An incremental-PID-controlled particle swarm optimization algorithm for EEG-data-based estimation of operator functional state

• Published in: Biomedical signal processing and control Vol. 14; pp. 272 - 284
• Main Authors: Zhang, Jianhua, Yang, Shaozeng
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2014
• Subjects: Benchmark mathematical function; EEG signals; Incremental PID control; Operator functional state; Particle swarm optimization; Benchmark mathematical function; EEG signals; Operator functional state; Particle swarm optimization; Incremental PID control
• ISSN: 1746-8094
• DOI: 10.1016/j.bspc.2014.08.005

•Accurate estimation of operator functional state (OFS) is crucial to improve the safety of human–machine system.•A simulated process control task was used to elicit different complexity levels of tasks which can induce different OFS.•OFS was assessed by using EEG signals and fuzzy modeling paradigm.•A newly developed IPID-PSO algorithm was proposed and verified.•The IPID-PSO algorithm was used to identify the fuzzy model of OFS. Operator functional state (OFS) is referred to as the ability of an operator to complete assigned tasks which may fluctuate over time. In the adaptive human–machine systems, it is required that the OFSs should be estimated in real-time in order to prevent the potential performance breakdown. To this end, an accurate OFS estimation model must be established. OFS can be reflected by operator's various physiological signals including EEG measures. In this paper, five subjects’ EEG signals were collected while working jointly with the AutoCAMS, a simulated software environment of human–machine cooperative control system. The fuzzy models are employed to estimate the OFS-related operator performance from three EEG-based input features. To derive the optimal fuzzy models, a new incremental-PID-controlled particle swarm optimization (IPID-PSO) algorithm is developed. The IPID-PSO algorithm is a combination of the standard PSO algorithm and the incremental PID control algorithm. The usefulness of the IPID-PSO algorithm is firstly validated by its application to eight benchmark function optimization problems. The superiority of IPID-PSO to the standard PSO algorithm is shown to be more significant especially for the optimization of multimodal functions with multiple local optima. Then, it is used to optimize the OFS model. The model parameters optimization accuracy and convergence rate of the IPID-PSO and standard PSO algorithm were compared when used for model-based OFS estimation. The IPID-PSO algorithm developed in this work has potential to be widely applied to other real-world optimization and model identification problems.