| Description |
This thesis presents the development and testing of the Voronoi-based path generation (VPG) algorithm for a mobile robot. The algorithm solves a variation of the coverage path planning (CPP) problem where complete coverage of an area is not possible due to path length limits caused by energy constraints on the robot. The problem of covering an area with a mobile robot has numerous applications, and this thesis focuses on sensorbased coverage - specifically, mapping chemical concentrations using an autonomous chemical-sensing aerial robot, such as a quadrotor helicopter (quadcopter) robot. There is little previous work on optimizing partial coverage paths for mobile robots with energy constraints, and this thesis aims to fill some of that gap. The presented algorithm works by modeling the path as a chained mass-spring-damper system and leveraging useful properties of Voronoi diagrams to generate a potential field that moves path waypoints to near-optimal configurations while maintaining path length constraints stemming from energy limitations of the robot. Simulation tests quantifying algorithm runtime show a linear time complexity with respect to the number of path waypoints, and tests in both convex and nonconvex areas demonstrate the method can generate paths in both. Comparison tests with other path generation methods demonstrate the VPG algorithm has runtimes 1-2 orders of magnitude better than direct optimization and a conceptually similar method proposed by Soltero et al. (2013), and coverage paths with costs 64-83% closer to the optimal path cost than a lawnmower-style coverage path and 45-55% closer than the Soltero method. Physical experiments demonstrate the applicability of the VPG method to a physical aerial robot, and comparisons between real-world results and simulations show that costs of the generated paths are within 1-3% of each other, implying that analysis performed in simulation will hold on real-world mobile robotic platforms, assuming the robot is capable of closely following the path and a good energy model is available. |
