Patent Application
20060058933
The present invention relates to rollover detection in motor vehicles, and more particularly to a method of detecting an impending rollover event due to a soil trip condition.
Various rollover detection methodologies have been developed for activating electrically deployed rollover safety devices such as air bags, side curtains, seat belt pretensioners and pop-up roll bars, and/or for activating visual, auditory or haptic warnings. Typical parameters used to detect rollover include the vehicle attitude rate of change or angular roll rate, the vehicle roll angle, the vehicle speed, the steering wheel angle, the vehicle yaw rate and the side-slip angle. For example, the U.S. Pat. No. 6,542,792 to Schubert et al., issued on Apr. 1, 2003, discloses a rollover detection algorithm that involves determining an angular roll rate vs. roll angle operating point of the vehicle, and comparing the determined operating point to one or more calibrated thresholds.
While algorithms such as the one disclosed in the aforementioned U.S. Pat. No. 6,542,792 to Schubert et al. can timely detect impending rollover events that occur relatively quickly with high roll rates, soil trip rollover events that occur when a vehicle slides sideways into soft or yielding roadside material such as soil, sand or gravel can be very difficult to detect in a timely fashion. Accordingly, what is needed is a method of reliably and timely detecting soil trip rollover events.
The present invention is directed to an improved method of detecting an impending and slowly occurring soil trip rollover event due to lateral sliding of a vehicle on soft or yielding roadside material. An impending soil trip rollover event is detected by recognizing a plow phase increase in lateral acceleration, coupled with a significant trip phase roll rate. The plow phase increase is recognized by modeling a typical soil trip lateral acceleration characteristic and computing a cross-correlation between the measured lateral acceleration and the modeled acceleration. The correlation is compared to a threshold that varies with the measured roll rate to reliably discriminate rollover events from near-rollover events while enabling timely deployment of suitable occupant restraints.
Referring to
As indicated above, the present invention is particularly directed to a method carried by out by the MCU 12 for detecting a class of rollover events generally referred to as soil trip events. These events involve sideways sliding of a vehicle into soft, compliant or yielding roadside materials, such as soil, sand or gravel. In soil trip events, the tires of the vehicle plow into the roadside material, carving furrows and building up a pile of material that eventually trips rollover of the vehicle.
In general, the method of the present invention detects an impending soil trip rollover event by recognizing a plow phase increase in lateral acceleration, coupled with a significant roll rate. The plow phase increase is recognized by modeling a typical soil trip lateral acceleration characteristic and computing a correlation between the measured lateral acceleration and the modeled acceleration. The correlation is compared to a threshold that varies with the measured roll rate to reliably discriminate rollover events from near-rollover events while enabling timely deployment of suitable occupant restraints.
Referring to
The lateral acceleration profile stored in buffer 36 is compared with a reference profile generated by block 40. The reference profile, referred to herein as Ay_SOIL, is representative of a lateral acceleration characteristic typically observed during soil trip rollover events. The profile may be defined over a predefined time window as graphically depicted in
The block 42 cross-correlates the buffered Ay signal with Ay_SOIL, while the block 44 cross-correlates the buffered Ay signal with a version of Ay_SOIL inverted by inverter block 46. The inverter 46 simply reverses the sign of each coefficient in the AySoil profile to provide paths for checking both polarities of Ay signals, depending on whether the vehicle slides to the left or to the right in a soil trip event. The cross-correlations implemented by blocks 42 and 44 preferably comprise a digital convolution technique, such as described for example in “Digital Signal Processing in VLSI”, Richard J. Higgins, Prentice-Hall 1990, incorporated by reference herein. In each case, the cross-correlation generates a correlation value which indicates the closeness of match between the buffered Ay signal and Ay_SOIL (or its inverse in the case of block 44). In the plow phase of soil trip event, the respective correlation output will gradually increase as the buffered Ay signal slides across Ay_SOIL, and then gradually decrease as the input signal progresses beyond the observation window of Ay_SOIL. The peak correlation in the positive path is captured by block 48, and the peak correlation in the negative path is captured by block 50.
The block 52 is responsive to the measured roll rate RR and generates a roll rate dependent correlation threshold CORR_THR on line 54 for comparison with the peak correlation values captured by blocks 48 and 50. As depicted in
Referring back to
In summary, the method of the present invention provides reliable and timely detection of a soil trip rollover event by discriminating both plow phase and roll phase portions of the event. While the invention has been described with respect to the illustrated embodiment, it is recognized that numerous modifications and variations in addition to those mentioned herein will occur to those skilled in the art. For example, matching of the correlation threshold with the maximum correlation value may be achieved by scaling the peak correlation values instead of the threshold, the correlation threshold may be non-linear, and so on. Accordingly, it is intended that the invention not be limited to the disclosed embodiment, but that it have the full scope permitted by the language of the following claims.
