Printer Friendly

The retention of balance: an exploratory study into the limits of acceleration the human body can withstand without losing equilibrium.


Everyone who regularly travels by public transportation has seen someone lose his or her equilibrium because of acceleration of the vehicle. In the Netherlands, passengers have been swung against the door and have fallen out of the bus in a sharp turn, resulting in serious injuries or even death. A recent poll (Consumentengids, 1995, pp. 128-131) shows that complaints regarding excessive braking and acceleration are common, especially in buses and trams. The records of the Consumer Safety Institute show that each year 2300 passengers need first aid treatment in a hospital because of accidents occurring in the Dutch public transportation systems (Mulder, 1993). In accidents in which people lost their balance (claimed in 1227 cases), the direct cause is not unequivocally clear. However, it seems reasonable to look for a clue in the nature of the accelerations or decelerations to which the victims were exposed.

Worldwide, some research has been done on the effect of acceleration on postural balance, but this research is of a fragmentary nature and has frequently focused on compensatory responses to very small disturbances (Allum, 1983; Dietz, 1986; Nashner, Woollacott, & Tuma, 1979). Guedry (1974) reported on measurements of the (lower) perceptual threshold for accelerations, and human reactions to extreme stresses are known from military aviation research. However, there is a paucity of data in the international scientific literature on experiences with accelerations in everyday life and their role in causing people to lose their postural balance completely.

In the 1940s, however, Jongkees and Groen (1942) investigated a series of sudden accelerations causing people to lose their balance. They used a small vehicle to accelerate 50 standing participants in a forward, backward, and side-ward direction. They found that healthy individuals who stand upright with closed eyes and feet together are able to endure an acceleration of up to 76 cm/[s.sup.2] in a backward direction, a forward acceleration of up to 48 cm/[s.sup.2], and a sideward acceleration of up to 33 cm/[s.sup.2]. (An acceleration in a backward direction is produced by a sudden backward movement of the floor surface - in their case a small vehicle - relative to the body.)

The stimulus Jongkees and Groen used in their experiments was a sudden, constant acceleration. It is arguable that higher limiting values might be obtainable if the level of acceleration were increased gradually rather than in steps: This would give the participant's balance reflexes some time to adapt to the disturbance.

In practice, different figures from those Jongkees and Groen (1942) reported are found. The standards for acceleration and deceleration in European road traffic define acceleration levels of 100 to 150 cm/[s.sup.2] in a longitudinal direction as manageable and decelerations of 150 cm/[s.sup.2] as comfortable. The most "sensitive" road users (i.e., standing bus passengers; Westerduin, 1974) were taken as the criterion in choosing these values. If we evaluate these data in light of Jongkees and Groen's findings, we can see that these bus passengers are at minimum assumed to be capable of holding on tightly. Accelerations in a transverse or vertical direction should not exceed a value of 50 or 25 cm/[s.sup.2], respectively (Westerduin, 1974). No justification is provided for the European road traffic standards figures, and we do not know of any fundamental research on this topic.

The chief aim of the present study is to explore the most extreme human limiting values for linear accelerations - that is, the values obtained under the most favorable sensory conditions. The scope of the study was restricted to stationary people (either standing still or moving steadily in relation to the surface of the earth, such as in the bus) exposed to a sudden (constant or gradual) acceleration or deceleration. The data obtained in the laboratory were compared with situations occurring during travel on public transport.



A treadmill with a conveyor belt (ENRAF NONIUS, Delft, Netherlands) was used to expose 22 standing people in succession to backward, sideward, and forward accelerations of the floor surface. The group consisted of 11 men and 11 women in normal health, ranging in age from 26 to 63 years, who were randomly selected from the population of the TNO Human Factors Research Institute. Because of the exploratory character of the study, it did not appear useful to consider a sample more refined and extensive than 22 normal, healthy people chosen by chance from the population of one of our institutes.

The total distance moved was 80 cm, 10 cm of which was taken up by the initial movement and 25 cm by the braking distance. (In order to ensure that the increase in the speed of the treadmill had a genuinely linear character - constant acceleration - the treadmill was first brought to a low constant speed to overcome the initial frictional resistance. Participants had to correct their postural balance briefly at this point, but they were fully stabilized again before the real acceleration occurred. The time window of this initial period differed slightly for each stimulus exposure. This way the participant could not build up an expectancy pattern about the moment of stimulus acceleration.) This therefore left 45 cm over which the participants were exposed to a single constant acceleration. The treadmill was controlled by a computer. The level of acceleration for each of the stimulus profiles was regularly calibrated with an accelerometer (triaxial acceleration transducer AS-2TG, KYOWA Electronic Instruments, Tokyo, Japan). Figure 1 illustrates the type of stimulus profile used.

The acceleration was varied in steps of 0.1 m/[s.sup.2] over a range from 0.3 to 1.6 m/[s.sup.2]. The sequence in which the acceleration levels were presented was randomly determined, but we sometimes omitted acceleration levels that would clearly fall outside the coping capacity of the individual participant. For instance, someone who was already having severe difficulties with an acceleration of 0.5 m/[s.sup.2] would not be exposed to an acceleration of 1.6 m/[s.sup.2].

Participants were told that they were standing in the bus with a bus driver who could be bad-tempered and were instructed to keep their eyes open, their hands free, their shoes on, the heels of their feet together, and their toes about 3 to 4 cm apart. Each participant was then exposed to accelerations of the floor surface, twice in a backward direction (-Gx), then twice in a sideward direction (+Gy), and then twice in a forward direction (+Gx). The stimulus profile used was the same throughout, but the participant was turned in the desired direction. This meant that the participant knew the direction of the next acceleration but did not know when it would occur.

We investigated whether the participants were capable of coping with the acceleration without showing a drastic loss of postural balance (large sway movements with their body and arms), being forced to take one or more steps, having to hold on, or even needing to be caught in order to avoid falling. We assumed that all the participants felt sufficiently safe in the experimental situation because of the support railing and the proximity of the leader of the experiment (to catch them, if necessary). The whole range of accelerations was presented quickly and in a relaxed atmosphere. The threshold values at which the participants were just able to retain their balance over two exposures without problems were noted.

The complete series of measurements of the limiting values for accelerations in a forward, sideward, and backward direction took 10 min. The complete procedure (a repeated-measures 3 x 14 design, with acceleration levels randomly mixed) was repeated one week later with a randomly chosen subgroup of 12 participants (of the 22) in order to identify possible learning effects. This same subgroup was asked to get back on the treadmill yet again a week after the one-week follow-up; on this occasion, however, they were allowed to choose their favorite pose for standing in the bus (whatever they preferred, but without holding on). The rest of the procedure remained the same.

In order to explore a possible relationship between postural behavior on the treadmill and during normal upright stance, we briefly examined the participants on a stabilimeter platform. The stabilimeter, on which the participant stands, is a stable horizontal platform that records the excursions of the body. The measure of stability used in this study was the RMS value over a period of 50 s. (For details of this method, see Bles and De Jong, 1986.)


Despite clear differences in approach between Jongkees and Groen's (1942) experimental design and our design (moving vehicle vs. treadmill; closed eyes vs. open eyes; barefoot vs. wearing shoes; threshold measurements made during the deceleration phase vs. the acceleration phase), the measurement data obtained in the two studies are very similar (see Table 1). (Although the two threshold measurement methods are equivalent in physical terms, they could create differences in physiological - for instance, visual - terms).

Group Means (and Standard Deviations, in m/[s.sup.2]) for the Three
Directions of Acceleration

N = 22                    Forward    Sideward    Backward    Sum

Jongkees & Green            0.48       0.33        0.76      1.57
De Graaf & Van Weperen      0.54       0.45        0.61      1.60
                           (0.16)     (0.12)      (0.16)

The differences between the directions of acceleration are significant for our sample: post hoc Newman-Keuls on analysis of variance (ANOVA) p [less than] .01. The participants were found to be most vulnerable to sideward accelerations and least vulnerable to backward accelerations.

For the purpose of further comparison, we determined a persistence index (a kind of global indication of one's resistance to acceleration, which also identifies someone's position relative to a peer group) for each of the participants, in accordance with the method used by Jongkees and Groen (1942). This was the sum of the highest acceleration levels recorded for each participant in the three directions, divided by the sum of the group's mean values, or I = ([a.sub.f] + [a.sub.s] + [a.sub.b])/(mean [a.sub.f] + mean [a.sub.s] + mean [a.sub.b]).

The persistence index (I) in our experimental group (from 0.44 to 1.5) shows a somewhat greater spread than in Jongkees and Groen's group (0.7 to 1.3). We were unable to detect any differences in persistence between the male and female participants, but there was a (negative) correlation between persistence and age. Within the age range of our participants (26-63), the persistence index value decreased significantly (p [less than] .01) as the participant's age increased [ILLUSTRATION FOR FIGURE 2 OMITTED]. (A recalculation with omission of the two individuals with the highest and lowest persistence indexes still yields a correlation coefficient r = -.41 and a probability level p = .07.)

A comparison of the data for the 12 (of 22) participants who repeated the complete procedure one week later suggests a slight learning effect: ANOVA: F(1, 11) = 27.1, p [less than] .01, 4.5% variance explained. This is shown in the first two rows of Table 2. Because the effect applies to the whole group (a structural effect), on average the persistence index remains the same (given that it is a relative measure). In other words, on average, someone's position relative to the group remains the same because the performance of everyone in that group changes in the same direction and by the same amount.

When the third test was carried out, during which the same 12 participants were tested in their own choice of pose for standing in the bus (feet further apart and not completely in line), the strategy adopted appeared to be particularly functional for sideward accelerations. The limiting values for accelerations in a forward and backward direction could not (on average) be improved any further (see Table 2, Test 3).

We found no correlation for our sample between the data from the posturographical examination and the persistence index. All the RMS values recorded by the stabilimeter for our group fell well within the norms for normal postural balance behavior (Roos, 1991) and, indeed, differed little from one another. Consequently, postural control during normal upright stance with the eyes open is not demanding enough of the equilibrium system to allow prediction of postural stability on the treadmill.


Despite the differences in approach, our data replicate those obtained by Jongkees and Groen (1942). It is true that the limiting values for linear accelerations in the three directions lie somewhat closer, but the sum for the group corresponds very closely to the value found in 1942 (Table 1). Bearing in mind this correspondence, and the fact that the experimental procedures differed substantially, this means that the human limiting values for linear accelerations are constant in nature.

The persistence indexes in our group had a somewhat greater degree of spread than did Jongkees and Groen's (1942), but this may be attributable to the spread of ages in our sample: Our group of participants included people who were either younger or somewhat older than were those in Jongkees and Groen's study. The older people had lower limiting values than the younger ones, and this could have had an effect on the persistence index.

Subgroup Means (and Standard Deviations, in m/[s.sup.2]) for the
Three Directions of Acceleration

N = 12       Forward       Sideward       Backward       Sum

Test 1         0.56          0.47           0.64         1.67
              (0.18)        (0.14)         (0.18)

Test 2         0.65          0.53           0.75         1.93
              (0.17)        (0.17)         (0.21)

Test 3         0.60          0.93           0.68         2.24
              (0.19)        (0.19)         (0.24)

The limiting values can be increased further with the benefit of some experience, but only marginally (Table 2, Tests 1 and 2). However, if the participants stand with their legs well apart, the limiting value for a sideward acceleration can be increased substantially (Table 2, Test 3).

We found no correlation between the stabilimeter data obtained from the posturographical examination and the limiting values found for applied accelerations. Such a relationship, if it exists at all, is probably too weak to be detected in the limited size of our sample.


In order to gain an impression of the situation in practice, we measured the acceleration levels that occur during travel by public transport in Amsterdam on a random weekday using accelerometers (g meters: AS-2TG, KYOWA Electronic Instruments). The drivers of the vehicles were not informed about these measurements.


A randomly selected sample [ILLUSTRATION FOR FIGURE 3 OMITTED] of the figures recorded on the tram, express tram, bus, and metro services reveals immediately that the amplitude of the acceleration levels on all these modes of transport is high enough to ensure that none of the people we tested would have been able to maintain their postural balance without support. The initial acceleration in a longitudinal direction regularly lies between 1 and 2 m/[s.sup.2], which is certainly higher than the level at which people can cope in an optimal situation, but without support, without being in danger of losing their balance. This applies primarily to the bus; not only is the acceleration measured in a longitudinal direction the highest (2.15 m/[s.sup.2]), but the associated lateral acceleration levels occurring on bends and when swerving into and out of stops are substantial (up to about 4 m/[s.sup.2]). For details, see de Graaf (1993).


The accelerations that are commonly encountered in practice appear to be impossible to endure without support. Handgrips will increase the coping ability of standing passengers in a longitudinal direction to 1.50 m/[s.sup.2] (Browning, 1974), but this does not eliminate all problems.

The initial impetus ("jerk") with which the acceleration begins is also relevant to postural balance. It is this jerk that can throw passengers off balance in the first place. The issue is seldom broached, but Vuchic (1981) recommended that vehicles for public transport should be designed in such a way that their acceleration rate does not change more quickly than 0.50 to 0.60 m/[s.sup.3]. At the moment, this recommendation does not appear to have been adopted. The initial acceleration levels found on public transport in our data set appeared to vary from 1.5 m/[s.sup.3] in practice in a longitudinal direction to 3.5 m/[s.sup.3] in a transverse direction. The jerk that occurred on the treadmill varied between 1.0 and 7.0 m/[s.sup.3]. The jerk on the treadmill was undoubtedly sometimes somewhat higher but continued for only 0.2 s, whereas the jerk measured on public transport could continue for several seconds.


It appeared worthwhile to carry out a second experiment in the laboratory focusing specifically on the jerk with which acceleration begins (the change in acceleration per unit of time). The question we investigated in this experiment was whether there was any benefit to be gained - that is, would the human limiting value for acceleration levels increase if the acceleration of the vehicle began somewhat less abruptly?


Ten new participants (five men and five women, haphazardly chosen from the population of the Human Factors Research Institute) were exposed on the treadmill to a standard forward acceleration (1.00 m/[s.sup.2]) that was higher than the mean limiting value of the former participants. This meant that, as a rule, the participant would have to take one or more corrective steps at each stimulus in order to avoid falling. The participant was now repeatedly exposed to this acceleration value but with differences in the onset of the stimulus (the jerk component; [ILLUSTRATION FOR FIGURE 4 OMITTED]). This produced the four conditions specified in Table 3.

Each participant was presented with the four conditions in a fixed sequence (balanced among participants so that the sequence could vary for each participant), and four measurements were taken per condition. The participant's score was determined by the number of times that he or she coped with a stimulus without problems (maximum score: 4 x 4 = 16). The score per condition is thus the sum of the individual scores in that condition.

Stimulus Conditions in the Second Laboratory Experiment

Jerk (m/[s.sup.3])                      (m/[s.sup.2])

1.0 ([+ or -]0.2)                            1.0
2.0 ([+ or -]0.2)                            1.0
5.0 ([+ or -]0.2)                            1.0
10.0 ([+ or -]0.2)                           1.0


The results, summarized in Table 4, indicate that some benefit can be gained from the human limiting values for accelerations in cases in which the onset of the acceleration is sufficiently smooth. It is clear that the recommendation formulated by Vuchic (1981) to limit the onset of accelerations to 0.50-0.60 m/[s.sup.3] must be taken seriously by anyone wanting to prevent accidents involving falls on public transport. Unfortunately, the stimulus equipment available in the laboratory did not permit an acceleration rate of 1 m/[s.sup.2] to be paired with a jerk level of less than 1 m/[s.sup.3], and hence the value proposed by Vuchic could not be tested. However, bearing in mind that 35% of the participants were unable to continue standing without difficulty under the mildest stimulus condition, one can assume that the optimal jerk level must be well below 1 m/[s.sup.3].

Scores of 10 Participants in Response to a Stimulus with a Constant
Acceleration But Variable Onset Levels

                                Jerk (m/[s.sup.3])

Participant         1.0       2.0         5.0         10. 0

1                   4         4           0           0
2                   2         2           0           0
3                   0         0           0           0
4                   0         0           0           0
5                   0         0           0           0
6                   4         4           0           0
7                   4         3           0           0
8                   4         0           0           0
9                   4         4           4           0
10                  4         2           1           1

of balance          65%       47.5%       12.5%       2.5%


The combined approach, using tests of participants in the laboratory and measurements of accelerations occurring in practice on public transport, appears to provide meaningful data for the issue of accidents involving falls.

Laboratory research reveals that the human limiting values for applied acceleration levels are constant (replication of data obtained in the 1940s) and that there is a negative correlation with age. A comparison between the human limiting value for acceleration levels and the acceleration levels measured on public transport services reveals that standing passengers will never be able to retain their balance during the journey without support. However, restricting or changing the onset of acceleration to 0.5-0.6 m/[s.sup.3] should in principle enable passengers to cope with the acceleration levels in a longitudinal direction that now occur on the metro, tram, and bus. (Any such restriction should never be allowed to affect the vehicle's ability to brake immediately and sharply.) In the case of the bus, sideward accelerations must also be taken into account.

These data, however, can be generalized only to apply to people standing "passively" without support. In the case of people behaving actively, the acceptable limiting values can sometimes turn out to be higher (anticipation), but they are frequently also lower (e.g., one has less supporting surface when walking toward a seat). Because the aim is to ensure that increasing numbers of people will make use of public transport, future research should investigate this question, together with the human limiting values for (combinations of) applied acceleration levels with the use of support.


Part of these data were presented at the Third International Conference on Product Safety Research, Amsterdam, March 6-7, 1995.


Allum, J. H. J. (1983). Organisation of stabilizing reflex responses in tibialis anterior muscles following ankle flexion perturbations of standing man. Brain Research, 264, 297301.

Bles, W., & De Jong, J. M. B. V. (1986). Uni- and bilateral loss of vestibular function. In W. Bles & T. Brandt (Eds.), Disorders of posture and gait (pp. 127-140). Amsterdam: Elsevier.

Browning, A. C. (1974). The tolerance of the general public to a speed differential between adjacent floors with special reference to pedestrian conveyors: Exploratory and preliminary experiments (Tech. Report 74076). Farnborough, England: Royal Aircraft Establishment.

Consumentengids. (1995, February). Brokken in het openbaar vervoer [Accidents during public transportation]. The Hague, Netherlands: Consumentenbond.

Dietz, V. (1986). Afferent and efferent control of posture and gait. In W. Bles & T. Brandt (Eds.), Disorders of posture and gait (pp. 69-81). Amsterdam: Elsevier.

Graaf, B., de. (1993). Versnelling en evenwicht; experimenteel onderzoek naar de grenswaarden voor versnellingen die het menseliik lichaam kan ondergaan zonder het houdingsevenwicht te verliezen [Acceleration and balance; an investigation into the limits of acceleration which the human body can withstand without losing equilibrium]. Amsterdam: Consumer Safety Institute.

Guedry, F. E. (1974). Psychophysics of vestibular sensation. In H. H. Kornhuber (Ed.), Handbook of sensory physiology (pp. 3-154). Berlin: Springer.

Jongkees, L. B. W., & Groen, J. J. (1942). De standvastigheid van de mens [The stability of the human body]. Nederlands Tiidschrift voor Geneeskunde, 86, 1401-1407.

Mulder, S. (Ed.). (1993). Jaaroverzicht prive ongevallen registratie systeem [Annual report home and leisure accident surveillance system]. Amsterdam: Consumer Safety Institute.

Nashner, L. M., Woollacott, M., & Tuma, G. (1979). Organisation of rapid responses to postural and locomotor-like perturbations of standing man. Experimental Brain Research, 36, 463-476.

Roos, J. P. W. (1991). Posturography in the tilting room. Unpublished Ph.D. thesis, Free University, Amsterdam.

Vuchic, V. (1981). Urban public transportation systems and technology. Englewood Cliffs, NJ: Prentice-Hall.

Westerduin, B. (1994). Richtlijnen voor het ontwerpen van wegen buiten de bebouwde kom [Guidelines for the design of roads outside built-up city centers]. Janssen, Utrecht, Netherlands.

Bernd de Graaf received his Ph.D in biology at the University of Utrecht in 1990. He is senior scientist in the Equilibrium & Orientation Research Group at TNO Human Factors Research Institute.

Willem Van Weperen received his M.S. in physics from the University of Groningen and is head of the Product Safety Department at the Consumer Safety Institute, Amsterdam.
COPYRIGHT 1997 Human Factors and Ergonomics Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1997 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Graaf, Bernd de; Weperen, Willem van
Publication:Human Factors
Date:Mar 1, 1997
Previous Article:Performance during positive pressure breathing after rapid decompression up to 72,000 feet.
Next Article:Practical operation of a biaxial goniometer at the wrist joint.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters