Coverage Path Planning for Inspecting Large Structures using Unmanned Aerial Vehicles (UAVs)

  • Randa Almadhoun

Student thesis: Master's Thesis


Inspection of large structures using autonomous Unmanned Ariel Vehicles (UAVs) has emerged as a challenging and unique robotic application. Inspecting structures (e.g. bridges, ships, wind turbines, aircrafts) is considered a hard task for humans to perform, and of critical nature since missing any detail could a?ect the structure's performance and integrity. Furthermore, structure inspection is a time and resource intensive task that should be performed as efficiently and accurately as possible. Different structure inspection algorithms have been surveyed addressing various challenges depending on the environment, the shape of the structure, and the level of the required details. The two main challenging research topics related to inspection are coverage path planning, and workspace 3D reconstruction. Coverage path planning is the process of computing a feasible path encapsulating a set of waypoints through which the robot must pass in order to completely scan the structure of interest. This thesis makes different contributions related to the coverage path planning process for large structures inspection application. In this thesis, we propose a coverage planning algorithm for inspecting large complex structures using a UAV. The proposed method follows a model based coverage path planning approach which generates an optimized path that passes through a set of admissible waypoints sampled uniformly to cover a complex structure. The algorithm provides a prediction of the covered volume percentage by using an existing model of the complex structure as a reference. We developed a search space coverage path planner (SSCPP) with a heuristic reward function that exploits our knowledge of the structure's model, and the UAV's on-board sensors' specifications to generate optimal paths that maximize coverage and accuracy and minimize travelled distance and turning angle. Furthermore, we developed an extension to SSCPP, Dynamic Search Space Path Planner (DSSCPP), which generates a set of admissible waypoints dynamically with different discretization levels. DSSCPP generates an optimal path by evaluating expected Information Gain (IG) taking the advantage of the structure's 3D model. The proposed work also supports the integration of multiple sensors and it generates a resolution driven coverage paths in order to provide an accurate 3D reconstruction based on the used sensors' specifications. Additionally, critical components of the algorithm were accelerated utilizing the Graphics Processing Unit (GPU) parallel architecture. A set of experiments were conducted in a simulated environment using different models to test the validity of the proposed algorithm. Different evaluation criteria such as search duration, path distance, number of selected viewpoints, and accuracy were used to verify and test the effectiveness of the proposed work.
Date of AwardNov 2016
Original languageAmerican English
SupervisorTarek Taha (Supervisor)


  • Coverage Path Planning
  • Viewpoints Generation
  • Robotic Inspection
  • 3D Reconstruction.

Cite this