An assessment of the development of a tracking system using concurrent Ada

The authors describe the development of a representative real-time system from problem analysis to final implementation on a parallel machine. The problem selected for implementation is a multiple-model Kalman filter tracking algorithm. The system was designed using the Layered Virtual Machine/Object-Oriented Design methodology proposed by K.W. Nielson and K. Shumate (Designing Large Real-Time Systems with Ada, Multiscience Press, 1988) and was implemented in Ada on a 16-processor Encore Multimax. The authors discuss speedup results and include commentary on the methodology used and the advantages of using Ada in terms of programmer productivity, portability, and reusability of the code. The present study is of interest in connection with realizing the goals of the Strategic Defense Initiative