Prospective control of lane changing and tau-dot

Many manual and locomotor tasks require the actor to reach a certain position in space, while reducing movement velocity close to zero. The 'tau- dot' variable, the first temporal derivative of the current time-to-contact, was suggested to play a critical role in the prospective control of deceleration (e.g. Lee, 1976; Yilmaz & Warren, 1995). Lane changing can be considered as a biphasic task, where in a first phase a lateral velocity is produced (pull-out), which then has to be minimized in a second phase (approach to the envisaged lane). This second alignment phase necessarily requires prospective control of the lateral deceleration. Here we asked, whether lane changing might be guided by tau-dot, resulting in a lateral approach to the new lane with tau-dot values kept constant at or regulated around -0.5. In a driving simulation using a large cylindrical projection screen (7m diameter), 46 participants were asked to drive in a 3.5 m wide straight lane and to perform consecutive lane changes to an adjacent lane. The deceleration of 1460 trajectories of the second phase were analyzed, computing fit and slope of a linear regression of time-to-contact with respect to time. We found that two-thirds of the lane changes can be described by a single tau-dot value (criterion r > 0.6, mean r = 0.94). These trials yield a mean tau-dot of -0.499 (sd = 0.132), which is in suprising agreement with the theoretical value of -0.5. Futhermore, we observed also a positive correlation (r = 0.68) between the time required for the completion of the lane change and the tau-dot values. Our results complement previous studies showing evidence for the use of the tau-dot variable, even when the point of termination of the approach was largely self-defined, as it was true for our task. Drivers prefered a path with a constant lateral deceleration, despite the non-linear relationship between lateral deceleration and steering-wheel amplitude.