The loss of tire braking friction due to road roughness was simulated and measured experimentally in a test machine designed to produce simultaneous wheel slip and vertical vibration of the contact surface. Equivalent roughness amplitudes as great as 0.7 in. And roughness frequencies (velocity/wavelength) as high as 14 hz were covered. Wheel slip was varied by discrete amounts to include the values normally associated with maximum friction. The dynamic friction force found in each test run was time-averaged and expressed as a percentage of the highest observed average to minimize the effects of temperature, surface conditions, and other secondary considerations. Results show simulated roughness amplitude and frequency to have a strong influence on the average force available for braking: friction losses were 30 percent at 0.04 in. And 14 hz, and 90 percent at 0.71 in. And 6 hz. Wheel slip and the dead-weight load on the tires were found to have a less dramatic effect over the range tested. The most important conclusion reached is that friction predictions without road profile consideration can result in gross errors and may be one of the causes of lack of correlation of friction data.
Abstract