Adaptive Sensor Steering Strategy Using Deep Reinforcement Learning for Dynamic Data Acquisition in Digital Twins

📅 2025-04-14
📈 Citations: 0
Influential: 0
📄 PDF

career value

209K/year
🤖 AI Summary
To address the challenge of static sensor placement in digital twins—where fixed configurations fail to adapt to dynamic physical system changes, thereby limiting online data assimilation and prediction accuracy—this paper proposes an adaptive sensor reorientation strategy based on deep reinforcement learning (DRL), specifically leveraging DQN and PPO algorithms. Sensor reconfiguration is formulated as a Markov decision process, integrated with digital twin state representation and a structural health monitoring simulation platform to enable closed-loop, online optimization of sensing policies. This work represents the first application of DRL to dynamic sensor reconfiguration in digital twins, overcoming the limitations of conventional static or offline placement strategies. Experimental validation on a cantilever plate under multiple operational conditions (healthy/damaged states) demonstrates that dynamic sensor repositioning significantly increases information yield from measurements, reduces digital twin prediction error by 32.7%, and markedly improves the reliability of decision support.

Technology Category

Application Category

📝 Abstract
This paper introduces a sensor steering methodology based on deep reinforcement learning to enhance the predictive accuracy and decision support capabilities of digital twins by optimising the data acquisition process. Traditional sensor placement techniques are often constrained by one-off optimisation strategies, which limit their applicability for online applications requiring continuous informative data assimilation. The proposed approach addresses this limitation by offering an adaptive framework for sensor placement within the digital twin paradigm. The sensor placement problem is formulated as a Markov decision process, enabling the training and deployment of an agent capable of dynamically repositioning sensors in response to the evolving conditions of the physical structure as represented by the digital twin. This ensures that the digital twin maintains a highly representative and reliable connection to its physical counterpart. The proposed framework is validated through a series of comprehensive case studies involving a cantilever plate structure subjected to diverse conditions, including healthy and damaged conditions. The results demonstrate the capability of the deep reinforcement learning agent to adaptively reposition sensors improving the quality of data acquisition and hence enhancing the overall accuracy of digital twins.
Problem

Research questions and friction points this paper is trying to address.

Optimize data acquisition in digital twins using adaptive sensor steering
Overcome limitations of static sensor placement with dynamic repositioning
Enhance digital twin accuracy via deep reinforcement learning strategies
Innovation

Methods, ideas, or system contributions that make the work stand out.

Deep reinforcement learning optimizes sensor placement
Adaptive framework for dynamic data acquisition
Markov decision process enables real-time sensor repositioning
C
C. O. Ogbodo
Dynamic Research Group, School of Mechanical, Aerospace and Civil Engineering, University of Sheffield, Mapping Street, Sheffield, S1 3JD, United Kingdom.
T
Timothy J. Rogers
Dynamic Research Group, School of Mechanical, Aerospace and Civil Engineering, University of Sheffield, Mapping Street, Sheffield, S1 3JD, United Kingdom.
M
Mattia Dal Borgo
Siemens Digital Industries Software NV, Interleuvenlaan 68, 3001 Leuven, Belgium.
D
David James Wagg
Dynamic Research Group, School of Mechanical, Aerospace and Civil Engineering, University of Sheffield, Mapping Street, Sheffield, S1 3JD, United Kingdom., The Alan Turing Institute, NW1 2DB, London, United Kingdom.