Feature Extraction of Low Dimensional Sensor Returns for Autonomous Target Identification
Christopher W. Lum, Rolf T. Rysdyk · AIAA Guidance, Navigation, and Control Conference and Exhibit · 2008
This work considers algorithms for maritime search and surveillance missions. During these type of missions, an agent searches for a target using its various sensors. Performing target identiflcation and classiflcation of sensor returns is a crucial component of the mission. This paper investigates a system to process returns from a low dimensional sensor and automatically classify the data. This system uses an algorithm that employs the sensor and motion model of the agent to augment the limited data from the sensor. Several difierentiating features are then extracted and used to train various classiflers and machine learning algorithms.