Adaptive Autonomous Architectural Framework for Intelligent Autonomous Subsystems

在交互、行为和管理方面，越来越需要更好地理解自治、分布式和智能系统。假设所有自治的目的都是满足不矛盾的总体目标，发现解决在协作环境中运行的交互协作自治（ICA）的新方法是很有趣的。未来的自动驾驶汽车有望在日益复杂和恶劣的环境中运行，响应不断增长的人类需求，并始终保持高弹性和适应性的最佳运行。为了实现这些崇高的目标，自主性应该建立在开放的架构上，使用嵌入式工具和算法来测量、评估和最佳地响应其环境中的变化。自治之间的相互作用需要测量和评估。如果ICA之间的信任交换违反了所交换的信任度量所达到的期望，则可能导致自治隔离。考虑到未来的动态、敌对和复杂的环境，为了构建、观察、分析和与未来的大规模智能ica交互，重新思考我们如何构建和合并现代自治是很重要的。实现未来ICAs的好处的基本概念仍然是促进开放、动态和变化的体系结构，其中体系结构的动态响应其环境的动态。由于智能解决方案倾向于从上下文知识中推断智能，因此构建和分析上下文感知框架对于ICA模型的开发至关重要。尽管ica有广泛的分类，但大多数被调查的协作自治集中在相同或几乎相同的自治之间的交互上，本建议超越了这一限制。此外，该提案在寻求将机器学习应用于交互过程本身的价值方面仍然是独一无二的。该提案的主要影响是引入了一种全新的概念方法来研究自治架构，除了对信息和安全行业有益之外，还产生了社会效益。该提案有望帮助建立加拿大机器学习领域的研发领导地位，这对我们信息部门经济的增长至关重要，并对加拿大整体经济和高素质人才（HQP）的培训产生影响。

There is a growing need to develop better understanding of autonomous, distributed, and intelligent systems in terms of interaction, behaviour, and management. It is intriguing to discovering novel ways to address Interacting Cooperating Autonomies (ICA) operating in a collaborative environment, assuming all autonomies aim to satisfy non-contradicting overarching goals. Future autonomies are expected to operate in increasingly complex and hostile environments, respond to growing human needs, and maintain optimal operations with high resilience and adaptability at all times. To achieve those lofty goals, autonomies shall be built over open architectures with embedded tools and algorithms to measure, evaluate, and optimally respond to changes in their environments. Interactions between autonomies need to be measured and evaluated. Exchange of trust between ICA may lead to autonomy isolation if it violates expectations reached by the exchanged trust metric. In order to build, observe, analyze, and interact with future massive intelligent ICAs considering their dynamic, hostile, and complex environments it is important to rethink how we construct and merge modern autonomies. The fundamental concept to realizing the benefits of future ICAs remains in facilitating open, dynamic, and varying architecture where the dynamics of the architecture respond to the dynamics of its environment. Since intelligent solutions tend to deduce intelligence from contextual knowledge, building and analyzing context-awareness frameworks are crucial and vital to the development of ICA models. While there are wide classifications of ICAs, most surveyed collaborating autonomies focus on interaction between identical or almost identical autonomies, this proposal goes beyond this limitation. Further, this proposal remains unique in seeking the value of applying machine learning to the interaction process itself. The principal impact of this proposal is the introduction of a fundamentally new conceptual approach to studying autonomous architecture and yields societal benefits in addition to benefits to the information and security industries. the proposal is expected to help build research and development leadership in the machine learning areas in Canada which is pivotal to the growth of our information sector's economy and has an impact on the Canadian economy in general and the training of Highly Qualified Personnel (HQP).