Skip to main content

Cross-layer cognitive optimization tools & methods for the lifecycle support of dependable CPSoS

Periodic Reporting for period 1 - CPSoSaware (Cross-layer cognitive optimization tools & methods for the lifecycle support of dependable CPSoS)

Reporting period: 2020-01-01 to 2021-06-30

We envision the CPSoSaware solution as a framework that will span horizontally and vertically on the overall CPSoS architecture thus providing a holistic, cognitive and decentralized way of designing, running and decommissioning CPS components of the system or even the CPSoS as a whole. Since such an endeavour is highly complex and hard to manage, in CPSoSaware we use AI and ML assistance in order to make the above-mentioned procedures feasible and pragmatic.

Ο1: Design Cognitive, Reconfigurable and autonomous CPSoS orchestration, that support CPSoS full lifecycle (requirements, design, test, operate and decommissioning) and Design Operation Continuum
Ο2: Provide a Decentralized, cooperative, autonomic control and management that is resilient, fail-safe and adaptable to unforeseen physical and cyber-changes
O3: Structure a CPSoS Design approach that can be modelled and Simulated at system level
O4: Provide vertically and horizontally Secure and Trusted Designs (Security -by-Design) and Provide Runtime Cybersecurity monitoring to protect against cyber-threats and respond to attacks
O5: Consider throughout the CPSoS lifecycle Human users and operators and provide Extended Reality solutions that increase their situational awareness (Human in the Loop)
Ο6: Integrate the CPSaware various tools into a unified solution and test it in two distinct use cases (connected autonomous cars and Manufacturing processes with Robotics and Human interaction
O7: To define evidence-based business and financing models along with a business plan for the post-project sustainable exploitation of the CPSoSaware framework.

Μeeting the aforementioned objectives will open up new market slots in Europe. It is therefore expected that the implementation and the success of this project will contribute to the development of European technical leadership in the CPS market.
O1: This Objective is linked to WPs 3, 4 and 5 and covers issues related to Cognition and ReconfigurationTo match the objective regarding Cognition, we have performed a review of best practices and tools for monitoring and autonomously understand whether CPSoS matches the functional and non-functional requirements expressed in KPIs (Section 7 of D1.1 Submitted on M6). Regarding the configuration, identification of appropriate tools to support reconfiguration eg. using Xlinx Vitis platform and the latest FPGA of Xilinx has been made.
O2: This objective is directly linked with WPs 3,4 and 5and covers issues related to scene analysis, 3D perception, user state monitoring, localization, path planning and their application in the automotive and manufacturing pillar. More importantly, the cooperative and decentralized fusioning of extracted information from different modalities, towards solving problems related to CPSs localization, user state evaluation and path planning have been thoroughly investigated.
O3: We conducted extensive charting of alternatives for simulation models (as part of T2.2) to extend and utilize in the CPSoSAware’s inter and intra communication modeling needs. A working demonstrator of the distributed event synchronization capability was implemented. In addition, a LLVM-based profiler, extracting the instruction characteristics of the OpenCL SW modules, has been developed. An initial planning for a HiTL supported architecture has been made, while the use of Application Simulators (e.g Carla) as a physical environment model and simulation mechanism is going to be deployed in order to feed synthetic data on the HiTL modelling approach.
O4: i) A review of security by design and run time security monitoring state of the art was performed.ii) KPI definition for security non functional requirements iii) Mechanisms have been specified regarding the security primitives to ensure security goals iv) An architecture and prototype implementation of the CPSoSaware Security Runtime monitoring tool has been created.
O5: This objective is directly linked with the implementation of increasing driver’s situational awareness through Augmented Reality occlusion rendering. Moreover, part of the work linked with this objective focused on the problem of users' motion tracking, lying in industrial environments, in order to increase the level of safeness while they co-work or collaborate with robots sharing the same workspace.
O6:The first year of the project few integration activities were made since integration is primarily done in WP5. However, there were several preparatory activities on how to link the project solution/outcomes with the two use case pilots. More specifically, efforts have been made to: i) Create appropriate use case requirements and scenarios, ii) Produce/implement plan for the experiments of the quantification campaigns, iii) Define preliminary version of the assessment protocol, iv) Define testing scenarios, identify the expected behaviour of the system in the two use-cases.
O7: A preliminary version on the IPR document issues which has been provided led by RTC, in line with the Market analysis and as a first yet thorough approximation planning to of the Exploitation strategy has been performed. During the next reporting period, the above document issues will be enriched in line with the Exploitation Plan. Several Business Model Canvas (BMC) have been included in the corresponding deliverables.
In WP2 we research on communication Models focusing on WiFi, Bluetooth and ZigBee communication technologies. Required interfaces between NS3 as the main intra-communication simulator and other CPSoSaware framework components and potential 3rd party components were defined. This work was reported in D2.4 that is under internal review and predicted to be submitted in time to the EU portal. In terms of IP modeling and component interfacing, a previously initiated component interface called AlmaIF v1 extended to a new version 2 and adopted to the needs of the CPSoSAware project. It was successfully used as a hardware IP wrapper to interface the new IP blocks (programmable and non- programmable) to the OpenCL-based distributed heterogeneous software platform and demonstrated by two different means of generating the wrapped IP: An ASIP one (based on openasip.org toolset) and Vitis HLS generated one.
In WP3 we research on Scene Analysis using Accelerated Deep Learning, on Driver’s drowsiness estimation based on facial analysis, on Cooperative localization and path planning, on Occluded objects and safe spaces to increase Situational awareness, on hardware assisted Security token and security sensos and on Run-time Management of Fault Data and Instruction Caches.
In WP4 research was done in optimization of HW-SW partitioning based on non-functional requirements using a CPSoSaware developed generic methodology for HW-SW partitioning modelling and optimization w.r.t non-functional requirements. Also, research was done the design of a intracommunication simulation workflow along with tools that will be used for the evaluation of various network configurations. In WP5 we focus on continuing the work of WP2 and WP3. The complete DSM application, as well various security components, have been transferred in the target platform. Also, research related to human-in-the-loop situational awareness using extended reality tools has been made in WP5.e System AI-Assisted maintenance by exploring the deployment of semantic data integration mechanisms
CPSoSAware Concept Architecture and Partner Roles