Analyzing Operator States and the Impact of AI-Enhanced Decision Support in Control Rooms: A Human-in-the-Loop Specialized Reinforcement Learning Framework for Intervention Strategies (2024-02-20T00:00:00.000000Z)

TL;DR

These findings are particularly relevant when predicting the overall performance of the individual participant and their capacity to successfully handle a plant upset and the alarms connected to it using process and human-machine interaction logs in real-time which resulted in a 95.8% prediction accuracy using hidden Markov model.

Abstract

Abstract In complex industrial and chemical process control rooms, effective decision-making is crucial for safety and efficiency. The experiments in this paper evaluate the impact and applications of an AI-based decision support system integrated into an improved human-machine interface, using dynamic influence diagrams, a hidden Markov model, and deep reinforcement learning. The enhanced support system aims to reduce operator workload, improve situational awareness, and provide different intervention strategies to the operator adapted to the current state of both the system and human performance. Such a system can be particularly useful in cases of information overload when many alarms and inputs are presented all within the same time window, or for junior operators during training. A comprehensive cross-data analysis was conducted, involving 47 participants and a diverse range of data sources such as smartwatch metrics, eye-tracking data, process logs, and responses from questionnaires. The results indicate interesting insights regarding the effectiveness of the approach in aiding decision-making, decreasing perceived workload, and increasing situational awareness for the scenarios considered. Additionally, the results provide insights to compare differences between styles of information gathering when using the system by individual participants. These findings are particularly relevant when predicting the overall performance of the individual participant and their capacity to successfully handle a plant upset and the alarms connected to it using process and human-machine interaction logs in real-time which resulted in a 95.8% prediction accuracy using hidden Markov model. These predictions enable the development of more effective intervention strategies.

References50 items

Exploring the Influence of Human System Interfaces: Introducing Support Tools and an Experimental Study

Enhancing Control Room Operator Decision Making

Experiment data: Human-in-the-loop decision support in process control rooms

Specialized Deep Residual Policy Safe Reinforcement Learning-Based Controller for Complex and Continuous State-Action Spaces

Hierarchical framework for interpretable and specialized deep reinforcement learning-based predictive maintenance

ACT-R based human digital twin to enhance operators’ performance in process industries

Post-saccadic Eye Movement Indices Under Cognitive Load: A Path Analysis to Determine Visual Performance

Application of Artificial Intelligence in Detection and Mitigation of Human Factor Errors in Nuclear Power Plants: A Review

Concept of an Intelligent Operator Support System for Initial Emergency Responses in Nuclear Power Plants

Eye Movement and Pupil Measures: A Review

Hierarchical Reinforcement Learning

Assessing Autonomic Function from Electrodermal Activity and Heart Rate Variability During Cold-Pressor Test and Emotional Challenge

Process Systems Engineering and the Human-in-The-Loop: The Smart Control Room

Optuna: A Next-generation Hyperparameter Optimization Framework

Deep Reinforcement Learning for Process Control: A Primer for Beginners

An Introduction to Deep Reinforcement Learning

A Review of Alarm System Design for Advanced Control Rooms of Nuclear Power Plants

Addressing Function Approximation Error in Actor-Critic Methods

Safe and sensible preprocessing and baseline correction of pupil-size data

Risk-Based Decision Making for the Management of Change in Process Plants: Benefits of Integrating Probabilistic and Phenomenological Analysis

The Differences and Similarities Between Two-Sample T-Test and Paired T-Test

The Option-Critic Architecture

T test as a parametric statistic

Understanding P values

Hierarchical reinforcement learning and decision making

A decision support system for identifying abnormal operating procedures in a nuclear power plant

Pearson’s correlation coefficient

Bayesian Networks and Influence Diagrams: A Guide to Construction and Analysis

Development of an Integrated Decision Support System to Aid the Cognitive Activities of Operators in Main Control Rooms of Nuclear Power Plants

Influence Diagrams

Time course of information processing during scene perception: The relationship between saccade amplitude and fixation duration

Learning Options in Reinforcement Learning

The effect of cognitive load on saccadic eye movements.

Dynamic programming and influence diagrams

A tutorial on hidden Markov models and selected applications in speech recognition

“Cisc-live-lab-3/dataset: v1.0.2,”

ammar-n-abbas/drl-based-decision-support: v1.0.0 (v1.0.0)

Dynamic Influence Diagram-Based Deep Reinforcement Learning Framework and Application for Decision Support for Operators in Control Rooms

Decision making for process control management in control rooms: A survey methodology and initial findings

hmmlearn: Hidden Markov models in python with Scikit-learn like API

Deep Residual Policy Reinforcement Learning as a Corrective Term in Process Control for Alarm Reduction; A Preliminary Report

Interpretable input-output hidden Markov model-based deep reinforcement learning for the predictive maintenance of turbofan engines

What Is the Bonferroni Correction ?

Chasparis earned his mechanical engineering degree from Athens, and M.Sc. and Ph.D. from UCLA

A set of conditional probability distributions P, where each distribution P ð X v j X pa ð v Þ Þ is associated with a discrete random variable X v given its parent variables X pa ð v Þ in the graph

“Impact of human system interfaces on process control room operators: An introduction of support tools and a design of experiment.”

iv. A set of utility functions U, with each utility function u ð X pa ð v Þ Þ associated with a node v in the subset of utility nodes V U � V

i) A Directed Acyclic Graph (DAG) G = ( V, E ) , where V represents the set of nodes and E is the set of directed edges, indicating dependency relations and information flow

Tobii Technology. (2024a)

“Tobii pro lab.”