Validation of speed perception and production in a single screen simulator

Objective: validate the simulator for speed. Method: A study was conducted in which 16 24-30 years old drivers drove on an actual 40 km desert route between two cities, and in a simulator in which the road geometry was replicated exactly and the sparse scenery was approximately reproduced. Actual speed ranged from 40 to 100 km/h. Half of the drivers drove in the simulator first, and half drove on the real road first. In each drive the drivers could not see the speedometer and had two tasks: speed production, in which they were told to achieve a predetermined speed, and speed estimation, in which they were told to accelerate or decelerate to a certain point, and then asked to estimate the speed. Results: The comparisons between the real driving and the simulated driving yielded very high correlations. The correlation between real speed and produced speed on the road was nearly 1.0 (r=0.997), and in the simulator it was 0.86. The correlation between real speed and estimated speed on the road was also nearly 1.0 (r=0.997), and in the simulator it was almost the same (r=0.98). In both the simulator and the road, the estimated speed was slightly lower than the true speed, while the produced speed was slightly higher than the requested speed. Conclusion: all the results were consistent in showing that the simulator quite accurately reproduces speed sensations as a function of changes in the simulator speed, and transfer functions to adjust simulator speed to reproduce speed sensations in actual driving are quite easy to achieve.