Path planning and redundancy resolution of structure-climbing redundant robots

This thesis addresses critical challenges in the planning and control of biped structure-climbing redundant robots that navigate three-dimensional truss structures such as bridges, building skeletons, and power transmission towers. Maintenance tasks in these environments, traditionally performed by...

Descripción completa

Detalles Bibliográficos
Autor: Fabregat Jaén, Marc
Tipo de recurso: tesis doctoral
Fecha de publicación:2026
País:España
Institución:Universidad Miguel Hernández de Elche
Repositorio:REDIUMH. Depósito Digital de la UMH
OAI Identifier:oai:dnet:rediumh_____::ae852274ae7a6f89744836884de8594b
Acceso en línea:https://hdl.handle.net/11000/39841
Access Level:acceso abierto
Palabra clave:Self-motion manifolds
Global redundancy resolution
Redundant manipulators
Motion planning
Obstacle avoidance
Path planning
Climbing robots
Redundant robots
Truss structures
Workspace
CDU::6 - Ciencias aplicadas::68 - Industrias, oficios y comercio de artículos acabados. Tecnología cibernética y automática
CDU::6 - Ciencias aplicadas::62 - Ingeniería. Tecnología::621 - Ingeniería mecánica en general. Tecnología nuclear. Electrotecnia. Maquinaria
CDU::5 - Ciencias puras y naturales::51 - Matemáticas
Descripción
Sumario:This thesis addresses critical challenges in the planning and control of biped structure-climbing redundant robots that navigate three-dimensional truss structures such as bridges, building skeletons, and power transmission towers. Maintenance tasks in these environments, traditionally performed by human workers, present significant safety risks that can be mitigated through robotic solutions. Two fundamental challenges are addressed in the present thesis: path planning for structure navigation and redundancy resolution for optimal joint trajectory generation. First, we develop a hierarchical path planning algorithm that decomposes complex 3D exploration into manageable subproblems: a first planner determines the global path in the complete structure, while a second planner derives the local path in each visited face of the structure. This approach is generalizable, and efficiently handles transitions between structural faces while avoiding collisions. Second, we present three complementary approaches to redundancy resolution. The first method is based on Feasibility Maps (FMs), which provide a representation of the redundant-time space, on which we deploy an RRT-based algorithm to generate feasible joint trajectories while optimizing performance criteria. Building upon this, we extend the concept to Augmented Feasibility Maps (AFMs), which incorporates the task dimension into the redundancy resolution process, effectively tackling both the task-trajectory generation and redundancy resolution problems simultaneously. For the third approach, we developed a global redundancy resolution framework based on Self-Motion Manifolds (SMMs), which characterize the complete set of inverse kinematics solutions.We perform a spatial and topological analysis on the transformations of the SMMs throughout the task trajectory, and derive optimal joint trajectories from this analysis. All of these methods are able to incorpórate task constraints, such as joint limits and obstacle avoidance, into the redundancy resolution process. Finally, we introduce a novel homotopy-based method for computing SMMs that addresses limitations of existing approaches. This method effectively identifies all disjoint manifold components without requiring closed-form kinematic solutions, demonstrating superior performance compared to traditional continuation methods. The contributions are validated through extensive simulation tests, showing significant improvements over existing methods in terms of computational efficiency, path quality, and practical applicability. This research advances the state of the art in structure-climbing robotics and redundancy resolution, bringing these climbing robots closer to practical deployment in industrial inspection and maintenance applications.