Ergonomics
ISSN: 0014-0139 (Print) 1366-5847 (Online) Journal homepage: http://www.tandfonline.com/loi/terg20
Effects of vertical vibration on passenger activities: writing and drinking COLIN CORBRIDGE & MICHAEL J. GRIFFIN To cite this article: COLIN CORBRIDGE & MICHAEL J. GRIFFIN (1991) Effects of vertical vibration on passenger activities: writing and drinking, Ergonomics, 34:10, 1313-1332, DOI: 10.1080/00140139108964870 To link to this article: http://dx.doi.org/10.1080/00140139108964870
Published online: 31 May 2007.
Submit your article to this journal
Article views: 24
View related articles
Citing articles: 53 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=terg20 Download by: [University of Cambridge]
Date: 05 November 2015, At: 23:02
Effects of vertical vibration on passenger activities: writing and drinking COLINCORBRIDGE* and MICHAEL J. GRIFFIN Human Factors Research Unit, Institute of Sound and Vibration Research, The University, Southampton SO9 5NH, UK
Downloaded by [University of Cambridge] at 23:02 05 November 2015
Keywordr: Whole-body vibration; Task difficulty; Handwriting ability; Grasping ability.
Two laboratory studies have investigated how handwriting ability and holding a cup of liquid depend on the characteristics of whole-body vertical vibration. The effects of vibration magnitude (0.16 to 2.5 msd2 r.m.s.), vibration frequency (0.5 to 10 Hz),and vibration duration (2 cycles to 10 s) on handwriting were studied with 20 subjects. Subjects were asked to copy letters of the alphabet by writing on a hand-held surface. Writing speed decreased and subjective ratings of writing difficulty increased with increasing vibration magnitude, particularly in the frequency range 4 to 8 Hz. Writing difficulty also increased with increasing duration of vibration. A 10 s exposure to 5 Hz vibration at 2.0 msd2r.m.s. resulted in subjective estimates corresponding to 'extremely difficult'. The effects of vibration magnitude (0.63 to 1.6 msm2r.m.s.), vibration frequency (0.5 to 10 Hz),and vibration duration (2 cycles to 10 s) on the spilling of liquid from a hand-held cup were also investigated in a group of 20 subjects. The probability of spilling the liquid, the quantity of liquid spilt, and subject's estimates of the probability of spillage were determined for all conditions. Greatest interference with the task occurred at 4 Hz, with the lowest vibration magnitude (0.63 ms-* r.m.s.) causing measured and estimated spillage probabilities of approximately 85%. The interference was much less at other frequencies, with 0.63 ms-I r.m.s. causing less than 10% measured probability of spillage,below3 Hz and above 5 Hz. The estimated probability of spillage was generally greater than the observed probability of spillage when the spillage probability was low, but less than the observed probability when the spillage probability was high. Increasing the duration of vibration increased the'probability of spillage, and also increased the volume of liquid spilt. 1. Introduction This paper is concerned with the effects of vertical vibration on two activities, writing and holding a cup of liquid, often undertaken by passengers in railway trains and in other vibration environments. Most of the published investigations of the effects of vibration of performance have concerned tasks performed by military personnel, such as the control of joysticks in aircraft or tanks. In a summary of the research into the effects of vibration on task performance McLeod and Griffin (1986) classified tasks into three categories:
(i) type A tasks, in which the 'subject controls the hand freely in space: examples include reaching and pointing. In some type A tasks the hand may hold an object which will itself be affected by motion, such as fluid in a cup'. (ii) type B tasks, in which the 'subject's hand manipulates a control at a fixed *Now at the Department of Psychology, University of Nottingham, University Park, Nottingham NG7 2RD, UK. 0014-0 13919 1 $3.00 0 1991 Taylor & Francis Ltd.
C.Corbridge and & J. Griffin i . position attached to the vibrating structure: examples include the operation of joysticks and knobs'. (iii) type C tasks, in which the subject 'performs a single, discrete operation, such as changing a switch setting or pressing a button'. This type of task may often be preceded by a type A task, in which the hand moves through space in order to locate the control.
Downloaded by [University of Cambridge] at 23:02 05 November 2015
Both holding a cup of liquid in space and writing on a hand-held surface might be classified as type A tasks. This type of task, together with type C tasks have been the subject of little empirical research. 1.1. Writing Huddleston (1964) found that the size of seated subjects' handwriting on a hand-held surface increased when they were exposed to vibration. Subjects experienced particular difficulty when writing while exposed to 4.8 and 6.7 Hz whole-body vibration at 4.9 r n r 2 r.m.s., with 4.8 Hz causing most difficulty. There was a progressive reduction in script height as the vibration frequency was increased to 9.5 and 16 Hz. Of the 540 words copied in each condition, the numbers illegible with 4.8, 6.7, 9-5, and 16 Hz vibration were 31, 31, 17, and 0 respectively. A second study (Huddleston 1965) confirmed the adverse effects of vibration on writing performance but failed to show any difference between the effect of vibration of 4.8 and 6.7 Hz. Murata (1 973) measured the time required to write a grid reference, part of a map reading and writing task, by 19 military personnel exposed to vertical random vibration. The subjects performed the writing task on a table. The vibration conditions were designed to simulate the motion encountered in armoured fighting vehicles moving over metalled surfaces and cobbles and were presented at a magnitude of 2 m r 2 r.m.s. For exposure durations up to 20 min there was little consistent effect of vibration on writing time. There was evidence from tests conducted on an additional three subjects that exposures up to 1 h may decrease performance. The author suggested that some subjects worked harder during exposure to vibration so as to maintain performance and that this could be sustained for durations of up to 20 min but not up to 1 h. Cambell (1 974) exposed 12 military personnel to sinusoidal vibration at frequencies of 1-5, 3.0, and 6.0 Hz for periods of 3 h at a vibration magnitude of 2.5 , m r 2 r.m.s. The test conditions were intended to be similar to those encountered in armoured fighting vehicles and subjects were required to perform a battery of reading and writing tests, again on a table. Vibration conditions were found to impair the performance of a maze drawing task as assessed by both the length of maze completed and the number of errors. Performance of map reading and writing tasks showed similar adverse effects of vibration: both the time required to complete the task and the number of errors increased. The adverse effects of vibration also increased as vibration frequency increased and 6 Hz vibration was found to be particularly detrimental to writing performance. (In common with findings reported by Huddleston [1964], the author noted that in adverse conditions the writing became larger yet more difficult to read.) A study using similar performance measures was reported by Farmilo and Bennett (1975). They reported that the effect of exposure duration was negligible when compared to the overall decrement due to vibration.
Vertical vibration and passenger activity
1315
Downloaded by [University of Cambridge] at 23:02 05 November 2015
The published studies suggest that considerable difficulty may be experienced when attempting to write while exposed to whole-body vibration in the frequency range 4 to 6 Hz. However, the magnitudes of vibration used in previous studies were severe. There has been no attempt to investigate the effect of vibration magnitude on writing performance, or the effect of vibration conditions typical of those encountered by fare-paying passengers. 1.2. Holding a cup of liquid Adverse effects of vibration are commonly observed when drinking in moving vehicles but there has been little relevant research. Only one previous study, Whitham and Griffin (1978), is known to have investigated the effects of vibration on this activity. In their study, 12 male subjects were exposed to vibration while seated on a rigid flat seat without a backrest. The subjects were exposed to motion of the seat in the three translational axes (fore-and-aft, lateral, vertical) and in the three rotational axes (roll, pitch, and yaw). Sinusoidal vibration at the preferred one-third octave band centre frequencies between 1.6 Hz and 6.3 Hz were used (1.6 to 10 Hz for the vertical axis). All stimuli had a duration of 10 s and, after each had been experienced, subjects were required to indicate if they had spilt the liquid. The findings were expressed as the magnitudes of vibration at which 25% and 50°h of the subjects spilt the liquid, as a function of vibration frequency in each axis. High sensitivity was found to 4 Hz vibration in all axes. In the rotational axes subjects were less affected by yaw vibration than roll or pitch vibration. Subjects were less affected by motion in the horizontal axes than in the vertical axis. Vertical vibration with a magnitude of 0.3 ms-I r.m.s. at 4 Hz resulted in 50% of the subjects spilling the liquid. The second experiment reported here extends the findings reported by Whitham and Griffin (1 978). The range of stimuli was extended to include frequencies down to 0.5 Hz. This was necessary since rail vehicle vibration has significant motion at the low end of the frequency range from 0.5 to 5.0 Hz (see Grifiin 1990). 2. Experiment 1: writing
2.1. Experimenlal design The method of magnitude estimation was used to compare writing difficulty caused by 'test' motions with that caused by a 'reference' motion. The 'reference' motion was an octave band of random motion centred on 1 Hz, magnitude 0.63 m r 2 r.m.s. and duration 20-5 s. The difficulty caused by this condition was assigned a writing difficulty of 100. Each subject was then presented with a series of five 'test' motions. After each 'test' motion the subject was required to assign a number to the motion corresponding to the degree of writing difficulty experienced relative to that of the 'reference' motion. For example, if the degree of writing difficult was twice that of the 'reference' motion, the 'test' motion would have been assigned a value of 200. Conversely, if the 'test' motion produced only half the writing difficulty of the 'reference' motion it would have been assigned a value of 50. Each of the 'test' motions consisted of the same octave bandwidth random motion centred on 1 Hz as the 'reference' motion, but with an added sinusoidal vibration component. The frequency, duration and magnitude of the 40 additional components shown in table 1 were added to the octave bandwidth random motion centred on 1 Hz to produce 40 'test' motions. The only constraint was that the
C. Corbridge and M.J. Griffin
1316
Downloaded by [University of Cambridge] at 23:02 05 November 2015
additional sinusoidal vibration did not occur in the first or the last 5 s period of the background motion. The background random motion, which defined the length of the writing task, was typical of that which occurs in modem high speed rail vehicles travelling over good track. The sinusoidal vibration was to provide information on the magnitudes and durations of vibration at various frequencies which would affect writing. The 40 'test' motions were randomly divided into eight groups of five with a presentation of the 'reference' motion preceeding each group. The 'reference' motion was therefore presented to each subject on eight separate occasions. The sequence of presentation was randomized across subjects. Table 1. Additional sinusoidal vibration components added to the background random motion to produce the 'test' motions. Frequency Hz Duration
0.5
1.0
2.0
3.15
4.0
5.0
6.3
8.0
10.0
Magnitudes ms->r.m.s.
2 cycles 2s
x
x
x
x
x
x x
x
x
x
x
x
x
x
x
x
x x x x
1.6 1.6 1.6 0.63, 1.0, 1.6
4s 10s
x x
In a separate part of the experiment the magnitude estimation technique was used to investigate the effect of the magnitude of random vibration on writing difficulty. Subjects compared the effect of the 'reference' motion with that of five random 'test' motions and a control condition without vibration. The random 'test' motions were centred on 1 Hz, had the same spectrum and duration as the 'reference' motion (20.5 s) but magnitudes of 0-16, 0.3 15, 0.63, 1-25, and 2.5 m c 2 r.m.s. In a further part of the experiment the subjects were asked to rate the writing dificulty produced by the above five octave band random test motions on the following semantic scale: a not difficult; a little difficult; fairly dimcult; a difficult; a very difficult; a extremely difficult.
The use of both a semantic scale and the magnitude estimation technique enabled the magnitude estimates of writing difficulty to be located on an 'absolute' scale. This was then used as an aid to interpreting the magnitude estimates of writing difficulty produced by the 'test' motions containing sinusoidal components. The determinations of the effect of vibration magnitude on writing difficulty using both the magnitude estimation technique and the semantic rating scale were carried out at both the beginning and the end of the experimental session. This was to enable the effect of experience of vibration exposure and practice at both the task and the psychophysical methods employed in the study to be investigated. The form of the full experimental session, which lasted approximately 1 h, is shown in table 2.
J'ertical vibration and passenger activity
1317
Table 2. Structure of the experimental session. Section of experimental session 1
2
Downloaded by [University of Cambridge] at 23:02 05 November 2015
3
4 5
.
Description
Assessment of random vibration magnitude on writing difficulty using the semantic scale. Assessment of random vibration magnitude on writing difficulty using the magnitude estimation ~echnique. Assessment of 'test' motions containing added sinusoidal vibration components on writing difficulty using the magnitude estimation technique. As section 2. As section 1.
2.2. Subjects Ten male and ten female subjects participated in the study. The subjects, who were paid, were students, research workers and technicians at the University of Southampton. Male subjects were aged 22 to 3 1 years (mean 25.8 years), weighed 6 1 to 78 kg (mean 7 1.1 kg), and were 1.68 to 1.88 (mean 1.80 m) tall. Female subjects were aged 20 to 28 years (mean 23-2 years), weighed 47 to 6 1 kg (mean 54-4kg), and were 1.58 m to 1.71 m (mean 1.66 m) tall. Five of the male subjects and one female subject used the left hand as the preferred hand when writing. 2.3. Writing task Subjects were required to copy a row of 39 lower-case letters of the alphabet printed on one line of a sheet of paper onto a line marked out below. The first 26 characters of the row consisted of the letters of the alphabet in random order. The remaining 13 characters were randomly selected from the alphabet with the constraint that a character did not occur more than twice in a row. Each sheet contained six rows of characters and six associated lines of spaces. The writing tasks performed during a 'reference' motion and the five subsequent 'test' motions were therefore contained on a single sheet. Upon completing one 'reference' motion and five 'test' motions subjects removed the sheet from the top of the clipboard and placed it at the bottom. The thickness of paper on which the subjects were writing therefore remained constant throughout the experiment. The subjects were required to copy the characters as quickly and accurately as possible during the presentation of each vibration stimulus, consistent with producing legible handwriting. They were instructed to use cursive script, that is, joining up the letters as in normal handwriting. The time allowed for the task was 20-5 s (the duration of thevibration stimuli); a 2 kHz tone of magnitude 65 dB(A) at the subjects' ears signalled the start and the end of each writing task. 2.4. Equipment A 1 m stroke vertical hydraulic vibrator located at the Southampton University Institute of Sound and Vibration Research was used to generate the vibration stimuli. This device was designed for human response to vibration studies and its typical performance when generating sinusoidal stimuli has been documented
C. Corbridge and M. J. Griffin
Downloaded by [University of Cambridge] at 23:02 05 November 2015
previously, Corbridge and Griffin (1986). Subjects were seated on a rigid flat seat with a backrest inside a wooden cabin mounted on the vibrator table; the interior was illuminated by a single 24 W bulb. Subjects had no external view of the laboratory. The rigid flat seat on which the subjects sat against a rigid flat backrest has been described previously by Corbridge and Griffin (1 986) and has been shown to have unity transmissibility to vibration in the frequency range used in this study.
2.5. Procedure Subjects were screened for medical contra-indications (BSI 1973) and seated inside the cabin. For safety reasons subjects were required to wear a loose-fitting lap strap during the experiment. Subjects were then given instructions relating to the conduct of the writing task, allowed to ask questions and provided with a clipboard, response sheets and a Parker Rollerball pen fitted with a new refill. Subjects then practised the task five times under static conditions. They were then provided with a copy of the semantic scale described in section 2.1 and required to 'consider you are travelling on a train. Estimate the writing dificulty that you experience.' Subjects were verbally instructed to attend specifically to the writing dificulty inherent in the task and not to any discomfort produced by the various motions. Subjects were then exposed to the five magnitudes of octave bandwidth random motion from 0.16 to 2.5 ms-2 r.m.s. and an additional control condition without vibration whilst performing the writing task. The six stimuli were presented to the subjects in random order; following each stimulus presentation the subjects assessed the writing difficulty using the semantic scale. Following completion of section 1 of the experimental session, subjects were given instructions regarding the use of the magnitude estimation technique. Subjects were given the opportunity to practice the technique in the assessment of the length of lines drawn on paper. Upon satisfactory completion of this training, subjects compared the writing difficulty experienced with an octave bandwidth random 'reference' motion, magnitude 0-63m r 2 r.m.s., with that produced by five 'test' motions. These 'test' motions consisted of the same octave bandwidth random motion presented at five vibration magnitudes between 0.1 6 and 2.5 msm2r.rn,s. On completion of section 2 of the session, subjects immediately commenced section 3, the main part of the study. In this section subjects used magnitude. estimation to assess the writing difficulty of the 40 'test' motions which consisted of the octave bandwidth random motion, magnitude 0-63 m r 2 r.m.s., with an added sinusoidal component. The investigation of the effect of random vibration magnitude on writing difficulty was then repeated using the magnitude estimation technique (session section 4) followed by use of the semantic rating scale (session section 5). 2.6. Results Vibration had two main effects on writing: it reduced both the number of characters copied and their legibility. The number of characters copied was selected as the objective measure of writing performance most amenable to consistent scaring. The maximum number of characters which could be copied during each vibration exposure was 39. The maximum score was achieved in fewer than 5% of all exposures. Each individual score for any 'test' motion was converted to a percentage of the score obtained when the subject experienced the preceeding 'reference' motion. This reduced the effect of any systematic learning. (Raw scores indicated
Vertical vibration and passenger activity
1319
Downloaded by [University of Cambridge] at 23:02 05 November 2015
that the median number of characters copied when exposed to the 'reference' motion increased from 31 on the first presentation to 33.5 on the eighth and final presentation-indicating that a slight learning effect occurred over the experimental session.)
2.6.1. Effect of magnitude of random vibration on writing diflculty. Few significant differences were observed between maIe and female subjects in either the subjective ratings or the magnitude estimates. In only one of the 40 'test' conditions did the subjective responses of the male subjects differ significant from those made by female subjects (pt0.05, Mann-Whitney U-test). No differences were obtained between the two groups in objective measures of writing performance. In the subsequent sections the results from the two groups have been combined. No significant differences (Wilcoxon matched-pairs signed rank test, p>0.10) existed between ratings made at the beginning and end of the experimental session. The octave bandwidth random motion centred on 1 Hz with a magnitude of 0-63 ms-* r.m.s. (the 'reference' motion used in section three of the experimental session) resulted in a median rating of 0.5 (i.e. between 'not dificult' and 'a little difficult'). Analysis of the results obtained from the combined group of subjects at the beginning and end of the experimental session showed that in only one condition did the results differ significantly (pt0.02). (Estimates of writing difficulty in response to random motion with a magnitude of 1.25 ms-I r.m.s. made at the beginning of the experimental session were higher than those made at the end.) Figure 1 shows the magnitude estimates of writing dimculty expressed as a function of the rated writing difficulty obtained from the 20 subjects at the beginning and end of the experimental session.
I
2
3
4
5
6
Semantic rating of wrHing dinicuhy
Figure I . Magnitude estimates of writing difficulty as a function of rated difficulty of writing obtained from 20 subjects with octave bandwidth random vibration centred on 1 Hz with ~ Semantic rating: 1 =not difficult; 2=a little vibration magnitudes up to 2.5 r n ~ ' r.m.s. difficult; 3=fairly difficult; 4=difficult; 5 =very dificult; 6 =extremely difficult.
The least squares regression analysis relating magnitude estimates Q to semantic ratings ( x ) isdescribed by the following function: 10g,o= 1.18 10g~fi-k1.72
'
C. Corbridge and M.J. Grifln
1320
Using this equation, magnitude estimates corresponding to the various semantic descriptors of rated writing dificulty were calcuIated as in table 3. Table 3. Semantic descriptors and calculated magnitude estimates of writing difficulty.
Downloaded by [University of Cambridge] at 23:02 05 November 2015
Semantic descriptor of rated writing dificulty (numerical value) Not difficult A little difficult Fairly difficult Difficult Very difficult Extremely difficult
Calculated magnitude estimate of writing difficulty
(1)
(2)
(3) (4) (5) (6)
None of the subjects rated any of the random motions as being a condition in which writing was 'Extremely Difficult.' The extrapolation of the function to calculate the magnitude estimate of writing difficulty corresponding to this descriptor as shown in table 3 is therefore based on the assumption that magnitude estimates of writing difficulty will continue to increase logarithmically with rated difficulty. In the following section, magnitude estimates corresponding to the semantic descriptors 'A Little Difficult' and 'Very Difficult' have been marked on the figures to aid the interpretation of the magnitude estimates. 2.6.2. Effect offeguency of sinusoidal vibration (a) 10 s duration: Figure 2 illustrates the variability in subjective estimates and objective measures of writing difficulty as a function of vibration frequency for sinusoidal components added to the background random motion. The curves show the effect of adding 10 s sinusoidal components with magnitudes of 0.63 and 1.6 msm2r.m.s. to the random motion.
0
1
2
3
4 5 6 7 Frequency (Hz)
8
9
1
0
Frequency (HI)
Figure 2. Median, 25th, and 75th percentile subjective estimates and objective measures of writing difficulty as a function of frequency for sinusoidal vibration components added to background random motion. Sinusoidal components of duration 10 s: -0.63 rns+ r.rn.s.; - 1.6 ms-I r.m.s.
1321
Vertical vibration and passenger activity
Median subjective estimates of writing difficulty when 10 s sinusoidal components were added to the background motion at magnitudes of 0.63, 1.0, and 1-6 ms-I r.m.s. are shown in figure 3. Friedman non-parametric analyses of variance showed that there was a significant effect of vibration frequency on magnitude estimates for each of the three magnitudes of added sinusoidal vibration @0.05). Median estimates of writing difficulty exceeded 1 18 ('a little difficult') at only one frequency: 3.15 Hz. Writing speed also showed no significant effect of vibration frequency when the duration was limited to 2 cycles of motion (p>0- 10). The median writing performance did not fall below 93% of the number of characters copied in the relevant 'reference' conditions.
2.6.3. Eflect of duration of sinusoidal vibration: The effects of increasing the duration of added 5 Hz and 10 Hz sinusoidal components from 2 cycles to 10 s on subjective and objective measures of writing performance are shown in figure 4. There was a significant effect of vibration duration at both 5 and 10 Hz (p
ISSN: 0014-0139 (Print) 1366-5847 (Online) Journal homepage: http://www.tandfonline.com/loi/terg20
Effects of vertical vibration on passenger activities: writing and drinking COLIN CORBRIDGE & MICHAEL J. GRIFFIN To cite this article: COLIN CORBRIDGE & MICHAEL J. GRIFFIN (1991) Effects of vertical vibration on passenger activities: writing and drinking, Ergonomics, 34:10, 1313-1332, DOI: 10.1080/00140139108964870 To link to this article: http://dx.doi.org/10.1080/00140139108964870
Published online: 31 May 2007.
Submit your article to this journal
Article views: 24
View related articles
Citing articles: 53 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=terg20 Download by: [University of Cambridge]
Date: 05 November 2015, At: 23:02
Effects of vertical vibration on passenger activities: writing and drinking COLINCORBRIDGE* and MICHAEL J. GRIFFIN Human Factors Research Unit, Institute of Sound and Vibration Research, The University, Southampton SO9 5NH, UK
Downloaded by [University of Cambridge] at 23:02 05 November 2015
Keywordr: Whole-body vibration; Task difficulty; Handwriting ability; Grasping ability.
Two laboratory studies have investigated how handwriting ability and holding a cup of liquid depend on the characteristics of whole-body vertical vibration. The effects of vibration magnitude (0.16 to 2.5 msd2 r.m.s.), vibration frequency (0.5 to 10 Hz),and vibration duration (2 cycles to 10 s) on handwriting were studied with 20 subjects. Subjects were asked to copy letters of the alphabet by writing on a hand-held surface. Writing speed decreased and subjective ratings of writing difficulty increased with increasing vibration magnitude, particularly in the frequency range 4 to 8 Hz. Writing difficulty also increased with increasing duration of vibration. A 10 s exposure to 5 Hz vibration at 2.0 msd2r.m.s. resulted in subjective estimates corresponding to 'extremely difficult'. The effects of vibration magnitude (0.63 to 1.6 msm2r.m.s.), vibration frequency (0.5 to 10 Hz),and vibration duration (2 cycles to 10 s) on the spilling of liquid from a hand-held cup were also investigated in a group of 20 subjects. The probability of spilling the liquid, the quantity of liquid spilt, and subject's estimates of the probability of spillage were determined for all conditions. Greatest interference with the task occurred at 4 Hz, with the lowest vibration magnitude (0.63 ms-* r.m.s.) causing measured and estimated spillage probabilities of approximately 85%. The interference was much less at other frequencies, with 0.63 ms-I r.m.s. causing less than 10% measured probability of spillage,below3 Hz and above 5 Hz. The estimated probability of spillage was generally greater than the observed probability of spillage when the spillage probability was low, but less than the observed probability when the spillage probability was high. Increasing the duration of vibration increased the'probability of spillage, and also increased the volume of liquid spilt. 1. Introduction This paper is concerned with the effects of vertical vibration on two activities, writing and holding a cup of liquid, often undertaken by passengers in railway trains and in other vibration environments. Most of the published investigations of the effects of vibration of performance have concerned tasks performed by military personnel, such as the control of joysticks in aircraft or tanks. In a summary of the research into the effects of vibration on task performance McLeod and Griffin (1986) classified tasks into three categories:
(i) type A tasks, in which the 'subject controls the hand freely in space: examples include reaching and pointing. In some type A tasks the hand may hold an object which will itself be affected by motion, such as fluid in a cup'. (ii) type B tasks, in which the 'subject's hand manipulates a control at a fixed *Now at the Department of Psychology, University of Nottingham, University Park, Nottingham NG7 2RD, UK. 0014-0 13919 1 $3.00 0 1991 Taylor & Francis Ltd.
C.Corbridge and & J. Griffin i . position attached to the vibrating structure: examples include the operation of joysticks and knobs'. (iii) type C tasks, in which the subject 'performs a single, discrete operation, such as changing a switch setting or pressing a button'. This type of task may often be preceded by a type A task, in which the hand moves through space in order to locate the control.
Downloaded by [University of Cambridge] at 23:02 05 November 2015
Both holding a cup of liquid in space and writing on a hand-held surface might be classified as type A tasks. This type of task, together with type C tasks have been the subject of little empirical research. 1.1. Writing Huddleston (1964) found that the size of seated subjects' handwriting on a hand-held surface increased when they were exposed to vibration. Subjects experienced particular difficulty when writing while exposed to 4.8 and 6.7 Hz whole-body vibration at 4.9 r n r 2 r.m.s., with 4.8 Hz causing most difficulty. There was a progressive reduction in script height as the vibration frequency was increased to 9.5 and 16 Hz. Of the 540 words copied in each condition, the numbers illegible with 4.8, 6.7, 9-5, and 16 Hz vibration were 31, 31, 17, and 0 respectively. A second study (Huddleston 1965) confirmed the adverse effects of vibration on writing performance but failed to show any difference between the effect of vibration of 4.8 and 6.7 Hz. Murata (1 973) measured the time required to write a grid reference, part of a map reading and writing task, by 19 military personnel exposed to vertical random vibration. The subjects performed the writing task on a table. The vibration conditions were designed to simulate the motion encountered in armoured fighting vehicles moving over metalled surfaces and cobbles and were presented at a magnitude of 2 m r 2 r.m.s. For exposure durations up to 20 min there was little consistent effect of vibration on writing time. There was evidence from tests conducted on an additional three subjects that exposures up to 1 h may decrease performance. The author suggested that some subjects worked harder during exposure to vibration so as to maintain performance and that this could be sustained for durations of up to 20 min but not up to 1 h. Cambell (1 974) exposed 12 military personnel to sinusoidal vibration at frequencies of 1-5, 3.0, and 6.0 Hz for periods of 3 h at a vibration magnitude of 2.5 , m r 2 r.m.s. The test conditions were intended to be similar to those encountered in armoured fighting vehicles and subjects were required to perform a battery of reading and writing tests, again on a table. Vibration conditions were found to impair the performance of a maze drawing task as assessed by both the length of maze completed and the number of errors. Performance of map reading and writing tasks showed similar adverse effects of vibration: both the time required to complete the task and the number of errors increased. The adverse effects of vibration also increased as vibration frequency increased and 6 Hz vibration was found to be particularly detrimental to writing performance. (In common with findings reported by Huddleston [1964], the author noted that in adverse conditions the writing became larger yet more difficult to read.) A study using similar performance measures was reported by Farmilo and Bennett (1975). They reported that the effect of exposure duration was negligible when compared to the overall decrement due to vibration.
Vertical vibration and passenger activity
1315
Downloaded by [University of Cambridge] at 23:02 05 November 2015
The published studies suggest that considerable difficulty may be experienced when attempting to write while exposed to whole-body vibration in the frequency range 4 to 6 Hz. However, the magnitudes of vibration used in previous studies were severe. There has been no attempt to investigate the effect of vibration magnitude on writing performance, or the effect of vibration conditions typical of those encountered by fare-paying passengers. 1.2. Holding a cup of liquid Adverse effects of vibration are commonly observed when drinking in moving vehicles but there has been little relevant research. Only one previous study, Whitham and Griffin (1978), is known to have investigated the effects of vibration on this activity. In their study, 12 male subjects were exposed to vibration while seated on a rigid flat seat without a backrest. The subjects were exposed to motion of the seat in the three translational axes (fore-and-aft, lateral, vertical) and in the three rotational axes (roll, pitch, and yaw). Sinusoidal vibration at the preferred one-third octave band centre frequencies between 1.6 Hz and 6.3 Hz were used (1.6 to 10 Hz for the vertical axis). All stimuli had a duration of 10 s and, after each had been experienced, subjects were required to indicate if they had spilt the liquid. The findings were expressed as the magnitudes of vibration at which 25% and 50°h of the subjects spilt the liquid, as a function of vibration frequency in each axis. High sensitivity was found to 4 Hz vibration in all axes. In the rotational axes subjects were less affected by yaw vibration than roll or pitch vibration. Subjects were less affected by motion in the horizontal axes than in the vertical axis. Vertical vibration with a magnitude of 0.3 ms-I r.m.s. at 4 Hz resulted in 50% of the subjects spilling the liquid. The second experiment reported here extends the findings reported by Whitham and Griffin (1 978). The range of stimuli was extended to include frequencies down to 0.5 Hz. This was necessary since rail vehicle vibration has significant motion at the low end of the frequency range from 0.5 to 5.0 Hz (see Grifiin 1990). 2. Experiment 1: writing
2.1. Experimenlal design The method of magnitude estimation was used to compare writing difficulty caused by 'test' motions with that caused by a 'reference' motion. The 'reference' motion was an octave band of random motion centred on 1 Hz, magnitude 0.63 m r 2 r.m.s. and duration 20-5 s. The difficulty caused by this condition was assigned a writing difficulty of 100. Each subject was then presented with a series of five 'test' motions. After each 'test' motion the subject was required to assign a number to the motion corresponding to the degree of writing difficulty experienced relative to that of the 'reference' motion. For example, if the degree of writing difficult was twice that of the 'reference' motion, the 'test' motion would have been assigned a value of 200. Conversely, if the 'test' motion produced only half the writing difficulty of the 'reference' motion it would have been assigned a value of 50. Each of the 'test' motions consisted of the same octave bandwidth random motion centred on 1 Hz as the 'reference' motion, but with an added sinusoidal vibration component. The frequency, duration and magnitude of the 40 additional components shown in table 1 were added to the octave bandwidth random motion centred on 1 Hz to produce 40 'test' motions. The only constraint was that the
C. Corbridge and M.J. Griffin
1316
Downloaded by [University of Cambridge] at 23:02 05 November 2015
additional sinusoidal vibration did not occur in the first or the last 5 s period of the background motion. The background random motion, which defined the length of the writing task, was typical of that which occurs in modem high speed rail vehicles travelling over good track. The sinusoidal vibration was to provide information on the magnitudes and durations of vibration at various frequencies which would affect writing. The 40 'test' motions were randomly divided into eight groups of five with a presentation of the 'reference' motion preceeding each group. The 'reference' motion was therefore presented to each subject on eight separate occasions. The sequence of presentation was randomized across subjects. Table 1. Additional sinusoidal vibration components added to the background random motion to produce the 'test' motions. Frequency Hz Duration
0.5
1.0
2.0
3.15
4.0
5.0
6.3
8.0
10.0
Magnitudes ms->r.m.s.
2 cycles 2s
x
x
x
x
x
x x
x
x
x
x
x
x
x
x
x
x x x x
1.6 1.6 1.6 0.63, 1.0, 1.6
4s 10s
x x
In a separate part of the experiment the magnitude estimation technique was used to investigate the effect of the magnitude of random vibration on writing difficulty. Subjects compared the effect of the 'reference' motion with that of five random 'test' motions and a control condition without vibration. The random 'test' motions were centred on 1 Hz, had the same spectrum and duration as the 'reference' motion (20.5 s) but magnitudes of 0-16, 0.3 15, 0.63, 1-25, and 2.5 m c 2 r.m.s. In a further part of the experiment the subjects were asked to rate the writing dificulty produced by the above five octave band random test motions on the following semantic scale: a not difficult; a little difficult; fairly dimcult; a difficult; a very difficult; a extremely difficult.
The use of both a semantic scale and the magnitude estimation technique enabled the magnitude estimates of writing difficulty to be located on an 'absolute' scale. This was then used as an aid to interpreting the magnitude estimates of writing difficulty produced by the 'test' motions containing sinusoidal components. The determinations of the effect of vibration magnitude on writing difficulty using both the magnitude estimation technique and the semantic rating scale were carried out at both the beginning and the end of the experimental session. This was to enable the effect of experience of vibration exposure and practice at both the task and the psychophysical methods employed in the study to be investigated. The form of the full experimental session, which lasted approximately 1 h, is shown in table 2.
J'ertical vibration and passenger activity
1317
Table 2. Structure of the experimental session. Section of experimental session 1
2
Downloaded by [University of Cambridge] at 23:02 05 November 2015
3
4 5
.
Description
Assessment of random vibration magnitude on writing difficulty using the semantic scale. Assessment of random vibration magnitude on writing difficulty using the magnitude estimation ~echnique. Assessment of 'test' motions containing added sinusoidal vibration components on writing difficulty using the magnitude estimation technique. As section 2. As section 1.
2.2. Subjects Ten male and ten female subjects participated in the study. The subjects, who were paid, were students, research workers and technicians at the University of Southampton. Male subjects were aged 22 to 3 1 years (mean 25.8 years), weighed 6 1 to 78 kg (mean 7 1.1 kg), and were 1.68 to 1.88 (mean 1.80 m) tall. Female subjects were aged 20 to 28 years (mean 23-2 years), weighed 47 to 6 1 kg (mean 54-4kg), and were 1.58 m to 1.71 m (mean 1.66 m) tall. Five of the male subjects and one female subject used the left hand as the preferred hand when writing. 2.3. Writing task Subjects were required to copy a row of 39 lower-case letters of the alphabet printed on one line of a sheet of paper onto a line marked out below. The first 26 characters of the row consisted of the letters of the alphabet in random order. The remaining 13 characters were randomly selected from the alphabet with the constraint that a character did not occur more than twice in a row. Each sheet contained six rows of characters and six associated lines of spaces. The writing tasks performed during a 'reference' motion and the five subsequent 'test' motions were therefore contained on a single sheet. Upon completing one 'reference' motion and five 'test' motions subjects removed the sheet from the top of the clipboard and placed it at the bottom. The thickness of paper on which the subjects were writing therefore remained constant throughout the experiment. The subjects were required to copy the characters as quickly and accurately as possible during the presentation of each vibration stimulus, consistent with producing legible handwriting. They were instructed to use cursive script, that is, joining up the letters as in normal handwriting. The time allowed for the task was 20-5 s (the duration of thevibration stimuli); a 2 kHz tone of magnitude 65 dB(A) at the subjects' ears signalled the start and the end of each writing task. 2.4. Equipment A 1 m stroke vertical hydraulic vibrator located at the Southampton University Institute of Sound and Vibration Research was used to generate the vibration stimuli. This device was designed for human response to vibration studies and its typical performance when generating sinusoidal stimuli has been documented
C. Corbridge and M. J. Griffin
Downloaded by [University of Cambridge] at 23:02 05 November 2015
previously, Corbridge and Griffin (1986). Subjects were seated on a rigid flat seat with a backrest inside a wooden cabin mounted on the vibrator table; the interior was illuminated by a single 24 W bulb. Subjects had no external view of the laboratory. The rigid flat seat on which the subjects sat against a rigid flat backrest has been described previously by Corbridge and Griffin (1 986) and has been shown to have unity transmissibility to vibration in the frequency range used in this study.
2.5. Procedure Subjects were screened for medical contra-indications (BSI 1973) and seated inside the cabin. For safety reasons subjects were required to wear a loose-fitting lap strap during the experiment. Subjects were then given instructions relating to the conduct of the writing task, allowed to ask questions and provided with a clipboard, response sheets and a Parker Rollerball pen fitted with a new refill. Subjects then practised the task five times under static conditions. They were then provided with a copy of the semantic scale described in section 2.1 and required to 'consider you are travelling on a train. Estimate the writing dificulty that you experience.' Subjects were verbally instructed to attend specifically to the writing dificulty inherent in the task and not to any discomfort produced by the various motions. Subjects were then exposed to the five magnitudes of octave bandwidth random motion from 0.16 to 2.5 ms-2 r.m.s. and an additional control condition without vibration whilst performing the writing task. The six stimuli were presented to the subjects in random order; following each stimulus presentation the subjects assessed the writing difficulty using the semantic scale. Following completion of section 1 of the experimental session, subjects were given instructions regarding the use of the magnitude estimation technique. Subjects were given the opportunity to practice the technique in the assessment of the length of lines drawn on paper. Upon satisfactory completion of this training, subjects compared the writing difficulty experienced with an octave bandwidth random 'reference' motion, magnitude 0-63m r 2 r.m.s., with that produced by five 'test' motions. These 'test' motions consisted of the same octave bandwidth random motion presented at five vibration magnitudes between 0.1 6 and 2.5 msm2r.rn,s. On completion of section 2 of the session, subjects immediately commenced section 3, the main part of the study. In this section subjects used magnitude. estimation to assess the writing difficulty of the 40 'test' motions which consisted of the octave bandwidth random motion, magnitude 0-63 m r 2 r.m.s., with an added sinusoidal component. The investigation of the effect of random vibration magnitude on writing difficulty was then repeated using the magnitude estimation technique (session section 4) followed by use of the semantic rating scale (session section 5). 2.6. Results Vibration had two main effects on writing: it reduced both the number of characters copied and their legibility. The number of characters copied was selected as the objective measure of writing performance most amenable to consistent scaring. The maximum number of characters which could be copied during each vibration exposure was 39. The maximum score was achieved in fewer than 5% of all exposures. Each individual score for any 'test' motion was converted to a percentage of the score obtained when the subject experienced the preceeding 'reference' motion. This reduced the effect of any systematic learning. (Raw scores indicated
Vertical vibration and passenger activity
1319
Downloaded by [University of Cambridge] at 23:02 05 November 2015
that the median number of characters copied when exposed to the 'reference' motion increased from 31 on the first presentation to 33.5 on the eighth and final presentation-indicating that a slight learning effect occurred over the experimental session.)
2.6.1. Effect of magnitude of random vibration on writing diflculty. Few significant differences were observed between maIe and female subjects in either the subjective ratings or the magnitude estimates. In only one of the 40 'test' conditions did the subjective responses of the male subjects differ significant from those made by female subjects (pt0.05, Mann-Whitney U-test). No differences were obtained between the two groups in objective measures of writing performance. In the subsequent sections the results from the two groups have been combined. No significant differences (Wilcoxon matched-pairs signed rank test, p>0.10) existed between ratings made at the beginning and end of the experimental session. The octave bandwidth random motion centred on 1 Hz with a magnitude of 0-63 ms-* r.m.s. (the 'reference' motion used in section three of the experimental session) resulted in a median rating of 0.5 (i.e. between 'not dificult' and 'a little difficult'). Analysis of the results obtained from the combined group of subjects at the beginning and end of the experimental session showed that in only one condition did the results differ significantly (pt0.02). (Estimates of writing difficulty in response to random motion with a magnitude of 1.25 ms-I r.m.s. made at the beginning of the experimental session were higher than those made at the end.) Figure 1 shows the magnitude estimates of writing dimculty expressed as a function of the rated writing difficulty obtained from the 20 subjects at the beginning and end of the experimental session.
I
2
3
4
5
6
Semantic rating of wrHing dinicuhy
Figure I . Magnitude estimates of writing difficulty as a function of rated difficulty of writing obtained from 20 subjects with octave bandwidth random vibration centred on 1 Hz with ~ Semantic rating: 1 =not difficult; 2=a little vibration magnitudes up to 2.5 r n ~ ' r.m.s. difficult; 3=fairly difficult; 4=difficult; 5 =very dificult; 6 =extremely difficult.
The least squares regression analysis relating magnitude estimates Q to semantic ratings ( x ) isdescribed by the following function: 10g,o= 1.18 10g~fi-k1.72
'
C. Corbridge and M.J. Grifln
1320
Using this equation, magnitude estimates corresponding to the various semantic descriptors of rated writing dificulty were calcuIated as in table 3. Table 3. Semantic descriptors and calculated magnitude estimates of writing difficulty.
Downloaded by [University of Cambridge] at 23:02 05 November 2015
Semantic descriptor of rated writing dificulty (numerical value) Not difficult A little difficult Fairly difficult Difficult Very difficult Extremely difficult
Calculated magnitude estimate of writing difficulty
(1)
(2)
(3) (4) (5) (6)
None of the subjects rated any of the random motions as being a condition in which writing was 'Extremely Difficult.' The extrapolation of the function to calculate the magnitude estimate of writing difficulty corresponding to this descriptor as shown in table 3 is therefore based on the assumption that magnitude estimates of writing difficulty will continue to increase logarithmically with rated difficulty. In the following section, magnitude estimates corresponding to the semantic descriptors 'A Little Difficult' and 'Very Difficult' have been marked on the figures to aid the interpretation of the magnitude estimates. 2.6.2. Effect offeguency of sinusoidal vibration (a) 10 s duration: Figure 2 illustrates the variability in subjective estimates and objective measures of writing difficulty as a function of vibration frequency for sinusoidal components added to the background random motion. The curves show the effect of adding 10 s sinusoidal components with magnitudes of 0.63 and 1.6 msm2r.m.s. to the random motion.
0
1
2
3
4 5 6 7 Frequency (Hz)
8
9
1
0
Frequency (HI)
Figure 2. Median, 25th, and 75th percentile subjective estimates and objective measures of writing difficulty as a function of frequency for sinusoidal vibration components added to background random motion. Sinusoidal components of duration 10 s: -0.63 rns+ r.rn.s.; - 1.6 ms-I r.m.s.
1321
Vertical vibration and passenger activity
Median subjective estimates of writing difficulty when 10 s sinusoidal components were added to the background motion at magnitudes of 0.63, 1.0, and 1-6 ms-I r.m.s. are shown in figure 3. Friedman non-parametric analyses of variance showed that there was a significant effect of vibration frequency on magnitude estimates for each of the three magnitudes of added sinusoidal vibration @0.05). Median estimates of writing difficulty exceeded 1 18 ('a little difficult') at only one frequency: 3.15 Hz. Writing speed also showed no significant effect of vibration frequency when the duration was limited to 2 cycles of motion (p>0- 10). The median writing performance did not fall below 93% of the number of characters copied in the relevant 'reference' conditions.
2.6.3. Eflect of duration of sinusoidal vibration: The effects of increasing the duration of added 5 Hz and 10 Hz sinusoidal components from 2 cycles to 10 s on subjective and objective measures of writing performance are shown in figure 4. There was a significant effect of vibration duration at both 5 and 10 Hz (p
