Fig. 4. Lateral tire force as a function of slip angle at various values of FY. Fig. 5. Cornering stiffness as a function of pF. Fig. 6. Model of normalized longitudinal force versus slip ratio for friction estimation algorithm. Fig. 7. Test vehicle- Volvo XC90 with GPS system. Moreover, a low-pass filter is also being inherently used in (9) to reduce the noise from the model-based value. This method also provides a minimal phase lag due to all signal frequencies being captured. Fig. 9. Bank angle estimation using two accelerometers. Fig. 10. Lateral acceleration measured on banked road. Fig. 11. Vertical acceleration measured on banked road. Fig. 14. Bank angle calculation using lateral and vertical accelerometers anc algorithm 2. Fig. 13. Bank angle calculation using lateral and vertical accelerometers and algorithm 1. Fig. 12. Real-time variation of the quantity 4,7 + 4,7 — g? ESTIMATION ERRORS FOR EXPERIMENTAL TESTS. table, the rms error values are of the order of 0.15—0.2° on dry road and of the order of 0.8° on icy road (when the slip angle i noted that the vehicle is just completing a cornering maneuver at this time and returning to a straight and flat road. The roll angle induced during this transient is likely the cause of the dip/ spike in the data. The transient roll angle affects both measurements A, and A,. Since these measurements have opposite signs in (13) and (14), one of the algorithms has a transient dip while the other has a transient spike. a Dr. Piyabongkarn is a member of SAE and ASME. He was a recipient of the 2007 O. Hugo Schuck Award for the best paper at the 2006 American Control Conference, the SAE 2006 Arch Colwell Merit Award, and the 2003-2004 Doc- toral Dissertation Fellowship from the University of Minnesota.
