基于隐马尔科夫模型的操作员功能状态分类
|
摘要
复杂人机系统中,执行高危控制任务操作员的功能状态(Operator Functional State, OFS)骤降或失效往往会引发十分严重的事故。如何对操作员功能状态进行准确地预测以避免该类事故的发生己逐渐成为时下研究者致力解决的难题之一。解决该问题的一个有效途径是寻找恰当的分类器,使用操作员的电生理信号,对操作员功能状态进行有效的分类。
在基于太空舱空气组分自动控制系统仿真软件(automation-enhanced Cabin Air Management System, aCAMS)的操作员功能状态实验数据的基础上,本文引入了一种基于相关性谱分析的被试个体最优电生理特征选取方法,结合其他电生理信号,作为经典的隐马尔科夫模型的输入进行分类建模。仿真结果表明隐马尔科夫模型以其优秀的时间序列信号建模能力,在OFS分类问题中有着良好的效果。最后本文对OFS,分类问题中不同参数下隐马尔科夫模型分类性能进行了对比,挖掘在OFS问题中模型参数选取的潜在规律。
High risk operating task in complex human machine system is vulnerable to the operator's break-down of functional state. To effectively estimate the OFS to avoid such serious issue is now one of the most challenging topics for researchers. One potential solution is to well classify the OFS.
The experiment is based on aCAMS simulation software (automation-enhanced Cabin Air Management System). In this paper, a method of choosing the best feature for the individual subject based on correlation spectrum analysis is raised. With the new features extracted, classical Hidden Markov Model is introduced to OFS classification problem. The result shows that HMM is decent in OFS classification applications with its strong capability of modeling time serial signals. At last, the parameter choosing principals for HMM in OFS classification is researched.
在基于太空舱空气组分自动控制系统仿真软件(automation-enhanced Cabin Air Management System, aCAMS)的操作员功能状态实验数据的基础上,本文引入了一种基于相关性谱分析的被试个体最优电生理特征选取方法,结合其他电生理信号,作为经典的隐马尔科夫模型的输入进行分类建模。仿真结果表明隐马尔科夫模型以其优秀的时间序列信号建模能力,在OFS分类问题中有着良好的效果。最后本文对OFS,分类问题中不同参数下隐马尔科夫模型分类性能进行了对比,挖掘在OFS问题中模型参数选取的潜在规律。
High risk operating task in complex human machine system is vulnerable to the operator's break-down of functional state. To effectively estimate the OFS to avoid such serious issue is now one of the most challenging topics for researchers. One potential solution is to well classify the OFS.
The experiment is based on aCAMS simulation software (automation-enhanced Cabin Air Management System). In this paper, a method of choosing the best feature for the individual subject based on correlation spectrum analysis is raised. With the new features extracted, classical Hidden Markov Model is introduced to OFS classification problem. The result shows that HMM is decent in OFS classification applications with its strong capability of modeling time serial signals. At last, the parameter choosing principals for HMM in OFS classification is researched.
引文
[1]G. R. J. Hockey. Operator functional state:The Assessment and Prediction of Human Performance Degradation in Complex Tasks[M]. US:IOS Press,2003.
[2]Ichiro Kageyama, Yukiyu Kuriyagawa. A modeling of heart rate variability to estimate mental work load[A]. IEEE SMC 99 Conference Proceedings:Systems, Man, and Cybernetics [C]. Tokyo, Japan:IEEE International,1999.294-299.
[3]O. Kubo, H. Yoshikawa, M. Tkahashi. Mutual Adaptive Interface:Basic Concepts[A]. Robot and Human Communication Proceedings [C]. Tokyo, Japan:2nd IEEE International Workshop,1993.454-459.
[4]Frederick G. Freeman, Mark W. Scerbo, Peter J. Mikulka. The Efficacy of Psychophysiological Measures for Implementing Adaptive Technology[M]. US: Hampton,2001.20-42.
[5]M. Takahashi, etc. Development of Realtime Cognitive State Estimator[J]. Control Engineering Practice,1995,3(2):275-280.
[6]苏令波.电气操作运行人员的心理负荷研究[D].北京：北京交通大学,2006,12.
[7]G. Conway. Effects of workload, effort and fatigue on complex performance: application and tests of compensatory[D]. Leeds, UK:PhD Thesis, University of Leeds, 2005.
[8]G. F. Wilson, F. Fisher. Cognitive task classification based upon topographic EEG data[J]. Biological Psychology,1995,40(1-2):239-250.
[9]A. Gevins, M. ESmith, H. Leong. Monitoring working memory load during computer-based tasks with EEG pattern recognition methods[J]. Human Factors,1998,40(1): 79-91.
[10]C. A. Russell, G. F. Wilson. Air traffic controller functional state classification using neural networks[J]. Proc of the Artificial Neural Networks in Engineering Conference. 1998, (8):649-654.
[11]G. F. Wilson, F. T. Eggemerier. Physiological measures of workload in multi-task environments [A]. Damos D. Multiple-task Performance [C]. Boca Raton, US:CRC Press,1991.329-360.
[12]A. Gevins, M. E. Smith. Detecting transient cognitive impairment with EEG pattern recognition methods [J]. Aviation, Space and Environmental Medicine,1999,70(1): 1018-1024.
[13]C. Dussault, J. C. Jouanin, C. Y. Guezennec. EEG and ECG changes during selected flight sequences[J]. Aviat Space Environ Med,2004,75(10):889-97
[14]C. Dussault, J. C. Jouanin, M. Philippe et al. EEG and ECG changes during simulator operation reflect mental workload and vigilance[J]. Aviat Space Environ Med,2005, 76(4):344-51
[15]A. R. Nikolaev, G. A. Ivanitskii, A. M. Ivanitskii.Reproducible EEG alpha-patterns in psychological task solving[J]. Human Physiology,1998,24(2):261-268.
[16]J. R. Comstock. The Multi-Attribute Task Battery for Human Operator Workload and Strategic Behavior Research[M]. US:NASA Langley Research Center National Technical Information Service,1992:127-139.
[17]A. T. Pope, E. H. Bogart, D. S. Bartolome. Biocybernetic system evaluates indices of operator engagerment in automated task[J]. Biological Psychology,1995,40(1): 187-195.
[18]F. G. Freedman, P. J. Mikulka, L. J. Prinzel etc. Evaluation of an adaptive automation system using three EEG indices with a visual tracking task[J]. Biological Psychology, 1999,50(1):61-76.
[19]G. F. Wilson. Operator functional state classification using neural networks with combined psychological and performance measures[A]. Proc of the 43rd HFES Annual Meeting[C].1999.1099-1102.
[20]G.R.J. Hockey, D.G. Wastell, J. Sauer. Effects of sleep deprivation and user inerface on complex performance:a multilevel analysis of compensatory control[J]. Human Factors, 1998,40(2):233-253.
[21]Chong Zhang, Chongxun Zheng, Xiaolin Yu, etc. VDT Mental Fatigue Using Multichannel Linear Descriptors and KPCA-HMM[A]. EURASIP Journal on Advances in Signal Processing[C]. NY, US:Hindawi Publishing Corp,2008.1155-1166.
[22]王政,李延军,严曲,王宏伟等,监控作业脑力疲劳综合评定方法研究[J],航天医学与医学工程,2010,23(1)：20-24.
[23]Donald L. Schomer, Fernando Lopes da Silva. Electroencephalography:Basic Principles, Clinical Applications, and Related Fields[M]. US:Lippincott Williams & Wilkins,2010:234-251.
[24]Wolfgang Klimesch, EEG alpha and theta oscillations reflect cognitive and memory performance:a review and analysis[J], Brain Research Reviews,1999,29(3):169-195.
[25]Mehmet Akin, Muhammed B. Kurt. Estimating vigilance level by using EEG and EMG signals[J], Neural Comput & Applic,2008,17(3):227-336.
[26]H. H. Jasper. Report of the committee on methods of clinical examination in electroencephalography[J]. Electroencephalography and Clinical Neurophysiology, 1958,10(2):370-375.
[27]E. Braunwald. Heart Disease:A Textbook of Cardiovascular Medicine.5th Edition[M]. Philadelphia:W.B. Saunders Co.,1997.108-109.
[28]P. Nickel, G.R.J. Hockey, A.C. Roberts, etc. Markers of high risk operator functional state in adaptive control of process automation[A], Proc. Of IEA[C].2006.37-43.
[29]Jian-Hua Zhang, Xing-Yu Wang, M. Mahfouf, etc. Fuzzy Logic based Identification of Operator Functional States Using Multiple Physiological and Performance Measures[A], International Conference on BioMedical Engineering and Informatics[C]. Sanya, China: 2008.77-79.
[30]Chin-Teng Lin, Ruei-Cheng Wu, Tzyy-Ping Jung, etc. Estimating Driving Performance Based on EEG Spectrum Analysis[J], EURASIP Journal on Applied Signal Processing, 2005,19(1):3165-3174
[31]闫春红.心率变异性非线性动力学分析的研究[D].重庆：重庆大学硕士学位论文,2004
[32]P. G. N. A. Jorna. Heart rate and workload variations in actual and simulated flight[J]. Ergonomics,1993,36(9):1043-1054.
[33]Gevins A, Smith ME. Detecting transient cognitive impairment with EEG pattern recognition methods[J]. Aviat Space Environ Med,1999,70(1):1018-1024.
[34]彭小迪,张建华,王行愚等.基于遗传算法的操作员功能状态模糊建模[J].电子测量与仪器学报,2009,1(1)：27-30.
[35]Jian-Hua Zhang, Xing-Yu Wang, M. Mahfouf, etc. Use of Heart Rate Variability Analysis for Quantitatively Assessing Operator's Mental Workload[A], BMEI[C].2008. 668-672.
[36]Lawrence R. Rabiner, A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition[J], Proceedings of the IEEE,1989,77(2):257-286.
[37]谢锦辉.隐Markov模型(HMM)及其在语音处理中的应用[M].武汉：华中理工大学出版社,1995,1-46.
[38]F Lotte, M Congedo. A Review of Classification Algorithms for EEG-based Brain-Computer Interfaces[J]. Journal of Neural Engineering,4(2007):R1-R13.
[39]A. W. Drake. Discrete-state Markov processes[M]. NY, US:McGraw-Hill,1967, Chapter 5.
[40]L.E. Baum and T.Petrie. Statistical interface for probabilistic functions of finite state Markov chains[J]. Ann. Math. Stat.,1966,37(1966):1554-1563.
[41]L.E. Baum and J.A. Egon. An inequality with applications to statistical estimation for probabilistic functions of a Markov process and to a model for ecology[J]. Bull. Amer. Metorol. Soc.,1967,73(1967):1554-1563.
[42]A.J. Viterbi. Error bounds for convolutional codes and an asymptotically optimal decoding algorithm[J]. IEEE Trans. Informat. Theory,1967, IT-13:260-269.
[43]G.D. Forney. The Viterbi algorithm[J]. Proc. IEEE,1973,61(1973):268-278.
[44]L.E. Baum and G.R.Sell. Growth functions for transformations on manifolds. Pac. J. Math.,1968,27(2):211-227.
[45]L.E. Baum, T. Petrie, G. Soules, etc. A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains [J]. Ann. Math. Stat., 1970,41(1):164-171.
[46]L.E. Baum. An inequality and associated maximization technique in statistical estimation for probabilistic functions of Markov processes[J], Inequalities,1972, 3(1972):1-8.
[47]F. Jelinek. Continuous speech recognition by statistical methods[J]. Proc. IEEE,1976, 21(1976):250-256.
[48]R. Bakis. Continuous speech word recognition via centisecond acoustic states[A]. Proc. ASA Meeting[C]. Apr.1976
[49]Guilherme A. Barreto, Rewbenio A. Frota. On the classification of mental tasks:a performance comparison of neural and statistical approches[J]. IEEE workshop on Machine Learning for Signal Process,2004.
[50]Leonard J. Trejo, Kevin Knuth, Raquel Prado, etc. EEG-Based Estimation of Mental Fatigue:Convergent Evident for a Three-State Model [J], Augmented Cognition, HCII, 2007, LNAI 4565:201-211.
[51]Hyekyung Lee, Seungjin Choi. PCA+HMM+SVM for EEG pattern classification[J]. IEEE. Signal Processing and Its Applications,2003,1(2003):541-544.
[52]Tom Dietterich. Machine Learning for Sequential Data:A Review[A]. Lecture Notes in Computer Science[C],2002.
[53]Samy Bengio. HMM and IOHMM modeling of EEG rhythms for asynchronous BCI systems[J]. European Symposium on Artificial Neural Networks,2004:199-204.
[54]B. Obermaier, C. Guger, C. Neuper. Hidden Markov models for online classification of single trial EEG data[J]. Pattern Recognition Letters,2001,22(1):1299-1399.
[55]T. M. Vaughan, W. J. Heetderks, L. J. Trejo. Brain-computer interface technology:A review of the second international meeting [J]. IEEE transaction on Rehabilitation Engineering,2003,11:94-109.
[56]赵春临,郑崇勋,赵敏.基于脑电功率谱-连续隐马尔科夫链的精神疲劳分级模型[J],西安交通大学学报,2007,41：1474-1478.
[57]Ali Ozgur Argunsah. AR-PCA-HMM Approach for Sensorimotor Task Classification in EEG-based Brain-Computer Interfaces[A]. International Conference Pattern Recognition[C],2010.113-116.
[58]顾燕红.自适应剖面隐马氏模型软件言之[D].上海：上海大学,2005.
[2]Ichiro Kageyama, Yukiyu Kuriyagawa. A modeling of heart rate variability to estimate mental work load[A]. IEEE SMC 99 Conference Proceedings:Systems, Man, and Cybernetics [C]. Tokyo, Japan:IEEE International,1999.294-299.
[3]O. Kubo, H. Yoshikawa, M. Tkahashi. Mutual Adaptive Interface:Basic Concepts[A]. Robot and Human Communication Proceedings [C]. Tokyo, Japan:2nd IEEE International Workshop,1993.454-459.
[4]Frederick G. Freeman, Mark W. Scerbo, Peter J. Mikulka. The Efficacy of Psychophysiological Measures for Implementing Adaptive Technology[M]. US: Hampton,2001.20-42.
[5]M. Takahashi, etc. Development of Realtime Cognitive State Estimator[J]. Control Engineering Practice,1995,3(2):275-280.
[6]苏令波.电气操作运行人员的心理负荷研究[D].北京：北京交通大学,2006,12.
[7]G. Conway. Effects of workload, effort and fatigue on complex performance: application and tests of compensatory[D]. Leeds, UK:PhD Thesis, University of Leeds, 2005.
[8]G. F. Wilson, F. Fisher. Cognitive task classification based upon topographic EEG data[J]. Biological Psychology,1995,40(1-2):239-250.
[9]A. Gevins, M. ESmith, H. Leong. Monitoring working memory load during computer-based tasks with EEG pattern recognition methods[J]. Human Factors,1998,40(1): 79-91.
[10]C. A. Russell, G. F. Wilson. Air traffic controller functional state classification using neural networks[J]. Proc of the Artificial Neural Networks in Engineering Conference. 1998, (8):649-654.
[11]G. F. Wilson, F. T. Eggemerier. Physiological measures of workload in multi-task environments [A]. Damos D. Multiple-task Performance [C]. Boca Raton, US:CRC Press,1991.329-360.
[12]A. Gevins, M. E. Smith. Detecting transient cognitive impairment with EEG pattern recognition methods [J]. Aviation, Space and Environmental Medicine,1999,70(1): 1018-1024.
[13]C. Dussault, J. C. Jouanin, C. Y. Guezennec. EEG and ECG changes during selected flight sequences[J]. Aviat Space Environ Med,2004,75(10):889-97
[14]C. Dussault, J. C. Jouanin, M. Philippe et al. EEG and ECG changes during simulator operation reflect mental workload and vigilance[J]. Aviat Space Environ Med,2005, 76(4):344-51
[15]A. R. Nikolaev, G. A. Ivanitskii, A. M. Ivanitskii.Reproducible EEG alpha-patterns in psychological task solving[J]. Human Physiology,1998,24(2):261-268.
[16]J. R. Comstock. The Multi-Attribute Task Battery for Human Operator Workload and Strategic Behavior Research[M]. US:NASA Langley Research Center National Technical Information Service,1992:127-139.
[17]A. T. Pope, E. H. Bogart, D. S. Bartolome. Biocybernetic system evaluates indices of operator engagerment in automated task[J]. Biological Psychology,1995,40(1): 187-195.
[18]F. G. Freedman, P. J. Mikulka, L. J. Prinzel etc. Evaluation of an adaptive automation system using three EEG indices with a visual tracking task[J]. Biological Psychology, 1999,50(1):61-76.
[19]G. F. Wilson. Operator functional state classification using neural networks with combined psychological and performance measures[A]. Proc of the 43rd HFES Annual Meeting[C].1999.1099-1102.
[20]G.R.J. Hockey, D.G. Wastell, J. Sauer. Effects of sleep deprivation and user inerface on complex performance:a multilevel analysis of compensatory control[J]. Human Factors, 1998,40(2):233-253.
[21]Chong Zhang, Chongxun Zheng, Xiaolin Yu, etc. VDT Mental Fatigue Using Multichannel Linear Descriptors and KPCA-HMM[A]. EURASIP Journal on Advances in Signal Processing[C]. NY, US:Hindawi Publishing Corp,2008.1155-1166.
[22]王政,李延军,严曲,王宏伟等,监控作业脑力疲劳综合评定方法研究[J],航天医学与医学工程,2010,23(1)：20-24.
[23]Donald L. Schomer, Fernando Lopes da Silva. Electroencephalography:Basic Principles, Clinical Applications, and Related Fields[M]. US:Lippincott Williams & Wilkins,2010:234-251.
[24]Wolfgang Klimesch, EEG alpha and theta oscillations reflect cognitive and memory performance:a review and analysis[J], Brain Research Reviews,1999,29(3):169-195.
[25]Mehmet Akin, Muhammed B. Kurt. Estimating vigilance level by using EEG and EMG signals[J], Neural Comput & Applic,2008,17(3):227-336.
[26]H. H. Jasper. Report of the committee on methods of clinical examination in electroencephalography[J]. Electroencephalography and Clinical Neurophysiology, 1958,10(2):370-375.
[27]E. Braunwald. Heart Disease:A Textbook of Cardiovascular Medicine.5th Edition[M]. Philadelphia:W.B. Saunders Co.,1997.108-109.
[28]P. Nickel, G.R.J. Hockey, A.C. Roberts, etc. Markers of high risk operator functional state in adaptive control of process automation[A], Proc. Of IEA[C].2006.37-43.
[29]Jian-Hua Zhang, Xing-Yu Wang, M. Mahfouf, etc. Fuzzy Logic based Identification of Operator Functional States Using Multiple Physiological and Performance Measures[A], International Conference on BioMedical Engineering and Informatics[C]. Sanya, China: 2008.77-79.
[30]Chin-Teng Lin, Ruei-Cheng Wu, Tzyy-Ping Jung, etc. Estimating Driving Performance Based on EEG Spectrum Analysis[J], EURASIP Journal on Applied Signal Processing, 2005,19(1):3165-3174
[31]闫春红.心率变异性非线性动力学分析的研究[D].重庆：重庆大学硕士学位论文,2004
[32]P. G. N. A. Jorna. Heart rate and workload variations in actual and simulated flight[J]. Ergonomics,1993,36(9):1043-1054.
[33]Gevins A, Smith ME. Detecting transient cognitive impairment with EEG pattern recognition methods[J]. Aviat Space Environ Med,1999,70(1):1018-1024.
[34]彭小迪,张建华,王行愚等.基于遗传算法的操作员功能状态模糊建模[J].电子测量与仪器学报,2009,1(1)：27-30.
[35]Jian-Hua Zhang, Xing-Yu Wang, M. Mahfouf, etc. Use of Heart Rate Variability Analysis for Quantitatively Assessing Operator's Mental Workload[A], BMEI[C].2008. 668-672.
[36]Lawrence R. Rabiner, A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition[J], Proceedings of the IEEE,1989,77(2):257-286.
[37]谢锦辉.隐Markov模型(HMM)及其在语音处理中的应用[M].武汉：华中理工大学出版社,1995,1-46.
[38]F Lotte, M Congedo. A Review of Classification Algorithms for EEG-based Brain-Computer Interfaces[J]. Journal of Neural Engineering,4(2007):R1-R13.
[39]A. W. Drake. Discrete-state Markov processes[M]. NY, US:McGraw-Hill,1967, Chapter 5.
[40]L.E. Baum and T.Petrie. Statistical interface for probabilistic functions of finite state Markov chains[J]. Ann. Math. Stat.,1966,37(1966):1554-1563.
[41]L.E. Baum and J.A. Egon. An inequality with applications to statistical estimation for probabilistic functions of a Markov process and to a model for ecology[J]. Bull. Amer. Metorol. Soc.,1967,73(1967):1554-1563.
[42]A.J. Viterbi. Error bounds for convolutional codes and an asymptotically optimal decoding algorithm[J]. IEEE Trans. Informat. Theory,1967, IT-13:260-269.
[43]G.D. Forney. The Viterbi algorithm[J]. Proc. IEEE,1973,61(1973):268-278.
[44]L.E. Baum and G.R.Sell. Growth functions for transformations on manifolds. Pac. J. Math.,1968,27(2):211-227.
[45]L.E. Baum, T. Petrie, G. Soules, etc. A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains [J]. Ann. Math. Stat., 1970,41(1):164-171.
[46]L.E. Baum. An inequality and associated maximization technique in statistical estimation for probabilistic functions of Markov processes[J], Inequalities,1972, 3(1972):1-8.
[47]F. Jelinek. Continuous speech recognition by statistical methods[J]. Proc. IEEE,1976, 21(1976):250-256.
[48]R. Bakis. Continuous speech word recognition via centisecond acoustic states[A]. Proc. ASA Meeting[C]. Apr.1976
[49]Guilherme A. Barreto, Rewbenio A. Frota. On the classification of mental tasks:a performance comparison of neural and statistical approches[J]. IEEE workshop on Machine Learning for Signal Process,2004.
[50]Leonard J. Trejo, Kevin Knuth, Raquel Prado, etc. EEG-Based Estimation of Mental Fatigue:Convergent Evident for a Three-State Model [J], Augmented Cognition, HCII, 2007, LNAI 4565:201-211.
[51]Hyekyung Lee, Seungjin Choi. PCA+HMM+SVM for EEG pattern classification[J]. IEEE. Signal Processing and Its Applications,2003,1(2003):541-544.
[52]Tom Dietterich. Machine Learning for Sequential Data:A Review[A]. Lecture Notes in Computer Science[C],2002.
[53]Samy Bengio. HMM and IOHMM modeling of EEG rhythms for asynchronous BCI systems[J]. European Symposium on Artificial Neural Networks,2004:199-204.
[54]B. Obermaier, C. Guger, C. Neuper. Hidden Markov models for online classification of single trial EEG data[J]. Pattern Recognition Letters,2001,22(1):1299-1399.
[55]T. M. Vaughan, W. J. Heetderks, L. J. Trejo. Brain-computer interface technology:A review of the second international meeting [J]. IEEE transaction on Rehabilitation Engineering,2003,11:94-109.
[56]赵春临,郑崇勋,赵敏.基于脑电功率谱-连续隐马尔科夫链的精神疲劳分级模型[J],西安交通大学学报,2007,41：1474-1478.
[57]Ali Ozgur Argunsah. AR-PCA-HMM Approach for Sensorimotor Task Classification in EEG-based Brain-Computer Interfaces[A]. International Conference Pattern Recognition[C],2010.113-116.
[58]顾燕红.自适应剖面隐马氏模型软件言之[D].上海：上海大学,2005.