The University of British Columbia

The paper addresses the problem of scheduling radar dwells in multi-function phased array radar systems. The timing constraints of radar tasks are usually modeled by the minimal and maximal temporal distance between any two consecutive dwells of a task. Such timing constraints make it difficult for traditional real-time scheduling techniques to provide predictable timing guarantees, without over-consuming resources. We propose a novel approach to model the dwells as periodic real-time tasks. The periods of the tasks are synthesized by the minimal and maximal temporal distance constraint of the dwells. The synthetic periods allow the template-based scheduling algorithm to compute efficient dwell schedules with low overhead. We evaluate the algorithms via extensive simulations. Simulation results show that this algorithm can significantly improve resource utilization, compared with traditional dwell scheduling algorithms.