Luận án: Hệ thống robot giám sát đường ống tự hành dựa trên cảm biến của Jong-Hoon Kim

Louisiana State University LSU Digital Commons LSU Doctoral Dissertations Graduate School 2011 Sensor-based autonomous pipeline monitoring robotic system Jong-Hoon Kim Louisiana State University and Agricultural and Mechanical College, hoonywiz@gmail.com Follow this and additional works at: https://digitalcommons.edu/gradschool_dissertations Part of the Computer Sciences Commons Recommended Citation Kim, Jong-Hoon, "Sensor-based autonomous pipeline monitoring robotic system" (2011). LSU Doctoral Dissertations.edu/gradschool_dissertations/2566 This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Doctoral Dissertations by an authorized graduate school editor of LSU Digital Commons. For more information, please contactgradetd@lsu.

SENSOR-BASED AUTONOMOUS PIPELINE MONITORING ROBOTIC SYSTEM Dissertation Submitted to the Faculty of Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Doctor of Philosophy in The Department of Computer Science by Jong-Hoon Kim B., Seoul National University of Science and Technology, Seoul, Korea 2005 M., Louisiana State University, Baton Rouge, USA 2008 December, 2011 Acknowledgements I would like to thank the many people who made this dissertation possible. I will remember and appreciate their contributions forever. First and foremost, I would like to deeply thank Prof. He has always encouraged me to overcome the many difficulties faced during my research, and has supported me with his unshakeable confidence.

Only with his support could I complete this dissertation. I would also like to thank the other members of my dissertation committee, Prof. Jianhua Chen, Prof. Gerald Baumgartner and Prof.

Arash Dahi Taleghani. They have provided valuable feedback and interest- ing perspectives on the ideas contained in this dissertation. Once again I would like to thank them all for being on my committee. I wish to express sincere thanks to Prof.

Brygg Ullmer for supporting many things in the Tangviz Lab generously as well as his advice. And also I would like to deeply appreciate Prof. Noureddine Boudriga for guiding my research and sharing many good memories with me. And, I would give special thanks Prof.

Sukhamay Kundu for being my teaching mentor. I would like to thank my friends, YoungPyo Jeon, Gokarna, Lohit, Rajesh, Forrest Osterman, Chris J Michael, Karthik, Bhaskar, Balachandran, Bharat, Srikanth, Monika, Sandeep, Robert Kelly and many others for making my life in Baton Rouge enjoyable. Finally, I would give great appreciation to my family. They always gave me unquestioning faith and encouraged me in every time.

Especially, I would love to express to Keyyoung Park, my wife, my gratitude for her enormous support and devotion. Without her sacrifices, I could not have completed this long journey. So, I would like to dedicate this dissertation to her. ii Table of Contents Acknowledgements.

ii List of Tables. v List of Figures .2 Motivation and Objective .3 Navigation within Pipelines .4 Pipeline Robot Classification .2 Autonomy Based Classification. 16 3 State of the Art .1 Sensor-Based Technologies for Pipeline Monitoring .2 Robot Agent-Based Technologies for Pipeline Monitoring .1 System Requirements for an Efficient Monitoring System .2 High-Level Description of the System .2 McRAIT: Multiple Channel Redundant Array of Independent RFID Tag .1 The McRAIT Architecture .2 Functions of the McRAIT Controller .3 The McRAIT Fault-tolerance .4 Design of a McRAIT System .3 HPMS: High Performance Mobile Sensor .1 Design of a HPMS .4 FAMPER: Fully Autonomous Mobile Pipeline Exploration Robot .2 Motion Planning of the Robot .3 Design of a FAMPER. 50 5 A Technique for Incident and Sensor Localization .1 Maximum Range Estimation .2 The McRAIT-based Localization.

58 6 Prototyping of the System .1 Prototype of the Robot Agent .1 Performance Evaluation of the Robot Agent .2 Performance Analysis of the Robot Agent .2 Performance Evaluation on the System .2 Limitations, Challenges, and Open Issues. 91 iv List of Tables 3.1 Comparison of various pipeline monitoring techniques. 25 v List of Figures 2.1 Different types of robots .2 Different types of pipelines .3 Oil pipeline inspection robots .7 Many troubles in pipeline failure .8 Mechanical classification of pipeline robots .9 Typical methods of steering in branch .10 Examples of non autonomous robots [RAUSCH Electronics USA LLC] .11 Examples of semi autonomous robot .12 Examples of fully autonomous robots .1 McRAIT system design .4 An application of the SPRAM system .5 The McRAIT architecture .6 High Performance Mobile Sensor (HPMS) interface board .7 The autonomous pipeline exploration robot .8 Operational architecture of the robot .9 Control architecture of the robot .10 Different types of motion in elbow, T-, and Y-branches .11 Motion planning at 90 degrees elbow, where H to H indicates Horizontal to Hori- zontal, V to H indicates Vertical to Horizontal, and H to V indicates Horizontal to Vertical motion .12 Motion planning at T-branch .13 Depiction of the pipeline surface contact by robot wheels at T-branch .14 Motion planning at Y-branch .15 Side view of the stretchable caterpillar .16 Sectional view of the robot agent .1 Sensor localization within the pipeline.1 Components of the robot .2 Electrical architecture of the robot .3 Controlling and RF video system .1 Experimental pipeline layout .2 Types of motion at 90 degrees elbow, where (a), (b) show H to V, (c), (d) show H to H, and (e), (f) show V to H motion .3 Types of motion at T-branch, where (a)-(d) show H to V, (e)-(h) show H to H, and (i)-(l) show V to H motion .4 Types of motion at Y-branch, where (a)-(c) show H to V, (d)-(f) show H to H, and (g)-(i) show V to H motion .5 Motion singularity problem, where (a) depicts singular motion and (b) depicts suc- cessful motion .6 Self-adjusting from motion singularity position to successful motion position .7 Illustration of a pipeline topology and a prototype robot mission .8 Illustration of a pipeline system used in experiments .9 Average occupancy of the marker storage for different parameter settings and Hop = 6 .10 Average occupancy of the marker storage for different history settings, and n = 100, s/s = 20, and Hop = 6 .11 Measured RFID entries concentration for 12 incidents using 4-tags-McRAITs with values of n = 50, H = 5, s/s = 10, and Hop = 6 .12 Maximum limited radio range between a reader and a marker for different trans- mitted power .13 Maximum relative error vs. threshold angle plot .14 The relation between distance d and diameter L of the pipeline .15 Effects of the number of incidents on the average error made on the reported distance 79 7.16 Comparison of the average distance the robot agent travels to find reported inci- dents using three strategies.

80 viii Abstract The field of robotics applications continues to advance. This dissertation addresses the com- putational challenges of robotic applications and translations of actions using sensors. One of the most challenging fields for robotics applications is pipeline-based applications which have become an indispensable part of life. Proactive monitoring and frequent inspections are critical in main- taining pipeline health.

However, these tasks are highly expensive using traditional maintenance systems, knowing that pipeline systems can be largely deployed in an inaccessible and hazardous environment. Thus, we propose a novel cost effective, scalable, customizable, and autonomous sensor-based robotic system, called SPRAM System (Sensor-based Autonomous Pipeline Moni- toring Robotic System). It combines robot agent based technologies with sensing technologies for efficiently locating health related events and allows active and corrective monitoring and mainte- nance of the pipelines. The SPRAM System integrates RFID systems with mobile sensors and autonomous robots.

While the mobile sensor motion is based on the fluid transported by the pipeline, the fixed sensors provide event and mobile sensor location information and contribute efficiently to the study of health history of the pipeline. In addition, it permits a good tracking of the mobile sensors. Using the output of event analysis, a robot agent gets command from the controlling system, travels inside the pipelines for detailed inspection and repairing of the reported incidents (e., damage, leakage, or corrosion). The key innovations of the proposed system are 3-fold: (a) the system can apply to a large va- riety of pipeline systems; (b) the solution provided is cost effective since it uses low cost powerless fixed sensors that can be setup while the pipeline system is operating; (c) the robot is autonomous and the localization technique allows controllable errors.

In this dissertation, some simulation ex- periments described along with prototyping activities demonstrate the feasibility of the proposed system. ix Chapter 1 Introduction 1.1 Pipeline Everything from water to crude oil even solid capsule is being transported through millions of miles of pipelines in the United States. The pipelines are vulnerable to losing their functionality by internal and external corrosion, cracking, third party damage and manufacturing flaws. If a small water pipeline bursts a leak, it can be a problem but it usually does’t harm the our environment.

However, if a petroleum or chemical pipeline leaks, it can be a environmental and ecological disaster. We can see many US pipeline accidental reports at the National Transportation Safety Board’s Internet site [7]. Thus, for keeping pipelines operating safely, periodic inspections are performed to find cracks and damage before they become cause for serious concern. When a pipeline is built, many inspection methods can be used to evaluate its quality such as visual, X-ray, magnetic particle, and ultrasonic.

These inspections are performed as the pipeline is being constructed so gaining access to the inspection area is not problem. Most pipelines are buried except some pipelines like the Alaskan oil pipeline. Once the pipeline is buried, it is undesirable to dig it up for any reason. Therefore, many remote visual inspection equipments to assess the condition of the buried pipe have been developed.

For inspection and recovery action of damaged pipeline, robotic crawlers of 1 all shapes and sizes have been developed to navigate the pipeline. The video signal is typically fed to a truck where an operator reviews the images and controls the robot.2 Motivation and Objective Proactive monitoring and frequent inspections are critical to maintain pipeline health, as gas, oil, water, and sewer pipelines have become an indispensable part of life. Hence, the continuous proactive monitoring and maintenance system for these pipelines is essential, however, deploy- ment, monitoring, and maintenance of them should remain cost effective, scalable, and easily customizable. A number of technologies, which are proposed and available to monitor, control, and maintain diverse types of pipelines, have still remained in unsatisfying those requirements due to their limitations.

In this dissertation, we aim at designing a cost-effective pipeline maintenance and monitoring system. Such a system would allow frequent inspection, early detection of problems, controllable- error localization, and planned recovery measures. To accomplish those goals, we believe that a monitoring system for pipelines should combine sensor technologies, which are well suited for event localization, and robotic techniques, which allow proactive and corrective monitoring. In addition, we argue that a more efficient technique for locating objects and incidents should be integrated in such systems.

Such a technique should use built-in objects that are powerless, easy to add, and densely deployed. Based on the hypothesis, we have developed a novel method, called SPAMMS (Sensor-based Autonomous Pipeline Monitoring and Maintenance System) in [35], which combines sensor and robotic techniques with radio-frequency identification (RFID) [67] technology for efficient event localization and proactive and corrective monitoring of a large spectrum of pipeline types. Besides providing efficient localization of objects and incidents, our technique have achieved the efficient localization with low cost and controllable errors. However, the SPAMMS system can be significantly improved by efficient localization tech- 2 nique and enhanced major components; Fixed Sensor, Mobile Sensor, and Robot Agent.

Thus, in this dissertation, we firstly propose a RFID-based localization technique, applicable to any kind of pipeline network. It allows controllable localization errors in the sense that the threshold it reaches are controlled by a fraction of the distance separating two successive localization objects.