measurement of mental workload associated with agricultural spraying
DESCRIPTION
Measurement of Mental Workload Associated With Agricultural Spraying. Asit K. Dey and Danny D. Mann. Department of Biosystems Engineering University of Manitoba Winnipeg, MB R3T5V6. Prepared for 2008 Annual Meeting of CSBE, Vancouver, BC, 13 th -16 th July. - PowerPoint PPT PresentationTRANSCRIPT
Measurement of Mental Workload Associated With
Agricultural Spraying
Asit K. Dey and Danny D. MannDepartment of Biosystems Engineering
University of Manitoba Winnipeg, MB R3T5V6
Prepared for 2008 Annual Meeting of CSBE, Vancouver, BC, 13th -16th July
Two Major Tasks of Agricultural Spraying
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Task 1: Steering a sprayer along a predefined path in response to a navigation device (PT)
Predefined Path
Task 2: Monitoring and controlling the rear- attached boom (ST)
Joystick
Boom
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Key findings of our previous study*• Agricultural spraying is a dynamic dual task (i.e. driving and
monitoring) conducted under day, dusk, and night illumination levels.
• A sprayer operator sprays 16.5 h in a day that includes all the above changing environments.
• Moreover, various terrain conditions (i.e., rolling, flat, or field with obstacle) imposes additional difficulty.
*Dey, A and D. Mann. 2008. A complete task analysis to measure the workload associated with operating an agricultural sprayer equipped with a navigation device. Submitted to Applied Ergonomics.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Key findings of our previous study (contd.)
• Most of the sprayer operators use a GPS lightbar to guide the sprayer along a pre-defined path.
• The modern cabs are equipped with a mapping display, application display, entertainment unit, and two-way radio communication unit.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
• Event detection and driving performance degrades whenever new information cues are introduced in the operators cab (Boer 2000).
• Degradation in driving performance will result in more skips and overlaps of crop inputs.
• Degradation in event detection may be linked to the safety of the operators.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Boer, E.R. 2000. Behavioral entropy as an index of workload. In Proceedings of the IEA 2000/HFES 2000 Congress, 3/125-3/128, San Diego, CA.
• The degradation of the driving performance can be minimized and operators comfort and safety can be enhanced by approaching a human centric design.
• The study of mental workload helps us achieving the above goal.
• Till today, there is no published literature that explored the effect of illumination, terrain difficulty, and task levels on the mental workload of an agricultural sprayer operators.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Mental Workload
Definition: The proportion of mental resources invested to meet task demand.
• Excessive workload can affect selective attention, lead towards in-efficient sampling, and results in poor performance.
• Therefore, the role of a design engineer should be to keep the mental workload in an optimum zone below the workload redlines.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
• To investigate the effect of illumination (day and night), difficulty (low and high), and task levels (Single vs. Dual) on the mental workload of agricultural sprayer operators guiding a sprayer in response to a GPS lightbar.
ObjectivesIntroduction Conclusions AcknowledgementResultsMethodology
Agricultural Driving SimulatorInside of the Driving Simulator Simulation of the Field View Rear Display
•The participants drove a fixed-base agricultural driving simulator in response to a red commercial lightbar.•The simulator was equipped with a torque and visual feedback unit
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
• Participants
16 male university graduate students.
• Age Group
<25 (8), 26-30 (2), 31-40 (3), 41-60 (3)
• Training: participants were trained to drive the above tractor simulator.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Mental Workload Measurement
Mental workload was measured by:• Driving performance: lateral root mean square
error (cm).• Monitoring Performance: Reaction Time (s)• 0.1 Hz power of hear rate variability (a.u.)• Dynamic Spectrogram• P300 latency (s)• Eye-glance behaviour (% time spent)• NASA-Task Load Index and Simplified
subjective workload assessment technique (SSWAT) (a.u.)
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Experimental DesignTasks Day Night
PT PTDAL REST
PTDAH REST
PTNL REST
PTNH REST
ST STDAL STDAH STNL STNH
DT DTDAL DTDAH DTNL DTNH
5 min 5 min
•Each participant received 12 randomized sessions.
•The experimental time was 3 h/participant.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Percentage change in workload measures due to change in illumination
and difficulty levels
Workload measures Da-N (%) L-H (%)
LRMSE 3.9 6.8
RT 3.1 26
0.1 Hz HRV 37.8 116.5
P300 latency (Fz site) 16.45* 1.5*
NASA-TLX 3.9 6
SSWAT 17.4 9.1*
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodologyObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Da= day, N= night, L= low, and H= high
Percentage change in workload measures due to change in task levels
PT-DT(%) ST-DT(%)
LRMSE 18.8* n.a.
RT n.a. 108.1*
HRV -8 -6.6
P300 latency -2.3* 11.7*
NASA-TLX 22* 18.4*
SSWAT 41.1* 92.3*
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodologyObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
PT= primary task, ST= secondary task, DT= dual task, n.a.= not applicable
Percentage of time spent looking at various sectors (outside, lightbar, left boom, right boom) for day (DA) and night (N) illumination, and low (L) and high (H) difficulty levels under driving only condition
40.2%42.4%45.6%45.7%
59.8%57.6%
54.4%54.3%
30.0%
35.0%
40.0%
45.0%
50.0%
55.0%
60.0%
65.0%
PTDAL PTDAH PTNL PTNH
Tim
e sp
ent
(%)
OUTSIDE
LIGTHBAR
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Percentage of time spent looking at various sectors (outside, lightbar, left boom, right boom) for day (DA) and night (N) illumination, and for low (L) and high (H) difficulty levels under monitoring only condition.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
16.2%
18.0%16.8% 15.1%
0.2% 0.2% 1.0% 0.4%
39.5% 38.5% 39.5%40.7%
42.7%42.3%44.5% 44.3%
0.0%5.0%
10.0%15.0%
20.0%25.0%
30.0%35.0%
40.0%45.0%
50.0%
STDAL STDAH STNL STNH
Tim
e sp
ent (
%)
OUTSIDE
LIGHTBAR
RIGHT
LEFT
34.8%33.6%29.7%34.0%
35.9%38.4%37.6%
36.5%
15.6%16.0%13.7%14.6%
14.9%15.9%
13.9%14.8%
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
45.0%
DTDAL DTDAH DTNL DTNH
Tim
e sp
ent
(%)
OUTSIDE
LIGHTBAR
RIGHT
LEFT
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Percentage of time spent looking at various sectors (outside, lightbar, left boom, right boom) for day (DA) and night (N) illumination, and for low (L) and high (H) difficulty levels under dual task condition.
PTDAL
PTDAH
PTNL
PTNH
DTDAH
DTDAL
DTNH
DTNL
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
STDAH
STDAL
STNH
STNL
DTDAH
DTDAL
DTNL
DTNH
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Conclusions• The performance measures, P300 latency,
subjective rating scales showed a trend that night illumination was more demanding. The main effect of illumination was significant for P300.
• Similarly, the above measures showed that high difficulty was more demanding. Only the P300 and SSWAT was able to differentiate between low and high difficulty at p<0.05.
• The 0.1 Hz HRV data showed driving under day illumination or under low difficulty were more demanding.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
Conclusions (contd.)
• Spectrogram showed that PT and DT under day driving, and ST under night driving was more demanding. Moreover, low difficulty was more demanding than day driving.
• All the measures significantly revealed that the dual task was more demanding than single task level.
• Under any illumination, difficulty, or task levels (except ST), participant spent more time looking at the lightbar. Therefore, lightbar is an important source of guidance information.
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology
• U of M Graduate Fellowship
• Department of Biosystems Engineering of the University of Manitoba.
• Participants
ObjectiveIntroduction Conclusions AcknowledgementResultsMethodology