Self-optimization of task allocation in shared man/computer control

A technique for autonomously allocating control between an "intelligent" aiding system and a human operator is presented. A criteria function based on operator utilities and task related performance measures is used to adaptively optimize the allocation under changing operational conditions. The technique incorporates the following functions: (1) Adaptively estimate operator utilities; (2) Search for, and obtain, an optimum control allocation policy; (3) Adapt criteria function parameters to time variations in operational conditions.