4.1 Measures obtained by a response-type system depend primarily on the vehicle design and condition, the load, the measuring speed, and a host of environmental conditions. Even with control of all significant variables, the response of every vehicle is unique. Thus, raw measures from such a system are not reproducible with other systems.

4.2 The calibration described in this practice provides a method for converting the raw output of a particular response-type system to a reproducible standard roughness scale.

4.2.1 The response of a vehicle to road roughness is a complex phenomenon that cannot be summarized in a laboratory test. Therefore, the calibration is made through correlation with standard roughness index values established for calibration sites situated on representative roads. The data from the calibration sites are analyzed to determine an equation to estimate the standard roughness index from an RTSN.

4.3 The estimate of the standard roughness index made by transforming an RTSN is subject to three types of error:

4.3.1 Random Error of the Response-Type System (Repeatability)-This error includes operator error and variability in the response of the vehicle and other components of the response-type system. It can be reduced by performing repeated measurements with the response-type system and averaging the individual measurements to estimate the true RTSN for a site. Appendix X1 describes a test method for determining the magnitude of in-use repeatability error.

Note 1: The length of the site or sites used to estimate in-use repeatability shall be equal to the minimum length of the test sections to be surveyed by the response-type system. This may require test sites that are longer than those profiled for the calibration.

4.3.2 Bias Error in the Calibration Equation-Estimates of the standard roughness index are biased if the calibration equation is incorrect or if no calibration equation is used. The purpose of this standard practice is to reduce bias to a negligible level. If desired, the magnitude of bias remaining after calibration can be estimated from data collected in the calibration.

4.3.3 Standard Error of the Estimate (Error Due to Interactions Between Site Effects and Response-Type System Effects)-This error is constant (a bias) for a particular combination of response-type system and site, but it is random with site selection. Ultimately it limits the accuracy of the estimate of the standard roughness of a site made with a response-type system. The error can be estimated from data collected in the calibration.

4.3.3.1 The standard error of the estimate estimates the error due to physical differences in response between a particular response-type system and the standard roughness index. It cannot be reduced by a mathematical transform.

4.3.3.2 Three physical variables that are controllable and that influence the standard error of the estimate are vehicle test speed, shock absorber damping stiffness, and vehicle tire pressure. For most vehicles, maximum reproducibility of standard roughness index estimates is obtained by adopting a test speed of 80 km/h [50 mph], by equipping the vehicle with stiff shock absorbers, and by maintaining a standard tire pressure. (See also 8.2.)

4.4 Periodic verification is essential to ensure that the calibration remains valid.