Dissertation > Excellent graduate degree dissertation topics show

Research on Mobile Localization Techniques for Wheeled Restaurant Service Robots

Author: YuQingXiao
Tutor: ZhaoYanZheng
School: Shanghai Jiaotong University
Course: Mechanical and Electronic Engineering
Keywords: restaurant service robots iterated extended Kalman particle filter multi-information fusion region matching binocular vision
CLC: TP242
Type: PhD thesis
Year: 2013
Downloads: 339
Quote: 0
Read: Download Dissertation

Abstract


Since the late20th century, many scholars have begun to study the servicerobots in indoor environment, and made great achievements in the theory andpractice. As the mobile positioning technology is the key of the navigation andpath planning, localization has become a hot topic and drawn greater attention.To enhance the reliability of the restaurant service robot system, the mobilepositioning technology of service robot has been profoundly studied in thispaper. And the localization accuracy has been improved to ensure that therestaurant service robot could serve and deliver dishes for the customers inrestaurant environment. Then begin with the establishment of restaurant servicerobot systems, based on it indoor mobile localization is done, those as follows.1. A type of wheeled restaurant service robot system is established based onthe wheeled robot principle and analyze the mechanical system and sensorsystems. Establish the kinematic model and odometer model of the restaurantservice robot and the kinematic model of the manipulator. According to thesimulation of the manipulator kinematic model, define the emphasis and difficulty of the serving dishes task of the robot in less degrees of freedom.According to the characteristics of the sensor systems, establish the observationmodels of binocular and global visual sensors, and lay the necessary foundationfor environment perception and positioning techniques.2. To enhance the positioning accuracy of the restaurant service robot, animproved iterative extended Kalman particle filter algorithm (PF-UIEKF) isproposed. The algorithm optimizes the iterative extended Kalman filter (IEKF)using the Levenberg-Marquardt method, to amend the problem that IEKF issometimes slow to convergence and resolve the impact of the initial state on thefilter effect. The algorithm uses the maximum posteriori estimates of theoptimized IEKF to generate the importance probability density function, andtakes full advantage of its handling capacity of the nonlinear problem to producethe better reference distribution. Simulate the various particle filter algorithmswith the typical nonlinear model, comparing with the simulation results of theseparticle filters, verify that the state estimation accuracy of the proposedalgorithm is much better than the other particle filters. A global visuallocalization algorithm based on PF-UIEKF is proposed to position the mobilerobot, and the algorithm is applied to restaurant service robot and experimentsverify that it meets the positioning accuracy and real-time requirements ofservice robot mobile navigation. 3. To improve the positioning accuracy and reliability of the restaurantservice robot, based on the study and discussion of multi-information fusiontheory, centralized, distributed and federal IEKF multi-information fusionalgorithms are proposed to fuse the multi-sensor information in nonlinearenvironment. Respectively simulate the proposed multi-information fusionalgorithms using the typical nonlinear model to verify their state estimationperformance. A localization method based on federal IEKF multi-informationfusion is proposed to fuse the information of odometer, visual landmark andglobal vision, and the practical application verifies that the algorithm couldeffectively improve the positioning accuracy and reliability.4. Based on the mean-shift color segmentation and the SIFT featurematching technology, a regional dense matching algorithm is proposed. Thealgorithm does image segmentation with mean-shift method, extracts andmatches the feature points using the SIFT operator. According to whethersegmentation region contains the SIFT feature points, separately match eachimage region using the different methods to obtain the more accurate disparitymap. Simulate and analyze the proposed matching algorithm using theprofessional stereo image pairs to verify the accuracy, reliability and stability ofthe matching algorithm. A fine positioning algorithm of robot target pose basedon binocular vision is proposed to realize mobile robot moving towards the target and the precise positioning of the service robot relative to the targetfeature. The method is applied to restaurant service robot for experimentalverification and the service robot can successfully complete the self-servingdelicious task.In this thesis, a restaurant service robot platform is developed, and thekinematic model of the robot and sensor observation model are also studied. Theaccuracy and reliability of the mobile positioning have been obviously enhancedby researching the particle filter, the multi-information fusion and the imagematching technology. And the proposed algorithms provide a favorablereference for indoor service robot localization. The proposed algorithms havebeen successfully applied in the restaurant service robot. Many practicalapplications verify that the proposed algorithms can effectively enhance thepositioning accuracy and stability of the service robot and possess a practicalreference value.

Related Dissertations

  1. Design of Intelligent Object Tracking Luggage Cart System Based on Binocular Vision,TP391.41
  2. Study of Vision Sensing System for All-Position Autonomous Welding Robot with Passing Obstacles Capability,TP242
  3. Research on Target Detection of Panoramic Infrared Search System,TP391.41
  4. Research on Extraction Technologies of Environmental Characteristics Based on Parallel Binocular Vision,TP391.41
  5. Highway traffic incident detection modeling and application research,U491.116
  6. Vision Based Simultaneous Localization and Mapping of A Mobile Robot in Unknown Environment,TP242
  7. Research of Fault Diagnosis for Power Transformer Based on Information Fusion,TM407
  8. Research of Obstacle Detection and Path Planning for Unmanned Aerial Vehicle,V275.1
  9. Based on GPS/INS High Accuracy Navigation Technology Satellite Research,P228.4
  10. Study on Determination of Pork Quality by Near Infrared (NIR) Spectroscopy and Computer Vision,TP391.41
  11. Research on Intelligent Detection Technology of Electrical Fire Based on Multi-information Fusion,TM76
  12. Robot Navigation Based on Vision,TP242
  13. Study on the Method of Apple Intelligent Grading Based on Multi-information Fusion,TP29-AG
  14. Research on Smart Garments Oriented Medical Diagnosis Application Based on Multi-fusion Information,R319
  15. Research on Object Depth Measurement System in Three-dimension Vision Technology,TP391.41
  16. Binocular vision - based moving target positioning,TP391.41
  17. Design and Study on Remote Supervision Network of Communication Center,TP277
  18. Light pen binocular vision measurement system,TP391.41
  19. Research of Character Recognition and Label in Information Process of News Videos,TP391.41
  20. The Applied Study of Binocular Stereo Vision Technique in Measuring Three Dimension Position,TP391.41
  21. Research and Implementation of CORBA distributed RTI server - based,TP393.03

CLC: > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Robotics > Robot
© 2012 www.DissertationTopic.Net  Mobile