Driving Intention Identification Based on Neural Network Optimized by Particle Swarm Optimization

Authors

  • Min Li Qingdao University of Technology, School of Mechanical and Automotive Engineering 266520
  • Xilong Zhang Qingdao University of Technology, School of Mechanical and Automotive Engineering 266520
  • Yongliang Zhang Qingdao University of Technology, School of Mechanical and Automotive Engineering 266520
  • Dayi Qu Qingdao University of Technology, School of Mechanical and Automotive Engineering 266520
  • Fuquan Pan Qingdao University of Technology, School of Mechanical and Automotive Engineering 266520

Abstract

Electroencephalograph (EEG) signals received fromdifferent areas of the human brainwere analyzed using a combination of theoretical analysis, experimentation and simulation. The driving simulation experiment was designed, and an acquisition system was set up to collect EEG signals when drivers in the experiment turned left, turned right, or proceeded straight within a specified time window. The collected EEG signals were processed, throughwavelet package transformand other signal processing methods, to extract their feature parameters. The models, based on a SupportVectorMachine (SVM) model optimized by Particle Swarm Optimization (PSO) and on a Neural Network (NN), were built to recognize motorists’ driving intentions through the processed EEG signals. The recognized driving intention with better recognition rate was transformed into corresponding instruction signal which can control the vehicle to achieve automatic drive. The analysis and result shows that the recognition rate of the model based on SVM optimized by PSO increases to 73.5%, and that of the model based on NN achieves a better rate of 92.9%.

Keywords: Brain–Computer Interface (BCI); Driving Intention Recognition; Particle Swarm Optimization (PSO); Support Vector Machine (SVM); Neural Network (NN); Automotive Driving

Cite As

M. Li, X. Zhang, Y. Zhang, D. Qu, F. Pan, "Driving Intention Identification Based on Neural Network Optimized by
Particle Swarm Optimization", Engineering Intelligent Systems, vol. 26 no. 4, pp. 169-173, 2018.





Published

2018-12-01