Quantifying Air Traffic Controller Mental Workload Nicolas Suarez nstetzlaff@e-crida.enaire.es Patricia López pmldefrutos@e-crida.enaire.es Eva Puntero epuntero@e-crida.enaire.es Sara Rodriguez srodriguezg@e-crida.enaire.es Fecha 1 Titulo
Titulo Fecha 2
Index 1. Introduction 2. Modelling ATCo workload 4. Validation 5. Conclusions 3. Workload Analysis Component (WAC) Titulo Fecha 3
Air Traffic Complexity Mental Workload INTRODUCTION Fecha 4 Titulo
Basic Concepts • Difficulty perceived by • Ability to perceive and ATCo respond to variables • Prior experience • Prior developed personal constructs Air Traffic ATCo Mental Complexity Complexity • Mental structures used to interpret & respond • How each one relates, overlaps & influences the others determines cognitive complexity Personal constructs
How do we define complexity Currently: • “…measure of the difficulty that a particular traffic situation will present to an air traffic controller…” SESAR • “number of simultaneous or near - simultaneous interactions of trajectories Factors impacting complexity: in a given volume of airspace” • Geometrical nature of the air traffic • Operational procedures and practices used to handle the traffic • Characteristics and behaviour of individual controller Titulo Fecha 6
Two basic hypotheses Assessing the ATCo Complexity inside a mental workload inside sector is a function of a sector will provide an the ATCo mental estimation of the workload associated complexity Titulo Fecha 7
What is ATCo Mental Workload? The ATCo Mental Workload Framework Multiple Resource Workload Model (MWM) Experimental system MODELLING ATCO WORKLOAD Fecha 8 Titulo
Primitive Operator Tasks Human Behavior can be represented/segmented in Primitive Operator Tasks Perception Responding Compre- Strategic Decision (Visual & (Manual & hension Thinking Making Auditory) Verbal) Visual & Auditory Processing Visual Auditory Search Scan Monitor Fixate Track with with Listen Audio Object Object pattern pattern Signal C. D. Wickens, “Multiple Resources and Mental Workload,” Hum. Factors J. Hum. Factors Ergon. Soc., vol. 50, no. 3, pp. 449– 455, Jun. 2008. 9
Primitive Operator Tasks Human Behavior can be represented/segmented in Primitive Operator Tasks Perception Responding Compre- Strategic Decision (Visual & (Manual & hension Thinking Making Auditory) Verbal) Central Processing Comprehension Strategic Thinking Decision Making Recall Recognise Select Compare Compute Decide C. D. Wickens, “Multiple Resources and Mental Workload,” Hum. Factors J. Hum. Factors Ergon. Soc., vol. 50, no. 3, pp. 449– 455, Jun. 2008. 10
Primitive Operator Tasks Human Behavior can be represented/segmented in Primitive Operator Tasks Perception Responding Compre- Strategic Decision (Visual & (Manual & hension Thinking Making Auditory) Verbal) Responding Processing Manual Verbal Press Move Reach Push & Say a with Write Type …… with Grasp Touch Object hold message foot pattern C. D. Wickens, “Multiple Resources and Mental Workload,” Hum. Factors J. Hum. Factors Ergon. Soc., vol. 50, no. 3, pp. 449– 455, Jun. 2008. 11
What is workload? Task load Required psychological resources Workload Available psychological Resources
Demanded Mental Available Mental Threshold: Resources (Task Load): Resources: • Physical and mental • Physical and mental • Value beyond which activities demanded abilities (perceptual Demanded Mental to carry out actions, cognitive Resources (Task perceptual actions, actions and motor Load) exceeds the cognitive actions and skills) that an ATCo Available Mental motor skills has available to Resources provide the control • Empirical research service and psychological theories of human • Psychological factors cognitive processes (e.g. fatigue, stress) shape the available resources Titulo Fecha 13
Mental Workload Framework Overload Threshold Underload Operating modes Available Mental Resources Response Operating Concept How a Controller does it Decision Making What a Controller needs Control to do Events Situationa Awarenes Flight What aircraft do l s Events Demanded Mental Resources (Task Load)
This is nice, but how do we use it? Develop an experimental system that is able to estimate the workload Identification of control events • Required cognitive channels • Interference matrix • Resolution of the Matrix Multiple Resources Identification of flight Workload Model events algorithm (MWM) Titulo Fecha 15
Flight Event Controller Event Cognitive Process Layer: Layer: Layer: • Identification & • Actions expected • Manner in which use of flight from an ATCo an ATCo performs events a specific action • e.g. solve • Traffic demand / conflict • Operating Traffic data or concept sector • Mental processes configuration data required to • Aircraft behaviour perform said within a specific actions airspace Titulo Fecha 16
What is WAC? Use of WAC WORKLOAD ANALYSIS COMPONENT (WAC) Fecha 18 Titulo
Workload Analysis component Stand alone workload estimation and measurement Calculates workload per sector / volume
A sample application of WAC Titulo Fecha 20
Use of mental workload to estimate sector capacity Assessment of the coherency of mental workload results through the comparison of Predicted and Perceived Workload Further work VALIDATION Fecha 21 Titulo
Validating the model Not easy Difficult to measure the actual workload Using indirect methods to perform the validation • Use of Mental Workload to Estimate Sector Capacity • Assessment of the coherency of mental workload results through the comparison of Predicted and Perceived Workload
Use of mental workload to estimate sector capacity Access data base to identify actual capacity • Similar or Equal • Representative sample of actual traffic workload assessment were accurate • Use WAC component to • Different workload estimate capacity assessment were accurate Estimate actual capacity using Compare both values workload Titulo Fecha 23
Assessment of the coherency of mental workload results through the comparison of Predicted and Perceived Workload 60 Record ISA (Instantaneous Self Assessment) • Normalized values 40 Calculate workload using WAC 20 Issues 0 • Low correlation (R = 0.75, p = 0.35) Cognitive WL ISA WL 60 Causes: 40 • Time shift • Need to improve calibration of non-nominal events 20 With correction 0 • Strong correlation (R = 0.93, p = 0.14) Cognitive WL ISA WL Titulo Fecha 24
Do we think that the validation work is completed? SESAR EXE-04.07.01-VP-003 NO Resolving complexity by dynamic management of airspace • December 2014 EXPERIMENT RESULTS BETWEEN WORKLOAD MEASURES Subjective Measures Physiological measures Blink-related parameters Fixation-related parameters Number of Correlation factors Blink Eye closure Blink Fixation time on NASA-TLX MCH score fixation on frequency fraction duration Regions Regions R 0,908 0,908 0,927 0,741 0,910 0,904 0,774 Information flow rate p 0,003 0,002 0,009 0,035 0,002 0,002 0,024 R 0,894 0,934 0,934 0,818 0,878 0,656 NASA-TLX p 0,003 0,001 0,001 0,013 0,004 0,077 R 0,858 0,858 0,840 0,902 0,819 MCH score p 0,006 0,006 0,009 0,002 0,001 Ha, Kim and Seong (2006) Titulo Fecha 25
Estimating complexity How to improve the model CONCLUSIONS Fecha 26 Titulo
Estimating complexity Mental Workload framework: Using workload as an Key enabler for several indicator of air traffic • Demanded resources SESAR concepts complexity • Available resources • Thresholds. Workload estimation Workload Analysis algorithm (MWM) Component (WAC)
Validation Needs to be improved • More calibration experiments • Identifying more direct workload On-going methods Results But… • Workload estimations are accurate • Improvements must be made on both the framework and the • Useful for ATM algorithm • Suggest relationship between perceived & predicted mental workload Titulo Fecha 28
Areas of improvement Introducing Enhancement of the psychological factors Introducing dynamic operating mode (fatigue, stress & thresholds definition emotion) Full development and Impact of system integration of situational automation features awareness and decision making processes Titulo Fecha 29
Centro de Referencia I+D+i ATM Titulo Fecha 30
Recommend
More recommend
