Dynamic Situational Awareness Discovery for Goal-Driven Operation of Resource-Constrained Connected Systems

#Dynamic #Situational-Awareness-Discovery #Goal-Driven #Resource-Constrained #Connected-Systems

The extensive development of wireless communication technologies and their integration with artificial intelligence have empowered diverse Industrial Internet of Things (IIoT) applications, enabling dynamic devices to work collaboratively toward goals that go beyond conventional focus on data rates. However, with the growing network scale, the dynamic network conditions and constrained resources, it becomes extremely challenging to optimize network decision-making. Driven by application goals in highly dynamic and complex environments, future connected systems are expected to rapidly and efficiently discover situational awareness, including location, time, topology structure, and resources to support adaptive decision-making. Therefore, this thesis designs top-down discovery mechanisms of situational awareness for future dynamic connected systems, i.e., rapid discovery in spatio-temporal domain, energy-efficient network topology virtualization and adaptation, and dependency-aware resource allocation under dynamic conditions.  

Firstly, to achieve rapid situational discovery in spatio-temporal domain for IIoT, a distributed, optimal measurement geometry directed integrated localization and synchronization (ILAS) framework for large-scale dynamic connected systems is proposed. Specifically, this framework comprises a collaborating nodes selection algorithm, where the collaborating nodes constructing optimal measurement geometry are selected for the best estimation accuracy with controlled complexity, and a sequential-state-stacking belief propagation (3SBP) implementation, where the computational complexity is further reduced by limiting matrix inversions and squared roots to the dimensions of only a subset of the overall stacked states. Simulation results demonstrate a significant enhancement in the robustness of the ILAS estimation and reduction in the computational complexity compared to the baseline schemes.

Secondly, to simplify the discovery of situational awareness under dynamic conditions, a virtual network topology (VNT) construction and adaptation framework is proposed. The proposed VNT construction transforms complex, time-varying physical parameters into stable performance indicators, thus reducing network parameter dimensions while preserving their key influences on data collection. Based on the simplified network monitoring tool, we develop a VNT adaptation algorithm for rapid optimal network operations to maximize the effectiveness of data collection, with the value defined by the sum of collected data weighted by spatiotemporal priority. Numerical simulations and benchmark comparisons demonstrate a significant performance improvement in maximizing the effectiveness of data collection using the proposed framework.

Thirdly, based on the situational awareness, to effectively manage the network topology and resource allocation, we propose a dependency-aware task offloading strategy for integrated air-ground networks. Specifically, demanded by various applications, full tasks define the dependencies among dynamic nodes, which is modeled by dynamic directed acyclic graphs. We define the value of task offloading as the average success rate of completing full tasks, and formulate topology management, bandwidth, and computing resources allocation as a distributed optimization problem to maximize this value. A multi-agent deep deterministic policy gradient (MADDPG)-based algorithm is developed to solve it. Numerical results demonstrate that the proposed strategy significantly improves averaged full task success rate compared to baseline approaches in complex and dynamic environments.