A study of the human ability to detect road surface type on the basis of steering wheel vibration feedback

Abstract

A study was performed to investigate the human ability to detect road surface type on the basis of the associated steering wheel vibration feedback. Tangential direction acceleration time histories measured during road testing of a single mid-sized European automobile were used as the basis for the study. Scaled and frequency-filtered copies of two base stimuli were presented to test subjects in a laboratory setting during two experiments that each involved 25 participants. Theory of signal detection (TSD) was adopted as the analytical framework, and the results were summarized by means of the detectability index d∞ and as receiver operating curve (ROC) points. The results of the experiment to investigate the effect of scaling suggested monotonic relationships between stimulus level and detection for both road surfaces. Detection of the tarmac surface improved with reductions in acceleration level, while the opposite was true of the cobblestone surface. The ROC points for both surfaces were characterized by gradual increases in detectability as a function of acceleration level, obtaining hit rates of nearly 100 per cent at optimum. The results of the experiment to investigate the effect of frequency bandwidth suggested a monotonically increasing relationship between detectability and the bandwidth of the vibration stimuli. Detection of both road surfaces improved with increases in bandwidth. Average hit rates exceeded 80 per cent for stimuli covering the frequency range from 0 to 80 Hz. Human detection of road surface type appears to depend on the long-term memory model, or cognitive interpretation mechanism, associated with each surface. The complexity of the measured response suggests the need to categorize and classify incoming data before an optimal choice of feedback stimuli can be made in automotive steering systems.

Facilitating the driver detection of road surface type by selective manipulation of the steering-wheel acceleration signal

“…In order to facilitate the comparison of the current results for road surface type detection to findings from other research fields, the data was normalised by means of the detectability index d' which is commonly used in signal detection applications [10,20]. For experimental protocols such as the current one in which the test subjects are requested to provide a simple "yes" or "no" response, the detectability index d' can be estimated from the experimentally determined hit rates and false alarm rates by means of the resulting normalised deviate values (the Z scores) using the relations provided below.…”

“…Within the context of this general situation a small number of studies [3][4][10][11] have been performed to date in order to begin to understand how the characteristics of automotive steering wheel vibration might influence the human detection task.…”

“…Giacomin and Woo [10][11] investigated driver detection of road surface type by measuring the sensitivity of the human detection to changes in the amplitude and frequency bandwidth of the steering wheel acceleration stimuli. Steering wheel acceleration time histories from automobiles which had been driven over selected road surfaces were manipulated in post processing so as to produce a set of test stimuli which differed from the original signals only in terms of the amplitude or the frequency bandwidth.…”

Monitoring Of Hand-Arm Vibration

“…Similarly, Shibata et al 24 developed a single axis vibration test system for the measurement of a biodynamic response of a human hand-arm system. Giacomin and Woo 25 also used a single axis accelerometer to study HAV in the zaxis of the steering wheel. The z-axis is perpendicular 2 or in tangential direction to the x-y plane, and is positive in the direction towards the steering column.…”

“…As only a single axis of vibration could be reproduced experimentally in the present study, the tangential direction was selected as the most representative and useful in light of possible future developments in automotive steering technology. 25 Table 2 shows a summary of various operating / tests conditions that have been performed by different investigators on various types of vehicles. This shows that there are many factors that determine the dominant axis for measuring and monitoring HAV of the hands and arms of the driver.…”