Academic journal article Human Factors

Time-to-Collision Estimation in a Simulated Driving Task

Academic journal article Human Factors

Time-to-Collision Estimation in a Simulated Driving Task

Article excerpt

INTRODUCTION

When operating a vehicle the driver is continually faced with the problem of avoiding collisions with other vehicles, pedestrians, and various obstacles that may lie in the path of travel. To prevent potential collisions, the driver may initiate steering or braking actions or some combination of both. Appropriate regulation of the timing and control of such actions requires the driver to anticipate the time of the impending collision. This time remaining before the collision, often termed time to collision or time to contact ([T.sub.c]), is critical information for the driver in enabling prospective control of braking or steering behavior.

Often a driver's ability to judge the time of a collision is made more difficult by the fact that he or she may not have continuous visual information from the obstacle. This may be the case when the driver has to allocate his or her visual attention to some other aspect of the environment or when the obstacle is momentarily hidden from view by a corner of the road or the brow of a hill. In these situations the driver must judge the time remaining before collision from visual information obtained before the obstacle was occluded.

One way a driver might determine this [T.sub.c] information could be based on the ratio s/v, in which v represents the visual perception of closing velocity and s represents the relative position of the obstacle from the driver. In this putative method the observer might compute [T.sub.c] from independently perceived distance and velocity, and so this method is often referred to as a computational or cognitive approach. In attempts to examine the viability of this approach, researchers have asked observers to verbally report estimates of a vehicle's velocity and distance. [T.sub.c] is then calculated from these estimates (e.g., Schiff, Oldak, & Shah, 1992; Scialfa, Guzy, Leibowitz, Garvey, & Tyrell, 1991; Scialfa, Kline, Lyman, & Kosnik, 1987).

Findings from such experiments have consistently shown substantial underestimation of closing velocity and an even greater percentage of underestimation of distances. Subsequent [T.sub.c] calculation from these estimates suggests that observers would greatly underestimate [T.sub.c] if they were using this computational method. Such findings cast doubt on the functional relevance of this putative computation. Alternatively, underestimates of [T.sub.c] may arise because observers simply do not have verbal access to perceptual information concerning velocity and distance that is used in [T.sub.c] computation.

In contrast to the suggested cognitive mechanism, the ecological optics approach (Gibson, 1966, 1979; Lee, 1976) proposes that computation of [T.sub.c] from low-order information is unnecessary because [T.sub.c] information is directly available in the optic flow fields created by the approaching object. Lee (1976, 1980) showed mathematically that the retinal image of an approaching object dilates and that under constant closing velocity, the inverse of this rate of dilation, designated tau ([Tau]), directly specifies [T.sub.c]. In studies in which computer and film simulations of approaching objects have been displayed, observers were able to accurately predict [T.sub.c] even though no distance cues were available (Schiff & Detwiler, 1979; Todd, 1981).

Furthermore, examination of a wide variety of motor skills has espoused the role of [Tau] in the visual regulation of action. Investigations of ball punching (Lee, Young, Reddish, Lough, & Clayton, 1983, Sardinha & Bootsma, 1993), ball catching (Savelsbergh, Whiting, & Bootsma 1991), table tennis (Bootsma & van Wieringen, 1990), and landings from jumps (Sidaway, McNitt-Gray, & Davis, 1989) have found considerable empirical evidence for the use of [Tau] in determining [T.sub.c].

The possible role of [Tau] has also been examined in [T.sub.c] estimation in vehicle driving skills (Cavallo & Laurent, 1988; McLeod & Ross, 1983). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.