Using performance maps to build a real time decision making framework

摘要

This paper presents a real time decision making framework used to create a decision process based on multiple performance maps composed of volatile functional values to assist in complex decisions for highly nonlinear systems by allowing human visualization and direct algebraic computations. The framework, based on map-based process is intended to extend other decision tools used in the synthesis process, such as neural networks, fuzzy logic and classical control and optimization by archiving knowledge learned and data uncertainty in geometric representation during the process execution. Time is critical to provide reliable and exact response as the systems become more complex. This real time decision making framework maintains the structure of previous research projects performed by the Robotics Research Group where the selected maps are combined into envelopes as functional decision surfaces, but including the real-time constraints in the decision process. The potential and applicability of this technology is used in different knowledge domains, ranging from ground vehicles navigation, unmanned aerial vehicles, health care management systems, and intelligent actuators.