汽車外文翻譯---駕駛者的轉(zhuǎn)向感

1、<p>　　Driver perception of steering feel</p><p>　　Abstract: Steering feel is optimized at a late stage of vehicle development, using prototype vehicles and expert opinion. An understanding of human perc

2、eption may assist the development of a good 'feel' earlier in the design process. Three psychophysical experiments have been conducted to advance understanding of factors contributing to the feel of steering syst

3、ems. The first experiment, which investigated the frames of reference for describing the feel (tic properties) of a steering wheel, indi</p><p>　　INTRODUCTION</p><p>　　Driving a car is a complex

4、 task and involves many interactions between the driver and the vehicle through the various controls. Good performance of the system depends on how well a car 1 s able to create the driver's tensions, and how well di

5、ffer- ences between those in-u-ns and the vehicle's response can be detected the driver. The steering system is one of the primary controls in a car, allowing the driver to control the direction of the vehicle. The s

6、teering system not only allows the driver </p><p>　　Forces originating at the road tyre interface (and related to the road wheel angle, vehicle red, and road adhesion), present themselves at the steering whe

7、el (subject to kinematic losses through the steer-ing system, and subject to various assist methods in steering systems, e.g. hydraulic and electric power assist) where the driver can interact with them and develop an in

8、ternal model of the steering propertiesand the environment</p><p>　　The relationship between the steering-wheel torque and the steering-wheel angle has been considered a useful means of describing steering f

9、eel Various metrics' of the relationship are used to define steering feel, and experiments have found that changing the relation between the steering-wheel force and steering-wheel angle can alter the driving experie

10、nce. Knowledge of the way in which haptic stimuli at the steering wheel are perceived by drivers may therefore assist the development of steerin</p><p>　　The perception of stiffness and the perception of vi

11、scosity seem to come from force, position, and velocity cues. Psychophysiological studies indicate that muscle spindle receptors, cutaneous mechano-receptors, and joint receptors provide the neural.inputs used in the per

12、ception of the movement and force applied by a limb </p><p>　　Psychophysics provides techniques to describe how subjects perceive stimuli. Classic measures include the difference threshold (the minimum chang

13、e needed to detect a change in a stimulus) and the psychophystcal function (the relationship between changes in stimulus magnitude and the perception of those changes). However, the first step in quanti-fying steering fe

14、el using psychophysical methods is to identify what aspects of the haptic feedback at the steering wheel are used by drivers.</p><p>　　Steering torque and steering angle describe the steady state characteris

15、tics of steering systems and their relationships have been identified as influencing steering feel. It seems appropriate to check whether subjects are judging what the experimenter 15 measuring. It has not been shown whe

16、ther the properties of steering system should be described in rotational frames of reference or translation frames of reference </p><p>　　This paper describes three experiments designed to study how drivers

17、perceive the steady state properties of steering wheels. The first experiment investigated whether rotational or translation frames of reference are more intuitive to subjects. It was hypothesized that, if asked to '

18、match' different steering-wheel sizes, either the rotational or the translation frame of reference would be matched more consistently. The second experiment deter- mined difference thresholds for the perception of s&

19、lt;/p><p>　　2 APPARATUS</p><p>　　A rig was built to simulate the driving position of a 2002 model year Jaguar S-type saloon car as shown. The framework provided a heel point for subjects and suppor

20、ted a car seat and steering column assembly. The cross-section of a Jaguar S-type steering wheel was used to create the grips of the experimental steering wheel, which was formed by a rapid prototype polymer finished wit

21、h production quality leather glued and stitched on to the subject posture was constrained the seat steering wheel, a</p><p>　　The steering-column assembly included an optical incremental encoder to measure a

22、ngle (resolution, 0.0440), a strain gauge torque transducer to measure torque (0.01 N accuracy), bearings to allow the wheel to rotate freely (isotonic control), and a clamp to lock the column in position (isometric cont

23、rol).</p><p>　　3 EXPERIMENTS</p><p>　　Three experiments were performed to investigate the response of the driver to steady state steering-wheel properties and to determine, firstly, the driver f

24、rame of reference, secondly; the difference thresholds for the perception of force and angle, and, thirdly, the</p><p>　　rate of growth of sensations of force and angle.</p><p>　　The experiments

25、 were approved the Human Experimentation, Safety and Ethics Committee of the Institute of Sound and Vibration Research at the University of Southampton</p><p>　　3.1 Driver's frame of reference</p>

26、<p>　　Frames of reference provide means for representing the locations and motions of entities in space. There are two principal classifications for reference frames in spatial perception: the allocentric (a framew

27、ork external to the person), and the egocentric (a frame- work centred on the person). For some tasks, the choice of reference frame may be merely a matter of convenience. In human spatial cognition and navigation the re

28、ference frame determines human perception. The haptic perception of stee</p><p>　　In engineering terms, it is convenient to describe the motion of a steering wheel in a rotational frame of reference using st

29、eering-wheel torque and steering- wheel angle. However, drivers may use a different frame of reference when perceiving the feel of a steering system; they may perceive steering-wheel force rather than steering-wheel torq

30、ue, and steering-wheel displacement rather than steering-wheel angle</p><p>　　Alternatively, drivers may use neither allocentric nor egocentric frames of reference and instead may employ some intermediate re

31、ference frame as suggested by Kappers </p><p>　　This experiment aims to test whether drivers sense steering-wheel force or torque, and whether they sense angle or displacement. The relationships between thes

32、e properties are</p><p><b>　　T=rF (1)</b></p><p><b>　　x=rB (2)</b></p><p>　　To investigate which variable is intuitively used by drivers, it is necessary to

33、uncouple the relationship between rotational and translation frames of referece. This can be achieved by altering the radius of the steering wheel. It was hypothesized that, when asked to 'match' a reference cond

34、ition using isometric steering wheels (i.e. wheels chat do not rotate)with varying radii, subjects would match either the force applied by the hand or the torque applied to the steering wheel. It was similar</p>&

35、lt;p>　　Subjects attended two sessions, one with isometric steering wheels and one with isotonic steering wheels. Four reference conditions were presented in each session: 5 N, 15 N, 1.5 N m, and 3 N m with the isometr

36、ic steering wheels, and 3`, 90, 10 mm, and 30 mm with the isotonic steering wheels. The forces and distances refer to the forces and dist-es at the</p><p>　　hm of the steering wheel.</p><p>　　Fo

37、r this experiment, 12 male subjects, aged between 18 and 26 years, took part using a within-subjects experimental design where all subjects participated in all conditions. The order of presentation of the reference condi

38、tions was balanced across subjects</p><p>　　For six subjects, the first session used the isometric steering wheels; for the other six subjects, the first session used the isotonic steering wheels</p>

39、<p>　　For each reference condition, a total of 18 trials were undertaken: nine trials to account for each combination of three reference wheels and three diameters of test wheel (small, medium, and large) including

40、matching to the same wheel, and a repeat of these nine conditions.</p><p>　　The length of time that subjects were required to hold a force or torque was minimized to prevent fatigue. Typically, subjects took

41、 10s to reach the desired force or angle. The view of their hands was obscured so that subjects did not receive visual feedback of their position or movement.</p><p>　　3.1.1 Method</p><p>　　Usin

42、g the 'method of adjustment' [11l, subjects 'matched' sensations from a 'reference' steering wheel to a 'test' steering wheel. When grasping the reference wheel, subjects were required to

43、achieve a desired stimulus magnitude師acting on the wheel in a clockwise direction using visual feedback from a fixed 11-point indicator scale on a computer monitor. Instructions on the computer monitor then instructed th

44、e subjects to move their hands to either the'small','medium; or'large steering wheel, an</p><p>　　Correlation coefficients between the physical magnitudes of the reference condition and the t

45、est condition are presented for each subject in Table 1 For isometric control, correlation coefficients were obtained for both torque and force at the steering- wheel rim. For isotonic control, correlati

46、on coefficients were obtained for both angle and displace ment at the steering-wheel rim. It was assumed that the variable with the greater correlation (i.e. either force or torque, or an</p><p>　　Over the 1

47、2 subjects, for isometric control, the correlation coefficients obtained for force were significantly higher than those obtained for torque (p<0.01, Wilcoxon matched-pairs signed-ranks test) Por isotonic control, the

48、correlation coefficients obtained for angle were significantly higher than those obtained for displacement (p<0.01)</p><p>　　3.1.3 Discussion</p><p>　　Lines of best fit to the data had gradie

49、nts of less than unity for 11 subjects. The single subject that achieved a slope greater than 1.0 did so only for angle data. The effect could have arisen from the reference being presented first (i.e. an order effect) A

50、lternatively, it could indicate that the physical variables do not reflect the parameters adjusted by is described in terms of a'Weber fraction' or percentage. Weber proposed that the absolute difference threshol

51、d is a linear function of st</p><p>　　The results suggest that with idealized isometric and isotonic controls, drivers have a better sense of steering-wheel force than steering-wheel torque and a better sens

52、e of steering wheel-angle than steeringwheel displacement. It seems that subjects used the forces in their muscles and the angles at the joints of their hands and arms to position the steeringwheels.</p><p>

53、　　To judge torque, subjects would need to combine estimates of force with knowledge of the distance between their hands and the centre of the steering wheel. To judge the displacement of the steeringwheel rrn， subjects w

54、ould need to combine estimates of their joint angles with the length of their limbs. The estimation of torque and distance requires more information and greater processing than the estimation offorce and angle. Cons

55、equently,it is not surprising that torque and distance result </p><p>　　3.2 Difference thresholds</p><p>　　A difference threshold is the smallest change in a stimulus required to produce a just

56、noticeable difference in seusation 111]. Differencecuts,山resholds can be described in absolute terms, where the threshold is described in the physical units of the variable under test, or in relative terms, where the thr

57、eshold often expressed as a percentage</p><p>　　Difference thresholds for the perception of force are available in a variety of forms. Jones 1121 reported (7 per cent) for forces generated at the elbow flexo

58、r muscles. Difference thresholds for lifted weights have been reported by Laming .based on an Experiment Fechner using weights from 300 to 3000 g, resulting in a Weber fraction of 0.059 (5.9 per cent), and Oberlin [151 m

59、easured difference thresholds for lifted weights from 50 to 550 gaving a Weber fraction of 0.043 (4.3 percent)</p><p>　　tic discrimination of finger span with widths varying from 17.7 to 100 mm have been rep

60、orted as 0.021 (2.1 per cent) by Gaydos h61. Discrimination of elbow movement has been reported as 8 per cent by Jones et al. [17], while discrimination of sinusoidal movements of the finger studied Rinker et al. produce

61、d difference thresholds that ranged from the present experiment investigated difference thresholds for steady state steering-wheel force(using an isometric steering wheel), and difference thres</p><p>　　3.2.

62、1 Method</p><p>　　Difference thresholds were determined with a two alternative forced-choice procedure using an up-and-down transformed response (UDTR) method. Subjects were required to act on the steering w

63、heel to achieve a reference force or reference angle, followed by a test stimulus. The required levels for both actions were presented on a character-less 11-point scale on a computer monitor. The reference stimulus and

64、a test stimulus were presented sequentially, and in random order, to subjects who were requ</p><p>　　For this experiment, 12 male subjects, aged between 18 and 28 years, took part using a within-subjects exp

65、erimental design. The order of presentation for the refer-- conditions was balan-d across sub- jects with six subjects starting with isotonic control, and six starting with isometric control.</p><p>　　3.2.2

66、Results</p><p>　　The median absolute and relative difference thresh olds are shown in Table 2. For both force and angle, the absolute difference thresholds increased signifi- cantly with increasing magnitude

67、 of the reference(p < 0.01, Friedman test)</p><p>　　The median absolute and relative difference thresh- olds for both force and angle are shown in Fig. 6 andFig. 7 respectively. The median relative differ

68、ence thresholds tended to decrease (from 16.5 per cent to 11.5 per cent) with increases in the reference force and decrease (from 17.0 per cent to 11.5 per cent)with increases in the reference angle. However, over-all,

69、 the relative difference thresholdsd not differ significantly over the three force references or over the three angle references</p><p>　　Driver pereepfiun of.waring feel produce. The bias causes magnitude p

70、roduction to</p><p>　　yield steeper slopes (i.e. higher values for n) than magnitude estimation.</p><p>　　The third experiment employed both magnitude estimation and magnitude production to deve

71、lop a force</p><p>　　The mean relative difference thresholds across the magnitudes of the reference stimuli were 15 percent when detecting changes in force and 14 per cent when detecting changes in angle. Th

72、is suggests no fundamental difference in the accuracy of detecting changes in force and angle, implying that force and angle provide equally discriminable changes in</p><p><b>　　feedback.</b><

73、/p><p>　　For the perception of force, the 15 per cent relative difference threshold was obtained with a correct performance level of 79.4 per cent. Direct comparison with the aforementioned studies of the perce

74、ption of force are not possible, as correct response levels are not presented in those studies. For the perception of angle, 14 per cent in the present study compares the a difference threshold for limb movement in the r

75、ange 10-18 per cent (for a 71 per cent correct performance level) according to </p><p>　　3.3 Rate of growth of sensation</p><p>　　The rate of growth of sensation of stimuli has often been determ

76、ined using Stevens' power law Where is the sensation magnitude,is the stimulus intensity, k is a scalar constant depending on the conditions, and n is the value of the exponent that describes the rate of growth of se

77、nsation of the stimulus and depends on the sensory modality (e.g perception of force, or perception of loudness).</p><p>　　Previous studies have reported rates of growth of sensation of force and weight with

78、 exponents between。and 2.0 over a variety of experimental conditions 121-2引A study of the haptic sensation of finger span Steven. and Stone using widths of 2.3-63.7 mm reported an exponent of 1.33 using magnitude estima

79、tion.</p><p>　　The value of the exponent n may be determined by either magnitude estimation or magnitude production. Magnitude estimation requires subjects to make numerical estimations of the perceived magn

80、itudes of sensations, whereas magnitude production requires subjects to adjust the stimulus to produce sensory magnitudes equivalent to given numbers These methods have systematic biases which Stevens called a 'regre

81、ssion effect. The biasese attributed to a tendency for subjects to limit the range of stimul</p><p>　　A subject then applied 11 different test forces (or angles) by applying a force or angle until the pointe

82、r was placed at the middle mark of the 11-point scale. The forces or angles required corresponded to 50 per cent, 60 per cent, 70 per cent, 80 per cent, 90 per cent, 100 per cent, 120 percent, 140 per cent, 160 per cent,

83、 180 per cent, and 200 per cent of the reference force or angle. For force, these stimuli ranged from 5.25 N to 21 N while, for angle, they ranged from 4.5` to IS`. After the </p><p>　　For this experiment, 1

84、2 male subjects, aged between 18 and 26 years, took part using a within- subjects experimental design. Subjects attended two sessions with the order of presentation of the force angle, and magnitude estimation, and magni

85、tude production conditions balanced across subjects.</p><p>　　The exponent indicating the rate of growth of sensation was determined by fitting Stevens' power law to the data. With the stimulus and sensa

86、tion</p><p>　　3.3.2 Results</p><p>　　Exponents for the rate of growth of sensation were obtained from least-squares regression between the median judgements of the 12 subjects for each test magn

87、itude and the actual test magnitude, with the apparent magnitude assumed to be the dependent variable (261. The calculated exponents were 1.14 (force magnitude estimation), 1.70 (force magnitude production), (1.91 (angle

88、 magnitude estimation), and 0.96 (angle magnitude production).</p><p>　　The median data, and lines of best fit from all subjects, are shown in Figs 8, 9, 10, and 11 for force estimation, force production, an

89、gle estimation, and angle production respectively and are compared inFig. 12</p><p>　　The Spearman rank order correlation coefficients between the physical magnitudes and the perceived magnitudes were 0.89 f

90、or force magnitude estimation, 0.65 for force magnitude production, 0.89 for angle magnitude estimation, and 0.87 for angle magnitude production. All correlations were significant (p < 0.01; N=132), indicating high co

91、rrelations between stimuli and the estimated or assigned magnitude.</p><p>　　3.3.3iscussfon</p><p>　　With magnitude estimation, the rank order of all median estimates of force and angle increase

92、d with of growth production apparent force using magnirude estimatml of apparent force using magni</p><p>　　Data from 12 subjects Incresing force and angle, except for the middle (100 and 120) force estimate

93、s. This deviation is assumed to have arisen by chance. To assess the impact that this deviation has on the exponent obtained from the median data, an exponent was regressed to all data points from all subjects. This yiel

94、ded an exponent of 1.14, which is the same as the exponent determined from the median data. Similarly, with magnitude production, the median forces and angles</p><p>　　increased with increasing required valu

95、e, except for the two lowest forces. The lowest median force was produced when subjects were asked to produce a rate of growth of sensation of apparent force and angle is taken as the geometric mean of the rates of growt

96、h for magnitude estimation and magnitude production. In this study, the means of the esti- mation and production slopes were 1.39 for steeringwheel force and 0.93 for steering-wheel angle.</p><p>　　The rate

97、of growth of sensation of steering wheel force ties within the range previously reported for force. A rate of growth of 1.39 means the sensation of force growmore rapidly than the force causing the sensation. For example

98、, a doubling of force will give rise to a 162 per cent increase in the perception of force. Steering-wheel angle had a mean rate of growth of 0.93; so the sensation of angle grows at a slower rate than the angle. For exa