Sensor scheduling using Intelligent Agents

As a result of the versatility of modern phased array radar systems and the complexity of the modern battlefield, it is necessary to design planning algorithms that allow the full exploitation of the sensor’s capabilities. Optimising the radar’s performance is shown to belong to the class of NP-complete problems, thereby excluding the posibility to find an algorithm capable of providing solutions in polynomial time or shorter. Based on this hypothesis, alternative approaches for designing the scheduler need to be explored. This paper presents an innovative radar scheduler based on an object-oriented method of generating tasks, which are prioritised by a Multi-Agent system. The resulting planning provides a near-optimum allocation of sensor time and energy to all maintenance tasks required for the Air Picture. The paper concludes with the presentation of a basic IA task scheduler and the key performance issues that will be addressed in the near future.