Narrow Results

The successful navigation of an autonomous vehicle heavily depends on the accuracy of the parameters of the vehicle and sensor models, which are determined before the vehicle is in use. One of the main sources of error is the present prior way of determining parameters, partly because the currently accepted procedure for determining the parameters is not sufficiently accurate and partly because the parameters vary as the vehicle is driven.

This chapter presents an application of multi-objective evolutionary algorithms to the sensor and vehicle parameter determination for successful autonomous vehicles navigation. Followed by the multi-objective formulation, a general framework for multi-objective optimization, two types of search methods to find solutions efficiently and a technique for selecting a final solution from the multiple solutions are proposed. The proposed parameter determination technique was applied to an autonomous vehicle developed by the authors, and an appropriate parameter set has been obtained.

The following sections are included:

• The SLAM Problem and Its Applications
  • Description of the SLAM Problem
  • Applications of SLAM

• Summary of SLAM Approaches
  • EKF/EIF based SLAM Approaches
  • Other SLAM Approaches

• Key Properties of SLAM
  • Observability
  • EKF SLAM Convergence
  • EKF SLAM Consistency

• Motivation

• Book Overview

The following sections are included:

• Information Matrix in the Full SLAM Formulation

• Information Matrix in the Conventional EIF SLAM Formulation

• Meaning of Zero Off-diagonal Elements in Information Matrix

• Conditions for Achieving Exact Sparseness

• Strategies for Achieving Exact Sparseness
  • Decoupling Localization and Mapping
  • Using Local Submaps
  • Combining Decoupling and Submaps

• Important Practical Issues in EIF SLAM

• Summary

The following sections are included:

• The D-SLAM Algorithm
  • Extracting Map Information from Observations
  • Key Idea of D-SLAM
  • Mapping
  • Localization

• Structure of the Information Matrix in D-SLAM

• Efficient State and Covariance Recovery
  • Recovery Using the Preconditioned Conjugated Gradient (PCG) Method
  • Recovery Using Complete Cholesky Factorization

• Implementation Issues
  • Admissible Measurements
  • Data Association

• Computer Simulations

• Experimental Evaluation
  • Experiment in a Small Environment
  • Experiment Using the Victoria Park Dataset

• Computational Complexity
  • Storage
  • Localization
  • Mapping
  • State and Covariance Recovery

• Consistency of D-SLAM

• Bibliographical Remarks

• Summary

The following sections are included:

• Structure of D-SLAM Local Map Joining Filter
  • State Vectors
  • Relative Information Relating Feature Locations
  • Combining Local Maps Using Relative Information

• Obtaining Relative Location Information in Local Maps
  • Generating a Local Map
  • Obtaining Relative Location Information in the Local Map

• Global Map Update
  • Measurement Model
  • Updating the Global Map
  • Sparse Information Matrix

• Implementation Issues
  • Robot Localization
  • Data Association
  • State and Covariance Recovery
  • When to Start a New Local Map

• Computational Complexity
  • Storage
  • Local Map Construction
  • Global Map Update
  • Rescheduling the Computational Effort

• Computer Simulations
  • Simulation in a Small Area
  • Simulation in a Large Area

• Experimental Evaluation

• Bibliographical Remarks

• Summary

The following sections are included:

• Structure of Sparse Local Submap Joining Filter
  • Input to SLSJF - Local Maps
  • Output of SLSJF - One Global Map

• Fusing Local Maps into the Global Map
  • Adding into the Global Map
  • Initializing the Values of New Features and in the Global Map
  • Updating the Global Map

• Sparse Information Matrix

• Implementation Issues
  • Data Association
  • State and Covariance Recovery

• Computer Simulations

• Experimental Evaluation

• Discussion
  • Computational Complexity
  • Zero Information Loss
  • Tradeoffs in Achieving Exactly Sparse Representation

• Summary

The following sections are included:

• Appendix A Proofs of EKF SLAM Convergence and Consistency

• Appendix B Incremental Method for Cholesky Factorization of SLAM Information Matrix

• Bibliography

The following sections are included:

• Preface

• Acknowledgments

• Contents

Current applications of robotics is distinguished from more traditional automation by the focus on machines that operate in relatively unstructured, difficult and often hazardous environments. The past decade has seen the deployment of a number of robotic systems in highly challenging application domains such as mining and agriculture. In particular, the potential safety, cost and health impacts from the use of robotics aids for periodic inspection and maintenance of civil infrastructure has led to a significant expansion of research activity in “infrastructure robotics”. Two of the prerequisites for deploying a robot in the field are the ability to acquire and maintain a representation of the environment, and efficient motion planning. In scenarios where a machine and a human must collaborate to perform a task, an intuitive user interface, a joint understanding of abilities and joint management of task execution are also essential. This talk will summarise current research on these key competencies that underpin robot deployments in unstructured environments, with the focus on simultaneous localisation and mapping, and human robot collaboration. Future potential and key challenges in infrastructure robotics will also be presented together with a review of the research activities in this field at the Centre for Autonomous Systems, University of Technology Sydney.