvious [Information Services, 1998], parts of a bridge being maintained need to be fully enclosed to avoid contamination of the environment and their potential health risk to the general public.
Confined spaces are common in civil infrastructure. Examples of confined spaces include box girders (Figure 1.2a,b) and pipelines. Inspection and maintenance inside a confined space are very dangerous for human workers. In many situations, for example inside the box girder shown in Figure 1.2c, accessing some parts of the confined space is difficult for a human inspector. Furthermore, as rescuing a human inspector in the event of an accident is near impossible, potential safety risk is therefore extremely high.
Inspection and maintenance of underwater infrastructure such as bridge piles, wharf piers, underwater pipelines, and offshore oil/gas platforms are also extremely challenging. These structures vary in shape and size and are normally covered with marine growth. For example, oysters and barnacles can grow up to 20 cm thick, obscuring the structure's geometry and condition (Figure 1.3a). A bridge may have a number of piles, which makes the area cluttered (Figure 1.3b). Deterioration and defects of underwater structures include local scour, deteriorated piles, cracks, and exposed steel reinforcements. In order to comprehensively inspect underwater infrastructure, marine growth must first be removed. This is currently done by human divers using either ultra-high-pressure water jet blasting or mechanical cleaning tools. Methodical planning is required for manual cleaning to ensure the safety of the divers and their support teams, especially because diving in turbid water and in high currents is dangerous. Given the low productivity and expense of having human divers performing such work, complete cleaning is rarely attempted. Rather, marine growth is removed from a strip of a few selected piles to allow partial inspection. Therefore, manual inspection of underwater infrastructure is essentially a sampling process, which is also an extremely dangerous, labor-intensive, and expensive exercise.
Figure 1.2 (a) An example of confined space with partition plates each with a standard-sized manhole; (b) A confined space with a diaphragm with spaces above and below; (c) A human inspector has to slide through a space under the diaphragm to access a part of the confined space.
Source: [Ward et al., 2014]
Figure 1.3 (a) Bridge piles covered by marine growth; (b) A bridge with a number of piles.
Truss structures such as electrical power transmission towers and telecommunication towers are another example of infrastructure. These towers can be about 100 m high. For inspection and maintenance, human workers must climb the towers, carry maintenance tools, and conduct inspection, rust removing, and painting. They often must resort to awkward poses during operation (Figure 1.4a,b).
Figure 1.4 (a) A team of human workers painting a transmission tower using brushes; (b) A human worker conducting waterjet cleaning at the top of a tower.
Supplementing manual labor with intelligent robotic aids has the potential to have a significant health, safety, and economic impact on infrastructure inspection and maintenance.
1.2 Infrastructure Robotics
Infrastructure robotics involves developing methodologies that enable robotic systems to construct, maintain, and repair various types of civil infrastructure. These include bridges, buildings, roads, tunnels, underwater structures, and underground water mains. Robotic systems can take various forms, such as legged climbing robots, wheeled robots, underwater robots, drones, and unmanned aerial vehicles (UAVs). The focus of this book is on infrastructure inspection and maintenance, excluding the exploration of robotics for the design and construction of civil infrastructure, which is a separate research topic.
It is reasonable to state that almost all robotics research has the potential to be applied in inspection and maintenance of civil infrastructure. Due to the nature and complexity of civil infrastructure and the specific requirements of inspection and maintenance operations, developing intelligent robotic systems for infrastructure has specific research and development challenges. These challenges include (but are not limited to) [Liu et al., 2014]:
Sensing technology and sensor networks that can be used in field infrastructural environments for perception and robot navigation, robot interaction with the environment or structural members, and assessment of surface and structural conditions;
Design of novel (including bio-inspired) mechanisms for robots to move, stay, and support themselves safely in complex environments such as steel bridges and truss structures;
Methodologies for real-time awareness of environments and operations, including navigation, localization, and map building;
Efficient algorithms for real-time robot motion planning and collision avoidance in complex 3D environments;
Design methodologies such as collaborative design and bio-inspired design;
Multi-robot systems and human-robot teaming;
Methodologies that facilitate safe and intuitive physical human-robot collaboration (HRC);
Robot learning and computer vision that allows robots to adapt to various structural and field environments and conduct various tasks safely.
Besides research challenges, there are many engineering challenges in developing robots for civil infrastructure inspection and maintenance. Examples include design of lightweight robots that can move and climb in compact and complex structures, high torque-to-weight ratio actuators that allow robots to have enough payload for performing maintenance tasks, and fail-safe robotic system design.
There is a significant amount of research on infrastructure robotics research and development of robotic systems for infrastructure maintenance. Examples include:
1.2.1 Inspection and Maintenance of Steel Bridges
One of the early works was the robotic system developed to remove corroded paint and rust from the surfaces of steel beams [Lorenc et al., 2000]. This system is composed of a large crane boom, an actuated platform, a robot arm, a vision system, and proximity sensors. It is suitable for small and open bridge structure. A CAD drawing of a bridge structure is assumed to be available and used to reduce the challenges in perception. A manipulator mounted to a crane was also used by Boeing for aircraft maintenance [Schmitz, 2003]. These types of robotic systems can be used for inspecting and maintaining large smooth surfaces but are not suitable for complex, compact, and enclosed areas.
An autonomous grit-blasting robot for surface preparation in steel bridge maintenance was developed and deployed in real applications (Figure 1.5) [Liu et al., 2008, Paul et al., 2010]. This robot consists of a mobile platform, a 6-DoF root arm mounted to the platform, a sensor package, and a grit-blasting tool mounted to the end-effector of the arm. When this robot is placed on a scaffolding platform under a steel bridge, it autonomously performs sensing and perception to get the geometric information on the environment around it, builds a complete 3D map of the environment [Sehestedt et al., 2013], localizes itself, identifies surface condition, plans its motion and blasting trajectories [Paul et al., 2013], and then starts blasting to remove rust and old paint. This robotic system was extensively tested and verified in a number of field trials on a steel bridge.
Figure 1.5 A robot for stripping rust and old paint using grit-blasting.
The Wall-pushing Autonomous Maintenance roBOT (WAuMBot), as can be seen in Figure 1.6, removes rust and old paint. It vacuums and paints inside a confined space in a steel bridge. This robot has four limbs, each with a scissor lift mechanism to push against the walls of the confined space. The robot body is a 3-DoF manipulator that allows the robot to move forward or backward. Using the sensors mounted on the robot and algorithms for navigation, mapping, planning, and control, the robot can operate in a confined environment fully autonomously.
1.2.2 Climbing and Wheeled Robots for Inspection of Truss Structures
Inspection of truss structures and confined spaces is challenging because it requires a robot that is able to move in all possible directions along truss structure members or inside a confined space. At the same time, the robot needs to maneuver sensors and tools on the surfaces of structural members or surfaces inside confined spaces. Due to the challenges of navigation in complex truss structures or inside confined spaces, UAVs may not be an effective solution. One option is to use a wheeled robot, but a wheeled robot needs to be able to safely support itself (against gravity) onto a truss structure or inside a confined space, climb and go through manholes, and negotiate obstacles such as rivet arrays and positive and negative corners between surfaces.
Figure 1.6 A Wall-pushing Autonomous Maintenance roBOT (WAuMBot) inside a still tunnel of the Sydney Harbor Bridge.
A more feasible strategy is to develop legged climbing robots for this application. Legged locomotion enables operation in difficult and complex terrains where the use of wheels is either inefficient or infeasible. Research on adapting legged locomotion for climbing started in the early 1990s with a strong focus on development of robots for specific applications [Berns et al., 2003]. With advances in robotics research, there has been increasing interest in the use of climbing robots. Examples include the SM2 robot designed to walk along the I-beam structure of a space station [Nechyba and Xu, 1994, 1995], a climbing robot that can move in complex 3-D metal structures [Balaguer et al., 2005, Balaguer, 2000], the Shady3D robot for climbing trusses [Yoon and Rus, 2007], the Stickybot – a gecko-inspired robot that climbs smooth vertical surfaces such as glass [Kim et al., 2008], the RiSE climbing robot [Spenko et al., 2008], and the climbing robots for maintenance and inspection of vertical structures [Schmidt and Berns, 2013].
Most of the climbing robots developed to date are for climbing large flat surfaces or cylindrical surfaces with a large diameter, including smooth walls of glass, brick, steel, or concrete. These robots have difficulties to transfer from one surface to another angled surface (e.g. from wall to ceiling), climb non-flat steel surfaces such as steelwork with rivet arrays, reinforcement plates, or intersections of structural members.
Commonly used adhesion mechanisms for legged or wheeled robots for climbing steel structures include grasping, vacuum, permanent magnet, and electromagnet. Permanent magnets are used in the climbing robotic system developed by Nguyen et al. [2020]. This robot has both a mobile mode and a transforming mode that allows it to adapt to a wide range of surfaces (e.g. flat, curved, or rough).
A novel flying-climbing mobile robot was designed for inspection of steel bridge structure [Pham et al., 2021]. This design combines a drone's maneuverability with the mobile robot's climbing capability, allowing the robot to fly and land to location in a steel structure and conduct inspection. Permanent magnets are used in this robot for adjusting the distance between the robot body and the steel surface, allowing the robot to switch from landing, taking-off, and moving modes.
Many researchers have worked on climbing mechanisms [Webster et al., 2017, Nguyen and Liu, 2017]. Prototype robots have been developed with some of them extensively tested or deployed in the field. One example is a climbing robot for inspection of steel bridges [Ward et al., 2014]. This robot consists of a 7DoF body and two footpads each with three magnetic toes (Figure 1.7). Once placed inside a steel structure, it autonomously explores the environment [Quin et al., 2016], navigates around, plans safe motions [Yang et al., 2016], localizes itself in the environment, goes through manholes, and collects visual information for condition assessment.
Figure 1.7 A climbing robot inspecting the steel condition inside a confined space tunnel.
1.2.3 Robots for Underwater Infrastructure Maintenance
Intelligent robots that could clean and inspect underwater infrastructure components, such as bridge piles, wharf piles, underwater pipelines, have significant safety, cost, and health benefits. Developing robots for inspecting underwater structures in harsh shallow underwater environments has many challenges that are caused by unpredictable water current, low visibility, and marine growths that obscure the objects' geometry and condition. Extensive research and development on remotely operated underwater vehicles (ROVs) [Jin et al., 2013] and autonomous underwater vehicles (AUVs) [Caccia et al., 2000] have been conducted globally. Most ROVs and AUVs are designed to operate at depths where water currents are relatively slow and consistent, for example for detecting submerged wrecks and rocks, and mapping underwater environments. Therefore, the aforementioned challenges make the use of ROVs problematic in near-surface and tidal environments. AUVs that have been designed for the ocean environment for survey, monitoring, and oceanographic research may not be suitable because they are often heavy and may not respond fast enough to water current.
Some of the ROVs and AUVs have manipulative abilities. These are typically not designed to operate at relatively shallow depths, nor in the intertidal zone where wave action can be strong and inconsistent [Papadopoulos and Kurniawati, 2014]. The near-surface and tidal environment is hostile. External disturbances can strongly affect the dynamics of a robotic system, particularly near fixed underwater structures. For cleaning and inspection of structures in shallow water environments, the robots must have perceptual capabilities under low visibility and potentially rapid platform motion. The control for robot motion needs to deal with thrusters that have complex nonlinear behavior, particularly in the presence of changing currents. Some seemingly simple operations easily achievable with a ground-based robotic system, such as holding position, can become very difficult for an underwater robot in such environments.
Figure 1.8 An underwater robot firmly attached to a bridge pile by grasping with its arms uses a waterjet to remove marine growth.
“Crawfish” is a recently developed underwater robot, modified from BlueROV2, for ultrasonic fatigue crack detection and coating of underwater structures. Hardware architecture and software architecture are discussed in Preuß et al. [2022], and Wendt et al. [2022], respectively. A snake eel-inspired multi-joint underwater robot was developed [Lyu et al., 2022] for inspection of undersea infrastructure. Due to its slim design, this robot has high movement flexibility and passing ability inside complex structures. An ROV was also designed [Kizor V et al., 2021] for biofoul cleaning and inspection of underwater structures such as bridge piles and dams.
Figure 1.8 shows an underwater robot developed for removing marine growth and inspection in harsh shallow water environments [Le et al., 2020]. This robot has a body with eight thrusters, four grasping arms, and a number of sensors. It is capable of autonomously navigating to a bridge pile, loosely grasping to a pile, moving around a pile for 3D mapping and marine growth identification, firmly grasping a pile to form a fixed body, and then removing marine growth using waterjet blasting. After the pile surface is cleaned, the robot collects visual information of the surface for inspection.
1.3 Considerations in Infrastructure Robotics Research
A number of specific aspects can be taken into consideration when developing intelligent robotic systems for infrastructure inspection and maintenance.
(1) The specific features of infrastructure environments such as similar structures and geometric shapes. As an example, truss structures are common in steel bridges, electrical transmission towers, telecommunication towers, and offshore oil/gas platforms. Another example is underwater structures such as bridge piles, wharf piles, underwater pipes. Taking these features into consideration and taking advantage of them will help robot design, and at the same time extend the scope of application of a robot.
(2) Robot design: legged robots are more suitable for applications in steel bridges and truss structures compared to wheeled robots. Flying robots (e.g. UAVs) are better for applications with more open spaces and do not require physical interaction with the structure. When a robot is required to physically contact the structure or environment and conduct operations on the surface of a structural member, e.g., removing rust from a steel beam, flying robots need to be able to land on a structure and then firmly grasp the structure member to support the robot and the tools, and resist reaction force from the tools, or wind force.
(3) Co-design: Cooperative design (Co-Design) refers to the creativity of researchers (or robot designers) and potential end-users working together in the design and development process [Sanders and Stappers, 2008]. With Co-Design, researchers, or robot designers view potential end-users as experts in the area of application and involve them in the design process. So, design decisions are made cooperatively.
For intelligent robots to be used in infrastructure maintenance, the application scenario (or work site) is often unique. The end-users' requirements are specific as well. The Co-Design approach is therefore particularly useful. By designing cooperatively with users who have firsthand knowledge of a specific work environment and work tasks, user-friendly results are likely to emerge. Taking a Co-Design approach can aid in reducing anxieties and increasing the levels of trust [Lie et al., 2012].
