Seite - 148 - in Maximum Tire-Road Friction Coefficient Estimation

E Longitudinal velocity estimation as parked cars, sewer covers or traffic signs) have a better reflection capability than the road surface, for example. In addition, the road conditions influence the received signal power. In particular, the presence of water is limiting, since it diffuses the signal, [Win09b]. Nevertheless, for many driving states, a sensorfusion of radar-based velocity estimation with on-board sensors seems to be sensible for this application. E.4. GNSS-based estimation of the longitudinal velocity Global navigation satellite system (GNSS) is the umbrella term for existing satellite- basednavigation infrastructures, suchas theglobalpositioningsystem(GPS)orGalileo, [WHW09b, p.677]. Starting in October 2015, it will be mandatory for new vehicles in the European Union to be able to automatically communicate with the European emer- gency call infrastructure system (eCall) after a severe accident (e.g. in case the driver is unconscious and unable to call an emergency service), [Dat14]. The data received by the emergency services include the time of the accident, the position of the vehicle, and its direction, which can be relevant on highways or in tunnels. The technical considerations for the vehicle’s on-board system to be approved include the use of GNSS, [EotEC14]. Thus, it can be assumed that GNSS will be available in series production vehicles in the near future. Belvy and Gerdes used GNSS velocity information to calculate the longitudinal slips sx,i, slip angles αi and the vehicle’s side slip angle β, [BG00]. To calculate the wheel slip, the GNSS velocity was directly used in combination with the wheel speed sensors. During free rolling of the tires, they propose using the GNSS velocity to calibrate the wheel speed sensors’ signals. The results showed high slip noise and an offset that could have been subject to miscalibration. Nevertheless, it was also shown that the noise was brought in by the wheel speed sensors rather than the GNSS velocity measurement. A similar approach was used by Rajamani et al., [RPPL12], where sx is calculated directly using the GNSS velocity and the wheel speeds. Miller et al. usedtheGNSSvelocity toestimate theeffective tire radiusandthe longi- tudinal stiffness of the tires, [MYM+01]. Beyond calculating sx directly using the GNSS velocity, it has been used to calculate the effective tire radius re,i of non-driven wheels during free rolling within a sub-millimeter accuracy. Then, sxwas calculated using only the wheel speed measurements, which excludes estimation of braking conditions (other than engine braking) and for 4WD vehicles. 148