A BOOLEAN-ALGEBRA APPROACH TO MULTIPLE SENSOR VOTING FUSION

A technique is described for integrating multiple-sensor data using a voting fusion process that combines the individual sensor outputs. An important attribute of the method is the automatic confirmation of the target by the fusion processor without the need to explicitly determine which sensors and what level of sensor participation are involved. A three-sensor system, with multiple confidence levels in each sensor, is discussed to illustrate the approach. Boolean algebra is used to derive closed-form expressions for the multiple sensor-system detection probability and false alarm probability. Procedures for relating confidence levels to detection and false alarm probabilities are described through an application example. The paper concludes with the hardware implementation for the sensor system fusion algorithm.