Creating accurate tracks of multiple airborne targets from multiple sensors, in real time, can be a computationally demanding process. Our approach is to perform hypothesis testing based upon the traditional method of Maximum Likelihood, but within a distributed filtering environment. This results in a large reduction in the number of floating point computations required to generate the complete set of likelihood function values. This final report describes results obtained over a 6 month Phase I project. The primary mathematical operation performed by the distributed filters is matrix triangularization. Thus, this research focused on understanding algorithms for performing this operation, as well as their parallelization.