Technical - Control & Monitoring

Lead partner: Leicester University (Ian Postlethwaite)

Contributing partner: Loughborough University (John Andrews)

Reconfiguration

A key objective of this research stream was to make optimal use of information provided by the monitoring and prognosis systems developed in the C&M research group processing the battlefield sensor data, in order to perform asset reconfiguration, together with the control and communication operations related to the implementation of the reconfiguration functionality. This research in reconfigurable systems of systems must rely on a theoretical framework defining the dynamics of the entities leaving and joining the NEC environment, system failure propagation through the NEC environment, including the human operators faults and their compensatory, recovery, reliability, confidentiality, integrity and availability counteractions needed for maintaining the NEC capability. Concepts for real-time distributed systems of systems consisting of networked entities were derived in order to allow the reconfiguration of the partial processes controlling the operational capacity allocation of UAV formations or the change of working modes in a certain module (determining a change of working mode in a different module). The responsibility of orchestrating the working modes switching is with the Group/Global Reconfiguration Resource Manager organized as a supervisory/coordination control system.

Health Management

We developed a monitoring and diagnostic capability, which when integrated with the prognostics developments, provides a health management (HM) system for use in complex networks. The challenge was to bring health management to any level across a network enabled system. When dealing with large-scale complex network enabled systems, the design of HM systems is difficult because of the large numbers of sensors, actuators and entities which must be taken into account in what are essentially uncertain environments.  It was important to develop methodologies for failure/fault diagnostics, to address both the hierarchical and distributed aspects of network enabled systems and to give careful consideration to the management of information across the network.

The objective was to develop methodologies for health management systems, which will provided appropriate information for prognostics and reconfiguration as well as for consideration by human decision-makers. In this way, network enabled capability (NEC) is achieved.

In developing the tools, experiments were designed to assess performance against the requirements of industry and to understand any ambiguities and inaccuracies.

Prognostics

In achieving Network Enabled Capability a number of entities or systems must work in collaboration in order to achieve a certain mission objective. The mission performed by this system of systems will be comprised of a number of phases, which are determined according to the phases that must be performed by the individual entities or platforms that participate in the mission. As such it is a multi-platform phased mission system. A prognostics methodology that is capable of predicting the probability of success for such missions will help to determine whether or not those missions should begin in the first place and whether or not they should continue after faults or changes to the mission environment are reported during the mission. Since the probability of mission success could affect the way that the present or future missions are implemented, the outputs produced by the prognostics methodology must be quickly and accurately obtained. In this way any decisions that are made based on the prognostic outputs can be made in a well-informed and timely manner.

At the time that the research began no methodology existed for determining the probability of success of multi-platform phased mission systems. The challenge involved for Prognostics is to develop a methodology that can do this and work towards the goal of obtaining fast, accurate results. A further challenge is to be able to apply the prognostics methodology quickly to any multi-platform phased mission system that is to be considered.

Research Questions

Reconfiguration
The system of systems which might be operational (e.g., military command and control) or organisational in NEC environments, must continue to operate effectively and predictably under conditions of constituent system faults or failures, battle-damage, or when system elements are up-graded as a result of technology maturation. The overall system functionality should be guaranteed in order to meet availability, reliability, surveillance capability, affordability, and agility requirements. The research to be developed will be connected to the following top level questions:

Health Management
Based on the NECTISE Baseline Requirements from industries, there are six research questions which have been identified. The major research questions are:

Prognostics
The major research questions being addressed by the Prognostics stream involve identifying and applying methodologies that are suitable for creating a reliability-based prognostics capability for multi-platform phased mission systems. Typical major research questions are:

Programme of Research

Reconfiguration
NEC related attributes, functionalities and operations needed for performing the reconfiguration (e.g., QoS measures, resilience, distributed interactions, supervisory resource allocation, location based control) are the basic ingredients characterising the battlefield operational requirements. Interacting networked UAV formations (QoS, interaction performance, interaction control, topology operations, reconfiguration matrix, operational capacity allocation and sharing, location based interactions, availability, survivability, reliability) will be considered in this respect. An algorithmic assessment for the whole mission planning and the possible reconfiguration tasks/scenarios will be the output of this activity. In a time varying environment, the interaction control mechanisms, distributed system operations, parameters and functions are needed for the implementation of the mission reconfiguration functionality subject to availability constraints. In this context, optimising the use of supervisory and autonomous command and control in conjunction with human decision making in uncertain NEC environment will ensure the real-time reconfiguration capability. The application of the developed framework to the reconfiguration under system upgrades and technology maturation must be an ongoing progressive redesign of reconfiguration functionalities. The important difference in this development is that abrupt, but timed and predictable changes on the UAV mission capability will occur.

Health Management
Based on the research questions and challenges, the following broad research tasks have been identified:

  1. The development of new methods for HM eg a new diagnostic tool using neuro-fuzzy techniques.
  2. The development of a data/information framework for HM.

Health Management has focussed on these two tasks to manage information at different levels across a Shared Access Network (SAN) environment. The HM data/information (HMDI) framework is based on an Network Centric Operations Conceptual Framework, which identifies and defines attributes and metrics for each concept of the health management system. The HMDI framework will be associated with the physical and information domains. Work has also been done to develop a simulation environment for assessing the HMDI framework in a SAN environment. The research has also focussed on designing the diagnosis tool for a single system as well as the system-of-systems. The diagnosis tool is based on a fuzzy logic approach with neural network modelling. This approach has the potential to integrate human knowledge into a knowledge-base for the mid and high levels of the HM system.

Prognostics
Prognostics research has concentrated on the use of Binary Decision Diagrams (BDD) to address the research questions. BDDs are simple structures that can be quickly quantified. However, relying on the speed of quantitative analysis of BDDs is insufficient to produce prognoses of mission success in real-time. Particular focus is also placed on the rapid construction of reliability models, which allows models representing new mission configurations to be quickly constructed. Developing the reliability models to incorporate diagnostics information as missions progress also forms an important part of the CM3 prognostics research.

Results

Reconfiguration
In the NEC uncertain environment, the networked control mechanisms seen as distributed system operations will implement the mission reconfiguration functionality subject to availability constraints. The central viewpoint is to use the supervisory and coordination control in conjunction with human decision making to ensure the real-time reconfiguration capability. Mission reconfiguration in a changing environment will rely on fast change detection in the NEC environment state due to asset failure or health condition degradation and a planning of the connectivity, control and communication activities within the UAV formation. Retention of signal propagation delays, transients and time gaps, are qualitative aspects of the reconfiguration mechanisms to be considered in order to guarantee the system operability after reconfiguration. As for the possible architectural choices the following qualitative aspects are to be addressed:

Health Management
A simulation environment has been developed for assessing C&M developments in network enabled systems. In preliminary experiments, multiple UAVs have been simulated and their sensor signals monitored and detected in a SAN environment. The SAN can provide an integrated environment for modelling a range of hybrid dynamical systems and the HM software can provide a test environment to develop the HM system. The neural-fuzzy based methodology for the diagnostics tool is under development. Because of the complexity of fuzzy systems, we have focussed on developing a rule-based compression method for fuzzy rule-base generation, optimisation and management of complexity. Initial simulation results show that the compression method outperforms significantly many other traditional methods for fuzzy rule base reduction.

Further research will continue to develop the diagnostic capability to provide the system health information for prognostics and reconfiguration as well as directly to human decision-makers.

Prognostics
During the progrnostics research a novel Binary Decision Diagram (BDD) based methodology for modelling multi-platform phased mission systems has been developed. The modular nature of this methodology gives flexibility that is invaluable in quickly constructing BDD models representing the failure logic of multi-platform phased mission systems. Rapid quantification of these BDD models means that real-time analysis is a distinct possibility. The development of the prognostics methodology has also involved the incorporation of diagnostics information (faults, changes to the mission environment) in the calculation of mission failure probabilities once a mission is underway. Coupling this with the rapid analysis means that accurate mission failure probability information can be quickly provided in cases where this information is required in order to determine the future course of missions.

 

© Copyright Loughborough University