This content is not included in your SAE MOBILUS subscription, or you are not logged in.
A Method for Data Alignment
ISSN: 0148-7191, e-ISSN: 2688-3627
Published November 16, 1998 by SAE International in United States
Annotation ability available
Engineering in the Motorsports arena of ten involves the detailed comparison of two laps of data. Plots of data acquisition channels as a function of distance allow analysis of driver and vehicle performance at the same point on the track. Distance is computed from the Start/Finish line or wherever a timing beacon is placed.
The distance function is usually defined as the integral of speed, which is calculated as the sum of discrete speed values. Speed is usually measured as the RPM of a wheel, usually a non-powered wheel, such as a front wheel for a rear-drive race car.
The calculated distance function is subject to errors that can harm the analysis. For the four-mile Road America track this means counting the revolutions of a wheel from the timing beacon to a location up to 6,400 meters away. A skilled driver considers a variation of 3 meters a large change in braking point. This represents rolling a 27.5-inch diameter tire 2,900 times and defining a point within one revolution. This demands an accuracy of 1 part in 2,900, or 0.03%. A ten-bit data acquisition system only records data, such as speed, to one part in 1,024 or 0.10%. An eight-bit system has a resolution of one part in 256 or 0.39%.
There are additional problems with this type of measurement. The measuring device is a rubber tire that can change effective rolling radius as a result of aerodynamic download or other factors. The tire also generates significant braking forces. The dynamics of braking distort the relationship between the rpm of the wheel and true vehicle speed. The percent slip of a racing tire may be in the 6% range. This further distorts the calculation of the distance function.
An additional problem is brake lockup where the wheel may literally stop while the vehicle continues as the tire slides across the pavement. This results in a wheel rpm near zero while the true vehicle speed is much higher. This is detrimental to the tire, usually causing flat-spots, and to the accuracy of the measured vehicle speed. This introduces error to the calculated distance function. Timing beacons are also a source of error. The point where beacons are marked in the data may vary ten meters.
A better way of aligning data is to match the patterns of the road profile. By aligning the pattern of suspension travel or shock velocity the data can be aligned with confidence. The techniques used here with 50-100 hertz data resulted in an alignment within one meter. This drastically improves the confidence in the calculated distance function and permits more accurate analysis. This permits comparison of braking points, turn-in points, and even shock velocities over the same bump. It also improves the time gain/loss function that is often used to compare two laps.
This paper will present a mathematical method for matching road profile data and properly aligning data.