A system concept expected to improve sonar multiple target detection and classification performance in severe directional interference environments is presented. Adaptive nulling of undesired strong interference sources and detection of desired weak sources are achieved by utilizing constrained adaptive processing techniquestogether with conventional processing techniques. The adaptive processing functions are incorporated in the overall system architecture in a manner which maintains the computational complexity and storage requirementslinearly proportional to the product of the number of interference sources and the number of adaptive weights per interference source. In addition, a detailed discussion is provided concerning the importance of certaina priori assumptions on constrained adaptive algorithm performance.