SBIR Phase I: COGNITIVE INTELLIGENCE CAPABILITIES IN REAL-TIME EDGE COMPUTING AND DISTRIBUTED INTERNET OF THINGS (IOT) PLATFORMS

这个小企业创新研究（SBIR）第一阶段项目的更广泛的影响/商业潜力是展示边缘原生物联网（IoT）平台的认知运行能力，该平台在增量操作智能、实时监控和自主功能方面提供了显著的收益。随着边缘计算能力渗透到工业物联网应用领域，以克服集中式计算（延迟、带宽和单点故障）、受限的硬件（计算和存储）资源和数据可用性的限制。拟议的技术可以随着不断变化的网络物理环境进行调整，扩展态势感知，防范网络安全威胁，最大限度地降低扩展网络和云基础设施的成本，并在关键任务生态系统（如智能电网监控、远程患者监测、自动驾驶车辆和国防武器系统）中提供极端的响应能力。该小企业创新研究（SBIR）第一阶段项目将研究认知物联网（IoT）运行时，以增强边缘本地M2M/IoT解决方案的流程自动化和分析能力。边缘计算的最新趋势创造了服务机会，以迎合更接近操作环境的实时用例。边缘节点上最常见的预定义基于规则的逻辑和结构化智能无法提供可伸缩性、自主性、互操作性和智能控制操作。人工智能(AI) /机器学习（ML）模型通常在批处理模式下的云或多条款服务器基础设施上表现更好。本发明缩小了现有的差距，并通过引入认知人工智能模型进一步推进了分布式边缘计算能力，该模型引入了嵌入式运行时内增量数据的语义学习，并通过连接环境中的交互开发推理。跨边缘节点点对点协作的研究与开发克服了资源约束的限制，实现了工作流的执行。运行时架构模拟了人类神经系统的解剖层和神经运动操作的系统协调。提出的分析模型和分布式智能的设计模式将在边缘引入更智能的决策支持系统的现实实现，该系统可以适应不断变化的操作环境，并通过协作学习引入多种智能操作。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to demonstrate the cognitive runtime capabilities of edge native Internet of Things (IoT) platform that offers significant gain in incremental operational intelligence, real-time monitoring, and autonomic functions. As the edge computing capabilities infiltrate into industrial IoT application domains to overcome the limitation of centralized computing (latency, bandwidth, and single point of failure), constrained hardware (compute and storage) resources, and data availability. The proposed technology may adjust with the changing cyber-physical environment, extend situational awareness, protect against cyber security threats, minimize the cost of scaling network and cloud infrastructures, and deliver extreme responsiveness within mission critical ecosystems such as smart grid surveillance, remote patient monitoring, autonomous vehicles, and defense weaponry systems.This Small Business Innovation Research (SBIR) Phase I project will research a cognitive internet-of-things (IoT) runtime that enhances process automation and analytical capabilities of edge native M2M/IoT solutions. The recent trends of edge computing have created service opportunities catering to real-time use cases closer to the operating environment. The most common predefined rule-based logic and structured intelligence on edge nodes fails to deliver scalability, autonomy, interoperability, and intelligent control operations. The artificial intelligence (AI) / Machine Learning (ML) models usually perform better on cloud or multiclause server infrastructure in a batch mode. This invention closes the existing gap and further advances the distributed edge computing capabilities by introducing Cognitive AI models that introduce semantic learning with incremental data from within embedded runtime and develop reasoning through interactions within the connected environment. The research and development on Per-to-Peer collaboration across the edge nodes overcomes the limitations of constrained resource and executes workflows. The runtime architecture simulates the human nervous system’s anatomical layers and systematic coordination of neuro-motor operations. The proposed design patterns of analytical models and distributed intelligence will introduce a realistic implementation of smarter decision support systems on the edge that could adapt to ever changing operational contexts and introduce diverse intelligent operations through collaborative learning.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.