| The GeoSurv II Unmanned Aerial Vehicle (UAV) project is focused on the construction of a UAV for geosurveying. It is intended to be able to fly with minimal operator interaction and as such must be able to detect, track and avoid obstacles automatically. Classically, obstacle detection and tracking are tightly linked to each other and to the sensors. This makes for an efficient system but poses certain problems if it becomes necessary to change one of the components, or to evaluate different approaches. The Obstacle Tracking and Classification (OTC) framework system developed in this research provides tracking and classification capabilities for a mobile sensor platform such as the GeoSurv II UAV and allows prediction and tracking methods to be evaluated and characterized so that an engineer can choose the most appropriate methods to solve their tracking problem. The OTC testbench system includes a scripted simulator for conducting the evaluation and characterization tests. A testing methodology and test scripts are presented and two prediction methods are tested as a demonstration. |
