Self-Aware Processing for Adaptive Resource Optimization of Advanced Computing Systems

Impact Technologies, in collaboration with the Georgia Institute of Technology and our commercialization partner Raytheon, propose to develop and demonstrate a suite of core self-aware processing technologies that can autonomously adapt available computing resources based on the real-time processing requirements and state awareness of the overall computing environment. Building on the team’s expertise in real-time prognostics and health monitoring for digital electronic systems, our proposed approach will integrate state-of-the-art electronic system fault detection with higher level cognitive modeling techniques to provide the constituent capabilities for advanced, reconfigurable computing. The self-aware system will utilize artificial intelligence technologies such as probabilistic neural networks as underlying software agents within a distributed reasoning process similar to cognitive capabilities in human thinking processes. These systems will balance resources, efficiencies, and mission priorities to efficiently use system resources to dynamically and effectively meet changing operational requirements. In doing so, it is anticipated that the performance of the systems can be enhanced several times when compared for such critical metrics as power consumption, fault-tolerance, and reliability.