Proceedings PaperA Real Time AI Approach to Discrimination Boost Phase Optical Sensor Systems in SDI Architectures
|Format||Member Price||Non-Member Price|
Interest has been rekindled in the potential utility of Ballistic Missile Defence (BMD) systems 1,2 and their ability to enhance the existing NATO strategic defence posture 3,4. Whereas in the past BMD systems have been thought to be vulnerable to relatively simple offence countermeasures, technological developments that have occurred over the past 20 years offer the potential to solve some of the main criticisms that have bedeviled BMD research since its inception in the early 1950s. One of the key areas where dramatic developments have taken place is in the field of electro-optic sensor technologies where developments in device sensitivity and packing density offer new solutions to threat detection, tracking and discrimination that complement data traditionally associated with radar based systems. Analysis has shown 5 that optical sensor systems can make a significant contribution to threat analysis in the boost and mid course phases of flight of ballistic missile systems. In the Boost phase the large amounts of energy contained within the plume of a ballistic missile system provides a signature that must be detected against cloud and earth backgrounds - necessitating viewing from space. The process of detection is complicated by reflected sunlight and other sources of false alarms. The optical sensor systems must therefore be adaptable and capable of reasoning about the location of the signatures, their persistence and temporal variations. Much of this processing is ideally carried out at the sensor system - in order to eliminate false alarms and reduce the communications bandwidths required to transfer the sensor data to centralised early warning and battle management facilities. In the mid course phase optical sensor systems can be used to detect warm objects against the background of deep space. These sensor systems can form tracks on these objects that can be merged into 3D tracks as data from individual sensor systems are combined. As closely spaced objects are resolved by sensor systems feature data can be extracted on individual objects that can be used by the defence system to attempt to discriminate between warheads, decoys and other penetration aids. This paper reviews work that has arisen from joint US SDIO and UK MOD research programmes into the feasibility of Theatre Missile Defence (TMD) systems that would be suitable for deploy ment and operation in a European theatre. The paper focuses on the problems of threat classification and discrimination in TtD systems and highlights the role of optical sensors. The paper discusses the integration of data derived from optical and radar sensors 6 and expands upon work previously reported into the use of an Artificial Intelligence (AI) approach to object classification and discrimination.