SHF: Medium: Conquering MPSoC Complexity with Principles of a Self-Aware Information Processing Factory

SHF：中：利用自我意识信息处理工厂的原理征服 MPSoC 复杂性

• 批准号: 1704859
• 负责人: Nikil Dutt
• 金额: $ 90万
• 依托单位: University of California-Irvine
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-10-01 至 2023-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1704859&HistoricalAwards=false
• 关键词: SHF; Medium; Conquering; MPSoC; Complexity

Self-aware systems are reactive and proactive, and capable of thinking outside the box, thereby coping with unforeseen situations in a dependable manner, unlike existing systems that are designed subject to deterministic procedural specifications. Self-awareness is a critical feature for future cyber-physical systems which must cope with scenarios that were not conceived when they were developed. This project will explore directions for overhead-efficient self-aware HW/SW systems while guaranteeing critical properties such as real-time behaviors, resilience, security, and energy efficiency. The project's study of self-awareness principles to enable autonomy in the face of dynamic and unforeseen changes will significantly impact a wide range of cyber physical application domains that critically depend on high-performance, low power and dependable computing. These systems play a pivotal role for conquering key societal challenges in healthcare, safety, transportation and industrial automation. The parallel collaborative nature of this project with German partners at TU Braunschweig and TU Munich will train a diverse group of students and researchers and enrich their learning experiences by exposing them to the challenges of delivering autonomous IT services in different societal and international contexts. The project will also motivate undergraduate and underrepresented students to pursue research and graduate studies. Students will benefit tremendously from exposure to and involvement in the design of new design and development paradigms for introducing cognition into a multi-tiered CPS and IoT infrastructure.The project introduces the notion of a self-aware Information Processing Factory (IPF) as a step towards autonomous Multi-Processor System-on-Chip (MPSoC) platforms in Cyber-Physical Systems (CPS) and the Internet of Things (IoT). The objective is to synergistically merge the situational flexibility of self-aware/self-organizing emergence with the predictability of hierarchical top-down control in a holistic multi-layer approach, resembling an Information Processing Factory. Sensory information at multiple scales -- on-chip, run-time systems, applications -- will be monitored, fused, and interpreted intelligently to assess health status, mitigate adverse effects, and optimize system goals of the MPSoC using a novel hierarchy of self-aware control loops that can both manage complexity, as well as achieve guarantees. Additionally, a unique scalable distributed run-time verification architecture will enable hardware-supported cross-layer monitoring of parametric specifications. The resulting approach synergistically covers the multi-tiered abstraction hierarchy from low-level VLSI circuits to runtime OS and dynamic application software. The project complements efforts at TU Braunschweig and TU Munich in Germany, and leverages research efforts on self-awareness and self-organization at the hardware and software layers of cyber physical multicore platforms.

自我意识系统是反应性和主动性的，能够跳出框框思考，从而以可靠的方式应对不可预见的情况，而不像现有的系统是根据确定性的程序规范设计的。自我意识是未来网络物理系统的一个关键特征，它必须科普开发时没有设想到的场景。该项目将探索开销高效的自我感知硬件/软件系统的方向，同时保证关键属性，如实时行为，弹性，安全性和能源效率。 该项目对自我意识原则的研究，使自主面对动态和不可预见的变化将显着影响广泛的网络物理应用领域，这些应用领域严重依赖于高性能，低功耗和可靠的计算。这些系统在克服医疗保健、安全、运输和工业自动化等关键社会挑战方面发挥着关键作用。该项目与布伦瑞克工业大学和慕尼黑工业大学的德国合作伙伴的平行合作性质将培养一批多样化的学生和研究人员，并通过让他们在不同的社会和国际环境中提供自主IT服务的挑战来丰富他们的学习经验。 该项目还将激励本科生和代表性不足的学生进行研究和研究生学习。 学生将从接触和参与设计新的设计和开发范例中受益匪浅，这些范例将认知引入到多层CPS和物联网基础设施中。该项目引入了自我感知的信息处理工厂（IPF）的概念，作为迈向网络物理系统（CPS）和物联网（IoT）中自主多处理器片上系统（MPSoC）平台的一步。 我们的目标是协同合并的情景灵活性的自我意识/自组织的出现与层次自上而下的控制在一个整体的多层次的方法，类似于一个信息处理工厂的可预测性。多个尺度的传感信息-片上，运行时系统，应用程序-将被智能地监控，融合和解释，以评估健康状况，减轻不良影响，并使用一种新型的自感知控制回路层次结构优化MPSoC的系统目标，该层次结构既可以管理复杂性，又可以实现保证。 此外，一个独特的可扩展的分布式运行时验证架构将使硬件支持的参数规格的跨层监控。由此产生的方法协同涵盖了从低级别的VLSI电路到运行时操作系统和动态应用软件的多层抽象层次结构。该项目补充了布伦瑞克工业大学和德国慕尼黑工业大学的工作，并利用了在网络物理多核平台的硬件和软件层的自我意识和自我组织的研究工作。

项目成果

Emergent Control of MPSoC Operation by a Hierarchical Supervisor / Reinforcement Learning Approach

通过分层管理器/强化学习方法对 MPSoC 操作进行紧急控制

• DOI: 10.23919/date48585.2020.9116574
• 发表时间: 2020
• 期刊: Automation & Test in Europe Conference & Exhibition (DATE
• 作者: Maurer, Florian;Donyanavard, Bryan;Rahmani, Amir M.;Dutt, Nikil;Herkersdorf, Andreas
• 通讯作者: Herkersdorf, Andreas

CHIPS-AHOy: a predictable holistic cyber-physical hypervisor for MPSoCs

CHIPS-AHOy：适用于 MPSoC 的可预测整体网络物理管理程序

• DOI: 10.1145/3229631.3229642
• 发表时间: 2018
• 作者: Mück, Tiago;Fröhlich, Antonio A.;Gracioli, Giovani;Rahmani, Amir M.;Reis, João Gabriel;Dutt, Nikil
• 通讯作者: Dutt, Nikil

Predicting Failures in Embedded Systems Using Long Short-Term Inference

使用长短期推理预测嵌入式系统中的故障

• DOI: 10.1109/les.2020.3007361
• 发表时间: 2021
• 期刊: IEEE Embedded Systems Letters
• 影响因子: 1.6
• 作者: Zhang, Tianyi;Seo, Minjun;Donyanavard, Bryan;Dutt, Nikil;Kurdahi, Fadi
• 通讯作者: Kurdahi, Fadi

Design methodologies for enabling self-awareness in autonomous systems

在自主系统中实现自我意识的设计方法

• DOI: 10.23919/date.2018.8342259
• 发表时间: 2018
• 期刊: Automation and Test in Europe Conference (DATE’18
• 作者: Sadighi, Armin;Donyanavard, Bryan;Kadeed, Thawra;Moazzemi, Kasra;Muck, Tiago;Nassar, Ahmed;Rahmani, Amir M.;Wild, Thomas;Dutt, Nikil;Ernst, Rolf
• 通讯作者: Ernst, Rolf

SPECTR: Formal Supervisory Control and Coordination for Many-core Systems Resource Management

• DOI: 10.1145/3173162.3173199
• 发表时间: 2018-03
• 期刊: Proceedings of the Twenty-Third International Conference on Architectural Support for Programming Languages and Operating Systems
• 作者: A. Rahmani;Bryan Donyanavard;T. Mück;Kasra Moazzemi;A. Jantsch;O. Mutlu;N. Dutt