Acoustic shooter localization systems are being rapidly deployed in the field. However, these are standalone systems—either wearable or vehicle-mounted—that do not have networking capability even though the advantages of widely distributed sensing for locating shooters have been demonstrated before. The reason for this is that certain disadvantages of wireless network-based prototypes made them impractical for the military. The system that utilized stationary single-channel sensors required many sensor nodes, while the multi-channel wearable version needed to track the absolute self-orientation of the nodes continuously, a notoriously hard task. This paper presents an approach that overcomes
the shortcomings of past approaches. Specifically, the technique requires as few as five single-channel wireless sensors to provide accurate shooter localization and projectile trajectory estimation. Caliber estimation and weapon classification are also supported. In addition, a single node alone can provide reliable miss distance and range estimates based on a single shot as long as a reasonable assumption holds. The main contribution of the work and the focus of this paper is the novel sensor fusion technique that works well with a limited number of observations. The technique is thoroughly evaluated using an extensive shot library.