Optimizing Sensor Locations to Improve the Worst Case Detection Performance of Sensor Detection Systems

Consider designing a sensor system to detect an emitter at an unknown location. For this scenario, it is common to place the sensors in a grid pattern. This paper shows that we can improve the worst case probability of detection over the grid pattern by optimizing the sensor locations. Using Torczon's local search algorithm, our results indicate that this algorithm can improve the worst case probability of detection by up to nearly 23% in the scenario considered.