Path Planning Software and Graphics Interface for an Autonomous Vehicle, Accounting for Terrain Features

摘要

A Navigation Test Vehicle (NTV) is being developed at the Center for Intelligent Machines and Robots at the University of Florida under the sponsorship of the Air Force Research Laboratory at Tyndall Air Force Base. This vehicle performs tasks to include surveying fields, laying mines, and teleoperation. The capability of the vehicle will be increased if its supporting software plans paths that take into account the terrain features. The objective of this work is to plan a survey path for a specified region taking terrain elevation data into account. A survey path is defined as a path that provides complete coverage of a defined polygonal region. This work is an extension of previous work that assumed that the region to be surveyed was flat. Now the characteristics of the vehicle, specifically the maximum allowable pitch and roll, will be accounted for when planning the path. Data that contains elevation points are available from government databases. Using three different programs the data are transformed from the vector format in which it is found to a Virtual Reality Modeling Language format in order to be viewed with a common web browser. Data are stored into a grid of m x n elements and the user selects the coordinates of the region that will be surveyed. The previous 2D algorithm is then executed in which the primary direction of the survey varies from zero to 180 degrees. For each path a cost is computed in which the cost is based on pitch and roll knowing the maximum that the vehicle can withstand. For the NTV the maximum roll and pitch angle is 15 degrees. If the pitch or roll is larger than 15 degrees a very high cost is associated with that part of the path. After each path is processed, the path with overall minimum cost is written to a file. If the minimum path contains any slopes that are inaccessible they are recorded as obstacles. The last step is an obstacle avoidance algorithm, which generates small sub-paths around the inaccessible regions.