Vehicle safety system

Abstract

A vehicle safety system includes an actuatable restraint device for helping to protect an occupant of a vehicle, and a controller for controlling actuation of the actuatable restraint device in response to a vehicle rollover event. The controller is configured to execute a discrimination algorithm that includes at least one classification metric that discriminates between at least one of a ramp rollover event and a ground rollover event from a bank rollover event using at least one of a vehicle pitch rate (P_RATE) and a vehicle roll acceleration (D_RATE). The discrimination algorithm determines a classification of the vehicle rollover event as one of the ramp rollover event, the ground rollover event, and the bank rollover event. The controller is further configured to select an application threshold for application of the actuatable restraint device. The application threshold corresponds to the classification of the vehicle rollover event.

Description

Technical Field

[0001] This invention relates to a vehicle safety system, and more particularly to a method and apparatus for controlling a vehicle's actuable occupant protection devices, and specifically to an enhanced discrimination method and apparatus for classifying several specific types of vehicle collision events. Background Technology

[0002] Vehicle safety systems include a central control unit (“CCU”) that uses sensors, both local and remote, to detect collision events involving the vehicle and determine whether these events warrant activation of actuated restraint devices, such as airbags and seatbelt retractors. Sensors used by the CCU may include accelerometers and other sensors, such as impact sensors, seatbelt buckle switches, seat pressure switches, steering angle sensors, etc. Using data from these sensors, the CCU can determine that a vehicle collision event has occurred and can execute a discrimination algorithm to classify the collision event into a specific type. The CCU can then actuate the actuated restraint devices based on the specific type of collision event.

[0003] For vehicle safety systems, it is desirable to distinguish between various collision events that a vehicle may be involved in. “Distinguishing” a collision event may mean classifying it into a specific type and differentiating it from other types. If a vehicle safety system can distinguish or identify a collision event as a specific type, it can actuate actuable restraints in a manner appropriate to that specific type of collision event. As used herein, “collision event” can be used to encompass a variety of events involving a vehicle. For example, a collision event might be a collision or impact in which the vehicle collides, strikes, or otherwise engages with a structure of a different type. These collision events could be collisions with deformable obstacles (such as another vehicle) or collisions with non-deformable obstacles (such as trees or utility poles). As another example, a collision event might also involve events such as rollover events, where a vehicle impact results from the vehicle rolling over. A rollover event might result from the vehicle skidding and striking a curb, sliding or otherwise moving away from the curb and falling onto an embankment, or sliding or otherwise moving away from the curb and climbing a slope (such as a hillside).

[0004] Vehicle safety systems can be configured or adapted to distinguish between events where the application of actuated restraints is expected (“applied events”) and events where the application of actuated restraints is not expected (“unapplied events”). Collision determination requires identifying the event type, such as deformable obstacles, non-deformable obstacles, frontal impact, rear impact, side impact, oblique impact, offset impact, rollover, etc. Collision determination also requires determining the severity of the collision and implementing safety functions, which manifest as checks or authorizations to ensure the safe application of actuated restraints.

[0005] Based on the above, it will be understood that it may be desirable to control the actuation and timing of actuable restraint devices in a safety system in response to the type and / or severity of a collision event involving the vehicle. To determine which occupant protection devices to actuate in response to a sensed collision event, the safety system may implement a collision assessment process to differentiate between types of collision events. If the identified collision event reaches or exceeds a severity threshold, and the safety functions agree, the actuable restraint devices can be actuated in a manner appropriate to the determined event type.

[0006] Over the years, safety standards have been revised and updated to "push the envelope" in automotive safety. As a result, automakers have been required to continuously improve the safety of their products to keep up with these standards. As standards become more stringent, safety systems adapt and become more complex and robust. Through the development of vehicle safety systems, it has been found that crash classification is one of the key aspects that helps determine system effectiveness. If a safety system can accurately and robustly identify crash scenarios defined by safety standards, then that safety system can be tailored to provide the best possible outcome for the occupants involved in the accidents that the standard was designed to address.

[0007] Although vehicle safety systems capable of differentiating between various collision events have been developed, further classification and differentiation between collision events remains necessary so that vehicle safety systems can take appropriate action. Collision events that may be expected to be differentiated include different types of side impact events, such as rollover events or events that could lead to rollover.

[0008] Side impact events are those that may warrant the activation of safety equipment such as side airbags (curtain airbags, chest airbags) and / or seatbelt pretensioners. Side impact events can occur in a variety of scenarios. For example, a vehicle may lose control and skid off the road onto adjacent grass / soil, fall onto an embankment, or climb a slope or hillside. As another example, a vehicle may lose control and skid onto a lower obstacle (such as a curb). In any of these scenarios, the severity of the resulting side impact event may warrant the activation of one or more vehicle safety devices. Summary of the Invention

[0009] A vehicle safety system includes an actuable restraint device for helping to protect vehicle occupants, and a controller for controlling the actuation of the actuable restraint device in response to a vehicle rollover event. The controller is configured to execute a discrimination algorithm including at least one classification metric, which uses at least one of vehicle pitch rate (P_RATE) and vehicle roll acceleration (D_RATE) to distinguish between at least one of a slope rollover event and a land rollover event and an embankment rollover event. The discrimination algorithm classifies the vehicle rollover event as one of a slope rollover event, a land rollover event, and an embankment rollover event. The controller is also configured to select an application threshold for applying the actuable restraint device. The application threshold corresponds to the classification of the vehicle rollover event.

[0010] According to one aspect, the at least one classification metric may include a classification metric that evaluates the vehicle roll acceleration (D_RATE) against the vehicle roll angle (R_ANGLE).

[0011] According to another aspect, alone or in combination with any other aspect, the at least one classification metric may include a classification metric evaluating vehicle roll acceleration (D_RATE) against vehicle roll angle (R_ANGLE). The controller may be configured to perform this classification metric evaluating vehicle roll acceleration (D_RATE) against vehicle roll angle (R_ANGLE) to distinguish between slope rollover events and embankment rollover events.

[0012] According to another aspect, alone or in combination with any other aspect, the at least one classification metric may include a classification metric that evaluates the vehicle pitch rate (P_RATE) against the vehicle roll angle (R_ANGLE). The controller may be configured to perform this classification metric, evaluating the vehicle pitch rate (P_RATE) against the vehicle roll angle (R_ANGLE), to distinguish between slope rollover events and embankment rollover events.

[0013] According to another aspect, alone or in combination with any other aspect, the at least one classification metric may further include evaluating the classification metric of the vehicle's lateral acceleration moving average (CCU_1Y_AMA) against the vehicle's roll angle (R_ANGLE). The controller may be configured to perform this evaluation of the vehicle's lateral acceleration moving average (CCU_1Y_AMA) against the vehicle's roll angle (R_ANGLE) to verify the distinction between slope rollover events and embankment rollover events.

[0014] According to another aspect, alone or in combination with any other aspect, the at least one classification metric may include a classification metric evaluating the vehicle's vertical acceleration moving average (CCU_6Z_AMA) against the vehicle's roll angle (R_ANGLE). The controller may be configured to perform this classification metric evaluating the vehicle's vertical acceleration moving average (CCU_6Z_AMA) against the vehicle's roll angle (R_ANGLE) to verify the distinction between slope rollover events and embankment rollover events.

[0015] According to another aspect, alone or in combination with any other aspect, the at least one classification metric may include a classification metric evaluating vehicle roll acceleration (D_RATE) against vehicle roll angle (R_ANGLE). The controller may be configured to perform this classification metric evaluating vehicle roll acceleration (D_RATE) against vehicle roll angle (R_ANGLE) to distinguish between land rollover events and embankment rollover events.

[0016] According to another aspect, alone or in combination with any other aspect, the discrimination algorithm may include a classification metric evaluating vehicle roll acceleration (D_RATE) against vehicle roll rate (R_RATE_2). The controller may be configured to perform this classification metric evaluating vehicle roll acceleration (D_RATE) against vehicle roll rate (R_RATE_2) to distinguish between land rollover events and embankment rollover events.

[0017] Depending on another aspect, alone or in combination with any other aspect, the controller can be configured to perform the evaluation of vehicle roll acceleration (D_RATE) against vehicle roll rate (R_RATE_2) as a classification metric to distinguish between hard soil rollover events, moderate soil rollover events, and soft soil rollover events.

[0018] According to another aspect, alone or in combination with any other aspect, the at least one classification metric may include a classification metric that evaluates the moving average of vehicle lateral acceleration (CCU_1Y_AMA) against the vehicle roll angle (R_ANGLE). The controller may be configured to perform this classification metric, evaluating the moving average of vehicle lateral acceleration (CCU_1Y_AMA) against the vehicle roll angle (R_ANGLE), to distinguish between land rollover events and embankment rollover events.

[0019] According to another aspect, alone or in combination with any other aspect, the at least one classification metric may include a classification metric evaluating the vehicle's vertical acceleration moving average (CCU_6Z_AMA) against the vehicle's roll angle (R_ANGLE). The controller may be configured to perform this classification metric evaluating the vehicle's vertical acceleration moving average (CCU_6Z_AMA) against the vehicle's roll angle (R_ANGLE) to verify the distinction between land rollover events and embankment rollover events.

[0020] According to another aspect, alone or in combination with any other aspect, the controller can be configured to perform an application threshold metric for determining whether to actuate the actuable restraint device. This application threshold metric evaluates the vehicle roll acceleration (D_RATE) against the vehicle roll angle (R_ANGLE), and wherein the controller is configured to apply the actuable restraint device in response to these application threshold metrics exceeding the application threshold.

[0021] According to another aspect, alone or in combination with any other aspect, the system may further include an accelerometer for sensing vehicle lateral acceleration and providing a signal indicating the sensed vehicle lateral acceleration (CCU_1Y). The system may also include an accelerometer for sensing vehicle vertical acceleration and providing a signal indicating the sensed vehicle vertical acceleration (CCU_6Z). The system may further include a roll rate sensor for sensing a vehicle roll rate value and providing a signal indicating the sensed vehicle roll rate value (CCU_4R).

[0022] According to another aspect, alone or in combination with any other aspect, the controller is configured to perform vehicle measurement calculations to:

[0023] • Determine the moving average value of vehicle lateral acceleration (CCU_1Y_AMA) based on the signal of vehicle lateral acceleration (CCU_1Y) sensed by this indication.

[0024] • Determine the moving average value of vehicle vertical acceleration (CCU_6Z_AMA) based on the signal of vehicle vertical acceleration (CCU_6Z) sensed by this indicator.

[0025] • The vehicle roll acceleration (D_RATE) is determined based on the signal of the vehicle roll rate value (CCU_4R) sensed by this indicator.

[0026] • The vehicle roll angle (R_ANGLE) is determined based on the signal of the vehicle roll rate value (CCU_4R) sensed by this indicator.

[0027] According to another aspect, alone or in combination with any other aspect, the system may also include a pitch rate sensor for sensing a vehicle pitch rate value and providing a signal indicating the sensed vehicle pitch rate (CCU_5P). The controller may be configured to perform vehicle metric calculations to determine the vehicle pitch rate (P_RATE) based on the signal indicating the sensed vehicle pitch rate (CCU_5P).

[0028] According to another aspect, alone or in combination with any other aspect, these actuable restraint devices may include at least one of a seatbelt anchor pretensioner, a seatbelt retractor pretensioner, a curtain airbag, a chest airbag, a side airbag, an emergency notification, a door unlocking command, and a high-voltage powertrain disconnect command.

[0029] According to another aspect, alone or in combination with any other aspect, the controller can be configured to issue at least one of an emergency notification, a door unlocking command, and a high-voltage power transmission disconnection command in response to actuation of these actuable restraints. Attached Figure Description

[0030] The foregoing and other features and advantages of the invention will become apparent to those skilled in the art after considering the following description of the invention and the accompanying drawings, in which:

[0031] Figure 1 It is a block diagram illustrating the vehicle's safety system.

[0032] Figure 2 It is a block diagram illustrating the measurement calculations implemented in a vehicle safety system.

[0033] Figure 3 This is a graph showing the application threshold metrics used to determine the occurrence of vehicle roll for different surface conditions.

[0034] Figure 4 This is a schematic diagram illustrating the slope detection algorithm implemented by the vehicle safety system.

[0035] Figure 5 This is a schematic block diagram illustrating the embankment discrimination algorithm implemented by the vehicle safety system.

[0036] Figure 6 This is a schematic diagram illustrating the hard ground discrimination algorithm implemented by the vehicle safety system.

[0037] Figure 7 This is a schematic diagram illustrating a moderate land discrimination algorithm implemented by a vehicle safety system.

[0038] Figure 8 This is a schematic diagram illustrating the soft soil discrimination algorithm implemented by the vehicle safety system. Detailed Implementation

[0039] This invention relates to a vehicle safety system implementing an enhanced discrimination algorithm that can distinguish and classify slope rollover collision events and embankment rollover collision events. The enhanced discrimination algorithm implemented by the vehicle safety system can also distinguish and classify embankment events and tripping events. Furthermore, the enhanced discrimination algorithm implemented by the vehicle safety system can further distinguish and classify hard ground tripping events, moderate ground tripping events, and soft ground tripping events.

[0040] Because this invention relates to enhanced discrimination of the events mentioned in the preceding paragraph, vehicle safety systems are shown and described herein as including certain components and implementing algorithms necessary to perform these specific enhanced discrimination functions. Those skilled in the art will understand that vehicle safety systems may include components other than those shown and described herein, and may perform discrimination algorithms other than those shown and described herein.

[0041] refer to Figure 1 According to one example configuration, the vehicle safety system 10 includes a central control unit (CCU) 50 operable to actuate one or more actuable restraint devices 20, such as left / right seatbelt pretensioners (anchors and / or retractors), left / right curtain airbags, left / right chest airbags, and left / right lateral airbags. The CCU 50 is also operable to control the actuation of other protective devices, such as front airbags and knee airbags.

[0042] The CCU 50 is also operable to control other vehicle safety features 22, such as emergency notifications, automatic door unlocking commands, and electric vehicle (EV) high-voltage power cut-off commands. Emergency notifications may include, for example, via vehicle-based emergency assistance systems (such as GM...). Ford and Chrysler A notification or request for emergency response (fire / EMS) issued by the vehicle's electrical system. For electric vehicles, a high-voltage disconnection command can decouple the vehicle battery from the vehicle's electrical system to reduce the risk of electric shock or fire due to electrical short circuits or faults.

[0043] CCU 50 includes one or more sensors operable to provide signals indicative of the vehicle's linear acceleration and / or angular acceleration and / or rate of movement in different directions and relative to different vehicle axes. These sensors may be locally mounted in or on the CCU 50 itself, or remotely connected to the CCU, for example, via wires. These vehicle axes include an X-axis that extends longitudinally within the vehicle along its forward / rearward direction of travel. A Y-axis extends laterally within the vehicle, perpendicular to the X-axis. A Z-axis extends vertically within the vehicle, perpendicular to both the X and Y axes. The X, Y, and Z axes can be considered to intersect at the vehicle's center of gravity.

[0044] CCU 50 includes an accelerometer 52 for sensing the vehicle's lateral (Y-axis) acceleration (CCU_1Y). CCU 50 also includes an accelerometer 54 for sensing the vehicle's vertical (Z-axis) acceleration (CCU_6Z). CCU 50 also includes a pitch rate sensor 56 for sensing the vehicle's pitch rate value (CCU_5P), i.e., the pitch rate around the vehicle's Y-axis. CCU 50 further includes a roll rate sensor 58 for sensing the vehicle's roll rate value (CCU_4R), i.e., the roll rate around the vehicle's X-axis. It may be desirable to position the sensors on or near the respective axes along or around which they sense vehicle motion. Since the sensors can be locally mounted on CCU 50, it may be desirable to mount the CCU at or near the vehicle's center of gravity.

[0045] The hardware and software configurations of the CCU implemented in the vehicle safety system are known in the art. Therefore, a detailed description of the hardware configuration of the CCU 50 is not required for a person skilled in the art to understand and recognize the vehicle safety system 10. Figure 1 The CCU 50 includes a central processing unit (CPU) 60, such as a microcomputer, which is configured to receive signals CCU_1Y, CCU_6Z, CCU_4R and CCU_5P from the respective sensors, perform vehicle measurement calculations 70 on these signals, and execute an enhanced discrimination algorithm 80 using the calculated measurements.

[0046] The vehicle metrics generated by calculation 62 include:

[0047] • Average shift of vehicle lateral Y-axis acceleration (CCU_1Y_AMA).

[0048] • Average vertical Z-axis acceleration of the vehicle (CCU_6Z_AMA).

[0049] • Vehicle roll rate, also known as roll acceleration (D_RATE).

[0050] • Vehicle pitch rate (P_RATE).

[0051] • Vehicle roll rate (R_RATE) and vehicle roll angle (R_ANGLE).

[0052] The enhanced discrimination algorithm 80 includes a rollover discrimination algorithm 82, a slope / embankment discrimination algorithm 84, and a hard soil / medium soil / soft soil discrimination algorithm 86. The CCU 50 is configured to perform vehicle measurement calculations 70 and the enhanced discrimination algorithm 80, and to determine which actuable restraint devices 20 (if any) should be actuated.

[0053] Figure 2 The vehicle measurement calculation 70 performed by CCU 50 is shown. Figure 2 The elements of the vehicle measurement calculation 70 shown are referred to herein as “functions” performed internally by the CCU 50.

[0054] Side roll rate measurement

[0055] Based on vehicle measurement calculation 70, analog-to-digital converter (ADC) function 100 converts the roll rate CCU_4R signal into a digital signal. ADC 100 can perform a 10-bit 8-sample summation, for example, at 125 μs. At track / bias function 102, track checks and bias adjustments are performed, for example, at 1 ms. The digitized and biased roll rate CCU_4R is passed to, for example, a high-pass filter (HPF) function 104, which can be selected to have a time constant that causes the filter function to be reset after a predetermined time period (e.g., T = 8 seconds). The high-pass filtered roll rate CCU_4R generated at HPF function 104 is passed to a low-pass filter (LPF) function 106 with a variable angle frequency (also called cutoff frequency) determined by rate transition function 108. For example, at rate transition function 108, the angle / cutoff frequency is selected from:

[0056] Rate_LPF_Shift Turning angle / cutoff frequency 3 21.25Hz 4 10.27Hz 5 5.05Hz

[0057] LPF function 106 is generated in Enhanced Discriminant Algorithm 80 (see...) Figure 1 The roll rate metric R_RATE implemented in the algorithm 80 has a value indicating the vehicle's roll rate (i.e., angular velocity). R_RATE is passed to an integral high-pass filter (IHPF) function 110, which includes an integrator function and a dual-time-constant high-pass filter function. The IHPF function 110 integrates the R_RATE signal to produce a value indicating a defined relative roll angle of the vehicle. The IHPF function 110 also performs high-pass filtering on the R_RATE signal. The IHPF function 110 generates the values ​​in the enhanced discrimination algorithm 80 (see...). Figure 1 The metric R_ANGLE implemented in ).

[0058] R_ANGLE indicates the vehicle's normalized roll angle, a measure of the vehicle's relative angular rotation in response to a sensed roll rate. The IHPF function 110 can reset R_ANGLE based on the time constant of the high-pass filter function, such that R_ANGLE provides an indication of angular rotation during the occurrence of the detected roll rate. Therefore, R_ANGLE may not indicate the vehicle's actual angular orientation relative to the ground. In this way, the determination of a vehicle rollover condition does not necessarily depend on the determination of the vehicle's initial angular orientation relative to the ground or road.

[0059] The high-pass filtered roll rate CCU_4R generated at HPF function 104 is also passed to a low-pass filter (LPF) function 112 (which is also a variable steering angle / cutoff frequency LPF) determined by rate transition function 108 (see table above). LPF function 112 produces a roll rate metric R_RATE_LPF, which has a value indicating the vehicle's roll rate (i.e., angular velocity). R_RATE_LPF is used in the enhanced discrimination algorithm 80 (see table above). Figure 1 Implemented in (). Since R_RATE_LPF is the product of high-pass and low-pass filtering, it can be regarded as the side-tilt rate value after band-pass filtering.

[0060] The high-pass filtered yaw rate CCU_4R generated at HPF function 104 is also passed to moving average function 120, and then to moving average function 122. Each moving average function 120, 122 can be, for example, adjustable to select the number of samples, for example, 1 to 32 samples. Moving average functions 120, 122 smooth the variation in yaw rate, thereby producing a result in the enhanced discrimination algorithm 80 (see...). Figure 1 The metric R_RATE_2 implemented in )

[0061] R_RATE_2 is provided to the difference function 124, in which the difference between the current sample and the previous sample is compared. This produces the differential roll rate metric D_RATE, which indicates the rate of change of the roll rate, i.e., acceleration. This roll acceleration D_RATE is the angular acceleration of the vehicle about the vehicle's X-axis. The roll acceleration D_RATE is used in the augmentation discrimination algorithm 80 (see...). Figure 1 Implemented in ).

[0062] Pitch rate measurement

[0063] Based on vehicle measurement calculation 70, analog-to-digital converter (ADC) function 130 converts the pitch rate CCU_5P signal into a digital signal. ADC 130 can perform 10-bit 8-sample summation, for example, at 125 μs. At track / bias function 132, track checking and bias adjustment are performed, for example, at 1 ms. The digitized and biased pitch rate CCU_5P is passed to, for example, a high-pass filter (HPF) function 134, which can be selected to have a time constant that causes the filter function to be reset after a predetermined time period (e.g., T = 8 seconds). The high-pass filtered pitch rate CCU_5P generated at HPF function 134 is passed to low-pass filter (LPF) function 136. LPF function 136 generates the signal in the enhancement discrimination algorithm 80 (see...). Figure 1 The pitch rate metric P_RATE implemented in ) has a value indicating the vehicle's pitch rate (i.e., angular velocity).

[0064] Lateral acceleration measurement

[0065] Based on vehicle measurement calculation 70, analog-to-digital converter (ADC) function 140 converts the lateral (Y-axis) acceleration CCU_1Y signal. ADC 140 can perform 10-bit 8-sample summation, for example, at 125 μs. At track / bias function 142, track checking and bias adjustment are performed, for example, at 1 ms. The digitized and biased lateral acceleration CCU_1Y is passed to, for example, a high-pass filter (HPF) function 144 with a time constant that causes the filter function to be reset after a predetermined time period (e.g., T = 8 seconds). The high-pass filtered lateral acceleration CCU_1Y generated at HPF function 144 is then passed to low-pass filter (LPF) function 146. The low-pass filtered lateral acceleration CCU_1Y value generated at LPF function 146 is passed to moving average boxes 148 and 150, which produce the lateral acceleration metrics CCU_1Y_AMA and CCU_1Y_AMA_SAFE. The number of samples included in each of the moving average functions 148 and 150 can be adjusted within a predetermined range, such as 1 to 32 samples. CCU_1Y_AMA and CCU_1Y_AMA_SAFE are used in the enhanced discrimination algorithm 80 (see...). Figure 1 The average lateral acceleration shift implemented in ).

[0066] Vertical acceleration measurement

[0067] Based on vehicle measurement calculation 70, analog-to-digital converter (ADC) function 160 converts the vertical (Z-axis) acceleration CCU_6Z signal. ADC 140 can perform 10-bit 8-sample summation, for example, at 125 μs. At track / bias function 162, track checking and bias adjustment are performed, for example, at 1 ms. The digitized and biased vertical acceleration CCU_6Z is passed to, for example, a high-pass filter (HPF) function 164 with a time constant that causes the filter function to be reset after a predetermined time period (e.g., T = 8 seconds). The high-pass filtered lateral acceleration CCU_6Z generated at HPF function 164 is then passed to low-pass filter (LPF) function 166. The low-pass filtered lateral acceleration CCU_6Z value generated at LPF function 166 is passed to moving average boxes 168 and 170, which produce the lateral acceleration metrics CCU_6Z_AMA and CCU_6Z_AMA_SAFE. The number of samples included in each of the moving average functions 168 and 170 can be adjusted within a predetermined range, such as 1 to 32 samples. CCU_6Z_AMA and CCU_6Z_AMA_SAFE are used in the enhanced discrimination algorithm 80 (see...). Figure 1 The average lateral acceleration shift implemented in ).

[0068] Lateral Threshold

[0069] The ability to differentiate between various types of rollover events allows for the customization of trigger-actuable restraint devices 20 (see...). Figure 1 The application thresholds of these thresholds. Figure 3 The figure illustrates the application of threshold metrics for actuating actuable restraint devices based on thresholds categorized according to different rollover events. For example... Figure 3 As shown, the threshold metric is applied based on a comparison of R_RATE and R_ANGLE. Figure 3 The threshold determination demonstrates left lateral tilt (i.e., tilt toward the driver's side) as indicated by the R_RATE and R_ANGLE values ​​in one direction (e.g., positive direction), and right lateral tilt (i.e., tilt toward the passenger's side) as indicated by the R_RATE and R_ANGLE values ​​in the opposite direction (e.g., negative direction).

[0070] like Figure 3 As shown, soft soil rollover events are classified with the lowest threshold for triggering the application of actuated restraint devices. Moderate land conditions have the second lowest threshold for triggering the application of actuated restraint devices, followed by hard soil, slopes, and embankments. Normal conditions (i.e., where no other thresholds are classified) have the highest application trigger threshold. These application threshold triggers can be latched, in which case, when the metric enters… Figure 3When the reset box is shown, the latch can be reset.

[0071] Furthermore, as shown in the figure, the trigger threshold may require the detection of a certain threshold R_ANGLE, as indicated by the vertical lines, i.e., the vertical lines to the left of the left lateral tilt threshold and to the right of the right lateral tilt threshold. Once the necessary R_ANGLE is met, the threshold will remain essentially unchanged, meaning that the lateral tilt rate (R_RATE) largely determines which thresholds (if any) are met.

[0072] Enhanced discrimination algorithm

[0073] According to the present invention, the enhanced discrimination algorithm utilizes... Figure 2 The vehicle metrics defined in the data are used to identify and classify vehicle rollover events, enabling the use of the correct rollover threshold. Figure 3 This determines whether actuable restraints should be actuated. Rollover events can be classified as slope events, embankment events, or land (hard soil, moderate soil, soft soil) events. "Discrimination" means that the classification of a vehicle rollover event not only identifies the type of rollover event that has occurred, but also identifies that the event is not one of the (multiple) other events that the algorithm is configured to discriminate among. Therefore, for example, in a case where the enhanced discrimination algorithm is configured to discriminate between slope rollover events and embankment rollover events, classifying the event as a slope rollover event also implies that the event is not an embankment rollover event.

[0074] A slope event is an event in which a vehicle engages with an upward-sloping surface that causes the vehicle to tilt / roll over. An embankment event is an event in which a vehicle engages with a downward-sloping surface that causes the vehicle to tilt / roll over. A land event is an event in which a vehicle moves laterally onto land that causes the vehicle to tilt / roll over. Land events can be hard land events, moderate land events, or soft land events. Soft land is soft ground that provides low resistance to further lateral movement of the vehicle, such as ordinary grass / lawn. Hard land is classified as solid ground, gravel, rock, curbs, or other road-level obstacles that cause a vehicle to "trip," where the tire / wheel road contact area is caught or otherwise blocked due to high resistance to further lateral movement of the vehicle. Moderate land is classified as providing a level or resistance to lateral movement of the vehicle that is between that of a hard land event and a soft land event.

[0075] When differentiating between various vehicle rollover events, it can be difficult to distinguish ramp events (e.g., left ramp and right embankment) from embankment events (e.g., left ramp and right embankment) because they exhibit similar characteristics. Differentiating land events from embankment events can also be difficult because they develop in a similar manner. In other words, embankment events can easily be confused with ramp and land events. The enhanced discrimination algorithm 80 implemented in vehicle safety system 10 utilizes pitch rate P_RATE and roll acceleration D_RATE to enhance the classification of these events. In this specification, this improved classification may be referred to as improved differentiation of these events. This is because classification metrics implemented in the enhanced discrimination algorithm have been developed such that their values ​​differ sufficiently to allow for differentiation between these events.

[0076] Slope rollover events and embankment rollover events develop slowly. Vehicles move uphill during slope rollovers and downhill during embankment rollovers. Physically, the vertical acceleration of a vehicle can distinguish between slopes and embankments. However, the vertical acceleration during embankment events is not significant. Therefore, using vertical acceleration as a determining factor may be difficult to differentiate between embankment events and slope or land events. Advantageously, the enhanced discrimination algorithm 80 utilizes... Figure 2 The vehicle metrics identified in the data are used to enhance the classification of ramps and embankments.

[0077] The enhanced discriminative algorithm implements a so-called classification metric to distinguish between various rollover events. The classification metric is a measure applied to two vehicles (see [link to relevant documentation]). Figure 2 The comparison, that is, the comparison between metric A and metric B, can be displayed graphically or as a graph. Although displayed graphically, it will be understood that the categorical metrics can be calculated in the CPU 60 through mathematical operations.

[0078] Slope incidents and embankment incidents

[0079] A ramp incident is an event that causes a vehicle to roll over when one side of it drives onto a structure that acts as a ramp for that side of the vehicle. For example, this can happen when one side of a vehicle is lifted up by a concrete roadway obstacle. An embankment incident is an event that causes a vehicle to roll over when one side of it drives onto an embankment on one side of the road. Comparing left-side rollover incidents, for example, a left ramp rollover is caused by the right side of the vehicle moving up / accelerating along a ramp structure, resulting in a leftward tilt about the vehicle's longitudinal X-axis. A left embankment rollover is caused by the left side of the vehicle moving down / accelerating along an embankment structure, resulting in a leftward tilt about the vehicle's longitudinal X-axis. Differentiating between ramp and embankment incidents may be advantageous to provide enhanced occupant protection.

[0080] Slope identification

[0081] Figure 4 The ramp discrimination algorithm 84 is shown. Ramp discrimination algorithm 84 is used to determine whether a ramp threshold should be used when applying an actuable constraint device (see [reference]). Figure 3 For left-side roll events (i.e., the vehicle rolls to the left or to the driver's side), the following is shown: Figure 4 The ramp detection algorithm 84. However, it should be understood that... Figure 4 The algorithm shown also applies to right-leaning events, the only difference being that the sign (+ / -) of the values ​​used in the classification metric is reversed. In other words, the classification metric used for right-leaning events will be... Figure 4 The classification measures shown are the same, except that the signs of the corresponding axes for different values ​​within a classification measure will be reversed, for example, positive becomes negative, and vice versa.

[0082] The ramp discrimination algorithm scored 84 points and implemented four different classification measures to discriminate ramp events. These four classification measures are:

[0083] ·CCU_1Y_AMA vs R_ANGLE

[0084] ·CCU_6Z vs R_ANGLE

[0085] ·D_RATE vs R_ANGLE

[0086] ·P_RATE vs R_ANGLE

[0087] In response to all classification measures 200, 202, 204, and 206 fed into AND box 208 being satisfied, a determination of using a ramp threshold (i.e., a Boolean value of one, referred to herein as ON) is made at box 212. The use-ramp threshold determination 212 can be a latch determination, as shown in latch box 210. Therefore, once AND box 208 is satisfied, the use-ramp threshold 212 is ON and remains ON due to latch 210, even after the classification measures fed into AND box 208 are no longer ON. The classification measures 200, 202, 204, and 206 fed into AND box 208 are described in the following paragraphs.

[0088] The lateral acceleration to roll angle classification metric 200 uses CCU_1Y_AMA and R_ANGLE to generate the output fed to box 208. As shown, the lateral acceleration to roll angle classification metric 200 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the lateral acceleration to roll angle classification metric 200 represents the metric when the vehicle is experiencing a hill rollover event. The lateral acceleration to roll angle classification metric 200 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the lateral acceleration to roll angle classification metric 200 does not produce a reliable distinction between hill rollover events and embankment events. Therefore, when the other classification metrics 202, 204, and 206 are consistent in distinguishing the occurrence of hill rollover events, the lateral acceleration to roll angle classification metric 200 is used as confirmation in the hill roll discrimination algorithm 84 (at box 208).

[0089] The vertical acceleration to roll angle classification metric 202 uses CCU_6Z_AMA and R_ANGLE to generate the output fed to box 208. As shown, the vertical acceleration to roll angle classification metric 202 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the vertical acceleration to roll angle classification metric 202 represents the metric when the vehicle is experiencing a hill rollover event. The vertical acceleration to roll angle classification metric 202 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the vertical acceleration to roll angle classification metric 202 does not produce a reliable distinction between hill roll events and embankment events. Therefore, when the other classification metrics 200, 204, and 206 are consistent in distinguishing the occurrence of hill roll events, the vertical acceleration to roll angle classification metric 202 is used as confirmation in the hill roll discrimination algorithm 84 (at box 208).

[0090] The angular acceleration or roll acceleration to roll angle classification metric 204 uses D_RATE and R_ANGLE to generate the output fed into box 208. As shown, the roll acceleration to roll angle classification metric 204 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the roll acceleration to roll angle classification metric 204 represents the metric when the vehicle is experiencing a hill rollover event. The roll acceleration to roll angle classification metric 204 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that comparing roll acceleration with roll angle reliably distinguishes between hill rollover events and embankment events. This is because the roll acceleration to roll angle classification metric 204 produces a sufficiently distinguishable metric value to reliably distinguish between hill rollover events and embankment events. Therefore, the roll acceleration to roll angle classification metric 204 is used as a positive metric for hill rollover events in the hill roll discrimination algorithm 84.

[0091] The pitch rate versus roll angle classification metric 206 uses P_RATE and R_ANGLE to generate the output fed into box 208. As shown, the pitch rate versus roll angle classification metric 206 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the pitch rate versus roll angle classification metric 206 represents the metric when the vehicle is experiencing a hill rollover event. The pitch rate versus roll angle classification metric 206 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that comparing the pitch rate with the roll angle reliably distinguishes between hill rollover events and embankment events. This is because the pitch rate versus roll angle classification metric 206 produces a sufficiently distinguishable metric value to reliably distinguish between hill rollover events and embankment events. Therefore, the pitch rate versus roll angle classification metric 206 is used as a positive metric for hill rollover events in the hill rollover discrimination algorithm 84.

[0092] Embankment identification

[0093] Figure 5 The diagram illustrates embankment discrimination algorithm 86. Embankment discrimination algorithm 86 is used to determine whether to use an embankment threshold when applying an actuable constraint device (see [link]). Figure 3 For left-side roll events (i.e., the vehicle rolls to the left or to the driver's side), the following is shown: Figure 5 The embankment discrimination algorithm 86. However, it should be understood that... Figure 5 The algorithm shown also applies to right-leaning events, the only difference being that the sign (+ / -) of the values ​​used in the metric is reversed. In other words, the metric used for right-leaning events will be the same as... Figure 5 The measurements shown are the same, but the signs of the corresponding axes for different measurement values ​​will be reversed, for example, positive becomes negative, and vice versa.

[0094] The embankment discrimination algorithm 86 implements four different metrics to discriminate slope events. These four metrics are:

[0095] ·CCU_1Y_AMA vs R_ANGLE

[0096] ·CCU_6Z vs R_ANGLE

[0097] ·D_RATE vs R_ANGLE

[0098] ·P_RATE vs R_ANGLE

[0099] In response to the satisfaction of all classification measures 220, 222, 224, and 226 fed into AND box 228, a determination of the use embankment threshold is made at box 232, i.e., a Boolean value of one, referred to herein as ON. The use embankment threshold determination 232 can be a latch determination, as shown in latch box 230. Therefore, once AND box 228 is satisfied, the use embankment threshold 232 is ON and remains ON due to latch 230, even after the measures fed into AND box 228 are no longer ON. The classification measures 220, 222, 224, and 226 fed into AND box 228 are described in the following paragraphs.

[0100] The lateral acceleration to roll angle classification metric 220 uses CCU_1Y_AMA and R_ANGLE to generate the output fed into box 228. As shown, the lateral acceleration to roll angle classification metric 220 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the lateral acceleration to roll angle classification metric 220 represents the metric when the vehicle is experiencing an embankment rollover event. The lateral acceleration to roll angle classification metric 220 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the lateral acceleration to roll angle classification metric 220 does not produce a reliable distinction between ramp events and embankment events. Therefore, when the other classification metrics 222, 224, and 226 are consistent in distinguishing the occurrence of an embankment event, the lateral acceleration to roll angle classification metric 220 is used as confirmation in the embankment discrimination algorithm 86 (at box 228).

[0101] The vertical acceleration to roll angle classification metric 222 uses CCU_6ZY_AMA and R_ANGLE to generate the output fed into box 228. As shown, the vertical acceleration to roll angle classification metric 222 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the vertical acceleration to roll angle classification metric 222 represents the metric when the vehicle is experiencing an embankment rollover event. The vertical acceleration to roll angle classification metric 222 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the vertical acceleration to roll angle classification metric 222 does not produce a reliable distinction between ramp events and embankment events. Therefore, when the other classification metrics 220, 224, and 226 are consistent in distinguishing the occurrence of an embankment event, the vertical acceleration to roll angle classification metric 222 is used as confirmation in the embankment discrimination algorithm 86 (at box 228).

[0102] The angular acceleration or roll acceleration to roll angle classification metric 224 uses D_RATE and R_ANGLE to generate the output fed into box 228. As shown, the metric 224 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the roll acceleration to roll angle classification metric 224 represents the metric when the vehicle is experiencing an embankment rollover event. The roll acceleration to roll angle classification metric 224 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that comparing roll acceleration with roll angle reliably distinguishes between ramp events and embankment events. This is because the roll acceleration to roll angle classification metric 224 produces a sufficiently distinguishable metric value to reliably distinguish between ramp events and embankment events. Therefore, the roll acceleration to roll angle classification metric 224 is used as a positive metric for embankment events in the embankment discrimination algorithm 86.

[0103] The pitch rate versus roll angle classification metric 226 uses P_RATE and R_ANGLE to generate the output fed into box 228. As shown, the pitch rate versus roll angle classification metric 226 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the pitch rate versus roll angle classification metric 226 represents the metric when the vehicle is experiencing an embankment rollover event. The pitch rate versus roll angle classification metric 226 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that comparing the pitch rate with the roll angle reliably distinguishes between ramp events and embankment events. This is because the pitch rate versus roll angle classification metric 226 produces a sufficiently distinguishable metric value to reliably distinguish between ramp events and embankment events. Therefore, the pitch rate versus roll angle classification metric 226 is used as a positive embankment event criterion in the embankment discrimination algorithm 86.

[0104] Distinguishing between slope events and embankment events

[0105] From the preceding paragraphs and viewing Figure 4 and Figure 5It can be seen that slope discrimination algorithm 84 and embankment discrimination algorithm 86 use P_RATE and D_RATE to enhance the distinction between slope rollover events and embankment rollover events. Slope discrimination algorithm 84 and embankment discrimination algorithm 86 use CCU_1Y_AMA and CCU_6Z_AMA to verify or confirm the discrimination between slope rollover events and embankment rollover events. Since embankment rollover has the smallest D_RATE among all types of rollover events, D_RATE can be implemented in the metric to distinguish embankment events from slope events. Therefore, if P_RATE is unavailable, D_RATE can distinguish embankment events from slope events; that is, AND gates 208 and 228 can be three-input gates with classification metrics 206 and 226 omitted, respectively. However, it should be noted that using both P_RATE and D_RATE (i.e., using...) Figure 4 and Figure 5 All four measures shown are more robust.

[0106] Slope events and embankment events produce similar lateral accelerations and roll rates. Therefore, as... Figure 4 and Figure 5 As shown, the trigger regions for lateral acceleration and roll rate classification measures 200 and 220 can be similar or identical. Because the lateral acceleration is similar in both types of events, classification measures 200 and 220, which compare these values, do not provide a distinction between slope events and embankment events. Since the lateral acceleration CCU_1Y_AMA responds predictably and repeatably to both slope and embankment events, they can be used to confirm or validate the occurrence of these events even if they cannot distinguish between slope and embankment.

[0107] Slope events and embankment events produce significantly different roll accelerations (D_RATE). Therefore, as... Figure 4 and Figure 5 As shown, the trigger region for the roll acceleration to the roll rate metric in slope discrimination 204 differs from that in embankment discrimination 224. Therefore, if a slope event occurs, classification metric 204 will be ON, while if an embankment event occurs, it will be OFF. Conversely, classification metrics 204 and 224 can thus discriminate between these events in a predictable, repeatable, and reliable manner.

[0108] Slope events and embankment events also produce significantly different pitch rates (P_RATE). Therefore, as... Figure 4 and Figure 5 As shown, the trigger region for the pitch rate to lateral tilt rate measure of slope discrimination 206 is different from that of the pitch rate to lateral tilt rate measure of embankment discrimination 226. Therefore, classification measures 206 and 226 can thus discriminate between these events in a predictable, repeatable, and reliable manner.

[0109] Embankment incident and land incident

[0110] As previously described, an embankment incident is an event in which a vehicle rolls over due to one side of it entering an embankment on the side of the road. A land incident is an event in which a vehicle that has laterally slid off the road comes into contact with land or other material, which then grips the tires / wheels, causing the vehicle to roll over. Comparing left-side rollover incidents, for example, a left embankment rollover is caused by the left side of the vehicle moving down / accelerating along the embankment structure, resulting in a leftward tilt about the vehicle's longitudinal X-axis. A left land rollover is caused by a sudden cessation or deceleration of lateral movement of the vehicle to the left along the vehicle's lateral (Y-axis) or toward the driver's side, resulting in a leftward tilt about the vehicle's longitudinal X-axis. To provide enhanced occupant protection, it may be advantageous to differentiate between embankment and land incidents.

[0111] Hard soil identification

[0112] Figure 6 The image shows a hard ground discrimination algorithm 88. The hard ground discrimination algorithm 88 is used to determine whether a hard ground threshold should be used when applying an actuable constraint device (see [reference]). Figure 3 For left-side roll events (i.e., the vehicle rolls to the left or to the driver's side), the following is shown: Figure 6 The hard soil discrimination algorithm 88. However, it should be understood that... Figure 6 The algorithm shown also applies to right-leaning events, the only difference being that the sign (+ / -) of the values ​​used in the metric is reversed. In other words, the metric used for right-leaning events will be the same as... Figure 6 The measurements shown are the same, but the signs of the corresponding axes for different measurement values ​​will be reversed, for example, positive becomes negative, and vice versa.

[0113] Land events refer to incidents where a vehicle laterally slides off the road and onto the ground. An enhanced discrimination algorithm categorizes land events into hard ground events, moderate ground events, and soft ground events, providing customized responses for the unique collision characteristics involved in these collision scenarios. Hard ground events can also be classified as tripping events; for example, when a vehicle laterally slides off the road and hits a curb, the curb can "trip" the vehicle, leading to a rapid rollover.

[0114] The hard soil discrimination algorithm 88 implements four different metrics to distinguish between slope events and embankment events. These four metrics are:

[0115] ·D_RATE to R_RATE_2

[0116] ·CCU_6Z vs R_ANGLE

[0117] ·CCU_1Y_AMA vs R_ANGLE

[0118] ·D_RATE vs R_ANGLE

[0119] In response to the satisfaction of both the classification metric 260 fed to AND box 252 and the latched hard soil box 254, a determination to use hard soil is made at box 250, i.e., a Boolean value of one, referred to herein as ON. The latched hard soil box 254 is latched, meaning that once AND box 256 is satisfied, latched hard soil 254 is ON and remains ON, even after the metric fed to AND box 256 is no longer ON. Classification metrics 262, 264, and 266 are fed to AND box 256. Classification metrics 260, 262, 264, and 266 are described in the following paragraphs.

[0120] The angular acceleration or roll acceleration to roll rate classification metric 260 uses D_RATE and R_RATE_2 to generate the output fed to box 252. As shown, the roll acceleration to roll rate classification metric 260 is ON when it is in the hard-triggered region defined by the solid line of the metric. The triggering region indicated by the classification metric 260 is the region (hard, medium, soft) into which the metric first enters from the shaded area. In other words, the triggering region is latched in the classification metric 260. Therefore, if the metric first enters the hard-triggered region and then moves to the medium and / or soft-triggered region, the hard-triggered region indication remains ON, while the medium and / or soft-triggered region indication remains OFF. The solid line in the roll acceleration to roll rate classification metric 260 represents the metric when the vehicle is experiencing a rollover event on hard ground.

[0121] Hard ground events (such as curb tripping) result in high-amplitude roll acceleration due to the high sliding resistance immediately provided by the hard ground surface (i.e., the curb). The roll acceleration-to-roll rate classification metric 260, configured to classify such high-amplitude roll acceleration in hard-triggered areas, provides a reliable discrimination for hard ground events. Furthermore, tests show that the roll acceleration-to-roll rate classification metric 260 produces a metric sufficiently distinguishable from embankment events to reliably differentiate them. Therefore, the roll acceleration-to-roll rate classification metric 260 can be used as a positive discrimination for hard ground events in the hard ground discrimination algorithm 88.

[0122] The vertical acceleration to roll angle classification metric 262 uses CCU_6ZY_AMA and R_ANGLE to generate the output fed to box 256. As shown, the vertical acceleration to roll angle classification metric 262 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the vertical acceleration to roll angle classification metric 262 represents the metric when the vehicle is experiencing a hard ground rollover event. The vertical acceleration to roll angle classification metric 262 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the vertical acceleration to roll angle classification metric 262 does not produce a reliable distinction between hard ground events and embankment events. Therefore, when other classification metrics 264 and 266 are consistent in distinguishing the occurrence of hard ground events, the vertical acceleration to roll angle classification metric 262 is used as confirmation in the hard ground discrimination algorithm 88 (at box 256).

[0123] The lateral acceleration to roll angle classification metric 264 uses CCU_1Y_AMA and R_ANGLE to generate the output fed into box 256. As shown, the lateral acceleration to roll angle classification metric 264 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the lateral acceleration to roll angle classification metric 264 represents the metric when the vehicle is experiencing a hard ground rollover event. The lateral acceleration to roll angle classification metric 264 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that comparing roll acceleration with roll angle reliably distinguishes between hard ground events and embankment events. This is because the roll acceleration to roll angle classification metric 264 produces a sufficiently distinguishable metric value to reliably distinguish between hard ground events and embankment events. Therefore, the roll acceleration to roll angle classification metric 264 is used as a positive metric for hard ground events in the hard ground discrimination algorithm 88.

[0124] The angular acceleration or roll acceleration classification metric 266 uses D_RATE and R_ANGLE to generate the output fed to box 256. As shown, the roll acceleration classification metric 266 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the roll acceleration classification metric 256 represents the metric when the vehicle is experiencing a hard ground rollover event. The roll acceleration classification metric 266 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the roll acceleration classification metric 266 does not produce a reliable distinction between hard ground events and embankment events. Therefore, when other classification metrics 262 and 264 are consistent in distinguishing the occurrence of hard ground events, the roll acceleration classification metric 266 is used as confirmation in the hard ground discrimination algorithm 88 (at box 256).

[0125] Moderate land identification

[0126] Figure 7 The image shows a moderate land discrimination algorithm 90. The moderate land discrimination algorithm 90 is used to determine whether to use a moderate land threshold when applying an actuable constraint device (see [link]). Figure 3 For left-side roll events (i.e., the vehicle rolls to the left or to the driver's side), the following is shown: Figure 7 The moderate land identification algorithm is 90. However, it should be understood that... Figure 7 The algorithm shown also applies to right-leaning events, the only difference being that the sign (+ / -) of the values ​​used in the metric is reversed. In other words, the metric used for right-leaning events will be the same as... Figure 7 The measurements shown are the same, but the signs of the corresponding axes for different measurement values ​​will be reversed, for example, positive becomes negative, and vice versa.

[0127] Land events refer to incidents where a vehicle laterally slides off the road and onto the ground. An enhanced discrimination algorithm categorizes land events into hard ground events, moderate ground events, and soft ground events, providing customized responses for the unique collision characteristics involved in these collision scenarios. Moderate ground events can also be classified as those that do not produce the high-amplitude acceleration associated with hard ground or tripping events, but produce acceleration higher than that associated with soft ground events (such as lawns or grass). Moderate ground events can be events between hard ground and soft ground events; for example, when a vehicle laterally slides off the road onto dry, hard ground and / or gravel, this results in a rollover development that is slower than a hard ground event but faster than a soft ground event.

[0128] The moderate land discrimination algorithm 90 implements four different metrics to distinguish between slope events and embankment events. These four metrics are:

[0129] ·D_RATE to R_RATE_2

[0130] ·CCU_6Z vs R_ANGLE

[0131] ·CCU_1Y_AMA vs R_ANGLE

[0132] ·D_RATE vs R_ANGLE

[0133] In response to the satisfaction of both the classification metric 280 fed to AND box 272 and the latched medium-soil box 274, a determination of use medium-soil (USE MID-SOIL) is made at box 270, i.e., a Boolean value of one, referred to herein as ON. The latched medium-soil box 274 is latched, meaning that once AND box 276 is satisfied, the latched medium-soil box 274 is ON and remains ON, even after the metric fed to AND box 276 is no longer ON. Classification metrics 282, 284, and 286 are fed to AND box 276. Classification metrics 280, 282, 284, and 286 are described in the following paragraphs.

[0134] The angular acceleration or roll acceleration to roll rate classification metric 280 uses D_RATE and R_RATE_2 to generate the output fed to box 272. As shown, the roll acceleration to roll rate classification metric 280 is ON when it is in the moderate trigger region defined by the solid line of the metric. The trigger region indicated by the classification metric 280 is the region in the region (hard, moderate, soft) that the metric first enters from the shaded area. In other words, the trigger region is latched in the classification metric 280. Therefore, if the metric first enters the moderate trigger region and then moves to the hard and / or soft trigger regions, the moderate trigger region indication remains ON, while the hard and / or soft trigger region indication remains OFF.

[0135] Moderate land events (such as hard, dry land and / or gravel) result in lower roll acceleration amplitudes than hard land events due to the rapid provision of high sliding resistance on the moderate land surface. A roll acceleration-to-roll rate classification metric 280, configured to classify such roll accelerations in moderate triggering areas, provides a reliable distinction for moderate land events. Furthermore, tests show that the roll acceleration-to-roll rate classification metric 280 produces a metric sufficiently distinguishable between moderate land events and embankment events. Therefore, the roll acceleration-to-roll rate classification metric 280 can be used as a positive distinction for moderate land events in the moderate land discrimination algorithm 90. The solid line in the roll acceleration-to-roll rate classification metric 280 represents the metric when a vehicle is experiencing a moderate land rollover event.

[0136] The vertical acceleration to roll angle classification metric 282 uses CCU_6ZY_AMA and R_ANGLE to generate the output fed to box 276. As shown, the vertical acceleration to roll angle classification metric 282 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the vertical acceleration to roll angle classification metric 282 represents the metric when the vehicle is experiencing a moderate land rollover event. The vertical acceleration to roll angle classification metric 282 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the vertical acceleration to roll angle classification metric 282 does not produce a reliable distinction between moderate land events and embankment events. Therefore, when other classification metrics 284 and 286 are consistent in distinguishing the occurrence of moderate land events, the vertical acceleration to roll angle classification metric 282 is used as confirmation in the moderate land discrimination algorithm 90 (at box 276).

[0137] The lateral acceleration versus roll angle classification metric 284 uses CCU_1Y_AMA and R_ANGLE to generate the output fed into box 276. As shown, the lateral acceleration versus roll angle classification metric 284 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the lateral acceleration versus roll angle classification metric 284 represents the metric when the vehicle is experiencing a moderate land rollover event. The lateral acceleration versus roll angle classification metric 284 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that comparing roll acceleration with roll angle reliably distinguishes between moderate land events and embankment events. This is because the roll acceleration versus roll angle classification metric 284 produces a sufficiently distinguishable metric value to reliably distinguish between moderate land events and embankment events. Therefore, the roll acceleration versus roll angle classification metric 284 is used as a positive criterion for moderate land events in the moderate land discrimination algorithm 90.

[0138] The angular acceleration or roll acceleration classification metric 286 uses D_RATE and R_ANGLE to generate the output fed into box 276. As shown, the roll acceleration classification metric 286 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the roll acceleration classification metric 286 represents the metric when the vehicle is experiencing a moderate land rollover event. The roll acceleration classification metric 286 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the roll acceleration classification metric 286 does not produce a reliable distinction between moderate land events and embankment events. Therefore, when other classification metrics 282 and 284 are consistent in distinguishing the occurrence of moderate land events, the roll acceleration classification metric 286 is used as confirmation in the moderate land discrimination algorithm 90 (at box 276).

[0139] Soft soil identification

[0140] Figure 8 The soft soil discrimination algorithm 92 is shown. The soft soil discrimination algorithm 92 is used to determine whether a soft soil threshold should be used when applying an actuable constraint device (see [reference]). Figure 3 For left-side roll events (i.e., the vehicle rolls to the left or to the driver's side), the following is shown: Figure 8 Algorithm 92 for identifying soft soil. However, it should be understood that... Figure 8 The algorithm shown also applies to right-leaning events, the only difference being that the sign (+ / -) of the values ​​used in the metric is reversed. In other words, the metric used for right-leaning events will be the same as... Figure 8 The measurements shown are the same, but the signs of the corresponding axes for different measurement values ​​will be reversed, for example, positive becomes negative, and vice versa.

[0141] Land events refer to incidents where a vehicle laterally slides off the road and onto the ground. Land events are categorized into hard ground events, moderate ground events, and soft ground events, providing customized responses for the unique collision characteristics involved in these collision scenarios. Soft ground events can also be classified as events that do not produce the high-amplitude acceleration seen in moderate ground, hard ground, or tripping events. For example, a soft ground event may occur when a vehicle laterally slides off the road and onto a lawn or grass, resulting in a slower rollover development compared to hard ground and moderate ground events.

[0142] The soft soil discrimination algorithm 92 implements four different metrics to distinguish between slope events and embankment events. These four metrics are:

[0143] ·D_RATE to R_RATE_2

[0144] ·CCU_6Z vs R_ANGLE

[0145] ·CCU_1Y_AMA vs R_ANGLE

[0146] ·D_RATE vs R_ANGLE

[0147] In response to the satisfaction of both classification metric 310 fed to AND box 302 and latched soft soil box 304, a determination to use soft soil is made at box 300, i.e., a Boolean value of one, referred to herein as ON. Latched soft soil box 304 is latched, meaning that once AND box 306 is satisfied, latched soft soil 304 is ON and remains ON, even after the metric fed to AND box 306 is no longer ON. Classification metrics 312, 314, and 316 are fed to AND box 306. Classification metrics 310, 312, 314, and 316 are described in the following paragraphs.

[0148] The angular acceleration or roll acceleration to roll rate classification metric 310 uses D_RATE and R_RATE_2 to generate an output fed to box 302. As shown, the roll acceleration to roll rate classification metric 310 is ON when it is in the soft-triggered region defined by the solid line of the metric. The triggering region indicated by the classification metric 310 is the region (hard, medium, soft) that the metric first enters from the shaded area. In other words, the triggering region is latched in the classification metric 310. Therefore, if the metric first enters the soft-triggered region and then moves to the hard and / or medium-triggered region, the soft-triggered region indication remains ON, while the hard and / or medium-triggered region indication remains OFF.

[0149] Soft soil events (such as hard, dry soil and / or gravel) result in lower roll acceleration amplitudes compared to hard soil events due to the rapid provision of high sliding resistance on the soft soil surface. A roll acceleration-to-roll rate classification metric 310, configured to classify such roll accelerations in soft-trigger areas, provides a reliable discrimination for soft soil events. Furthermore, tests show that the roll acceleration-to-roll rate classification metric 310 produces a metric sufficiently distinguishable from embankment events to reliably differentiate them. Therefore, the roll acceleration-to-roll rate classification metric 310 can be used as a positive discrimination for soft soil events in the soft soil discrimination algorithm 92. The solid line in the roll acceleration-to-roll rate classification metric 310 represents the metric when a vehicle is experiencing a rollover event on soft soil.

[0150] The vertical acceleration to roll angle classification metric 312 uses CCU_6ZY_AMA and R_ANGLE to generate the output fed to box 306. As shown, the vertical acceleration to roll angle classification metric 312 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the vertical acceleration to roll angle classification metric 312 represents the metric when the vehicle is experiencing a soft soil rollover event. The vertical acceleration to roll angle classification metric 312 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the vertical acceleration to roll angle classification metric 312 does not produce a reliable distinction between soft soil events and embankment events. Therefore, when other classification metrics 314 and 316 are consistent in distinguishing the occurrence of soft soil events, the vertical acceleration to roll angle classification metric 312 is used as confirmation in the soft soil discrimination algorithm 92 (at box 306).

[0151] The lateral acceleration versus roll angle classification metric 314 uses CCU_1Y_AMA and R_ANGLE to generate the output fed into box 306. As shown, the lateral acceleration versus roll angle classification metric 314 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the lateral acceleration versus roll angle classification metric 314 represents the metric when the vehicle is experiencing a soft soil rollover event. The lateral acceleration versus roll angle classification metric 314 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that comparing roll acceleration with roll angle reliably distinguishes between soft soil events and embankment events. This is because the roll acceleration versus roll angle classification metric 314 produces a sufficiently distinguishable metric value to reliably distinguish between soft soil events and embankment events. Therefore, the roll acceleration versus roll angle classification metric 314 is used as a positive metric for soft soil events in the soft soil discrimination algorithm 92.

[0152] The angular acceleration or roll acceleration classification metric 316 uses D_RATE and R_ANGLE to generate the output fed to box 306. As shown, the roll acceleration classification metric 316 is ON when it is in the shaded trigger region defined by the dashed line. The solid line in the roll acceleration classification metric 316 represents the metric when the vehicle is experiencing a soft soil rollover event. The roll acceleration classification metric 316 is a latch-free metric, meaning it is ON only when it is in the trigger region. Tests show that the roll acceleration classification metric 316 does not produce a reliable distinction between soft soil events and embankment events. Therefore, when other classification metrics 312 and 314 are consistent in distinguishing the occurrence of soft soil events, the roll acceleration classification metric 316 is used as confirmation in the soft soil discrimination algorithm 92 (at box 306).

[0153] Rollover detection

[0154] Based on the above, it will be understood that... Figures 4 to 8 The enhanced discrimination algorithm shown and described herein can be implemented in vehicle safety systems to detect hill-end rollover events. Figure 4 ), embankment collapse incident ( Figure 5 Hard ground rollover incident ( Figure 6 ), moderate land rollover incident ( Figure 7 ) and soft soil rollover incidents ( Figure 8 The discrimination is performed between these events. As mentioned above, these discriminations are performed with a high degree of mutual exclusion, meaning that the discrimination of any one of these events highly accurately excludes the others. Therefore, the enhanced discrimination algorithm is able to identify and classify rollover events into one of these categories. Vehicle safety systems can then select thresholds for applying actuated restraint devices (such as airbags and seat belts) based on the classified rollover events, thereby improving the level of occupant protection provided by the vehicle safety system.

[0155] Based on the above description of the invention, those skilled in the art will understand that the described vehicle safety systems and methods implement algorithms capable of distinguishing between ramp rollover events and embankment rollover events, and also capable of distinguishing between land rollover events and embankment rollover events. Those skilled in the art will also recognize improvements, variations, and modifications to the disclosed systems and methods that fall within the spirit and scope of the invention. The appended claims are intended to cover such improvements, variations, and / or modifications.

Claims

1. A vehicle safety system, characterized in that, include: Actuable restraint devices used to help protect vehicle occupants; as well as A controller for controlling the actuation of the actuable restraint device in response to a vehicle rollover event; The controller is configured to execute a discrimination algorithm including at least one classification metric, which uses at least one of vehicle pitch rate (P_RATE) and vehicle roll acceleration (D_RATE) to distinguish at least one of slope rollover events and land rollover events from embankment rollover events. The at least one classification metric includes evaluating the vehicle roll acceleration (D_RATE) against the vehicle roll angle (R_ANGLE). The discrimination algorithm determines the classification of the vehicle rollover event, and the classification is one of a slope rollover event, a land rollover event, and an embankment rollover event. The controller is further configured to select an application threshold for applying the actuable restraint device, the application threshold corresponding to the classification of the vehicle rollover event.

2. The system as claimed in claim 1, wherein: The controller is configured to perform the classification metric of evaluating vehicle roll acceleration (D_RATE) against vehicle roll angle (R_ANGLE) to distinguish between slope rollover events and embankment rollover events.

3. The system as described in claim 2, wherein: The at least one classification metric further includes a classification metric that evaluates the vehicle pitch rate (P_RATE) against the vehicle roll angle (R_ANGLE); and The controller is configured to perform the classification metric of evaluating the vehicle pitch rate (P_RATE) against the vehicle roll angle (R_ANGLE) to distinguish between slope rollover events and embankment rollover events.

4. The system as claimed in claim 3, wherein: The at least one classification metric further includes a classification metric that evaluates the vehicle's lateral acceleration moving average (CCU_1Y_AMA) against the vehicle's roll angle (R_ANGLE); and The controller is configured to perform the classification metric of evaluating the moving average of the vehicle's lateral acceleration (CCU_1Y_AMA) against the vehicle's roll angle (R_ANGLE) to verify the distinction between slope rollover events and embankment rollover events.

5. The system as claimed in claim 3, wherein: The at least one classification metric further includes a classification metric that evaluates the vehicle's vertical acceleration moving average (CCU_6Z_AMA) in relation to the vehicle's roll angle (R_ANGLE); and The controller is configured to perform the classification metric of evaluating the vehicle's vertical acceleration moving average (CCU_6Z_AMA) against the vehicle's roll angle (R_ANGLE) to verify the distinction between slope rollover events and embankment rollover events.

6. The system of claim 1, wherein: The controller is configured to perform the classification metric of evaluating vehicle roll acceleration (D_RATE) against vehicle roll angle (R_ANGLE) to distinguish between land rollover events and embankment rollover events.

7. The system of claim 6, wherein, The discrimination algorithm further includes evaluating the vehicle roll acceleration (D_RATE) as a classification metric for the vehicle roll rate (R_RATE_2); and The controller is configured to perform the classification metric of evaluating vehicle roll acceleration (D_RATE) against vehicle roll rate (R_RATE_2) to distinguish between land rollover events and embankment rollover events.

8. The system of claim 7, wherein, The controller is configured to perform the classification metric of evaluating vehicle roll acceleration (D_RATE) against vehicle roll rate (R_RATE_2) to distinguish between hard soil rollover events, moderate soil rollover events, and soft soil rollover events.

9. The system of claim 7, wherein: The at least one classification metric further includes a classification metric that evaluates the vehicle's lateral acceleration moving average (CCU_1Y_AMA) against the vehicle's roll angle (R_ANGLE); and The controller is configured to perform the classification metric of evaluating the moving average of the vehicle's lateral acceleration (CCU_1Y_AMA) against the vehicle's roll angle (R_ANGLE) to distinguish between land rollover events and embankment rollover events.

10. The system of claim 7, wherein: The at least one classification metric further includes a classification metric that evaluates the vehicle's vertical acceleration moving average (CCU_6Z_AMA) in relation to the vehicle's roll angle (R_ANGLE); and The controller is configured to perform the classification metric of evaluating the vehicle's vertical acceleration moving average (CCU_6Z_AMA) against the vehicle's roll angle (R_ANGLE) to verify the distinction between land rollover events and embankment rollover events.

11. The system of claim 1, wherein, The controller is configured to execute an application threshold metric for determining whether to actuate the actuable restraint device, wherein the application threshold metric evaluates vehicle roll acceleration (D_RATE) versus vehicle roll angle (R_ANGLE), and wherein the controller is configured to apply the actuable restraint device in response to the application threshold metric exceeding the application threshold.

12. The system of claim 1, further comprising: An accelerometer used to sense vehicle lateral acceleration and provide a signal indicating the sensed vehicle lateral acceleration (CCU_1Y); An accelerometer used to sense the vertical acceleration of a vehicle and provide a signal indicating the sensed vertical acceleration of the vehicle (CCU_6Z); as well as A roll rate sensor for sensing vehicle roll rate values ​​and providing a signal indicating the sensed vehicle roll rate value (CCU_4R); The controller is configured to perform vehicle measurement calculations to: The vehicle lateral acceleration moving average (CCU_1Y_AMA) is determined based on the signal of the vehicle lateral acceleration (CCU_1Y) sensed by the indicated signal; The moving average value of vehicle vertical acceleration (CCU_6Z_AMA) is determined based on the signal of the vehicle vertical acceleration (CCU_6Z) sensed by the indicated signal; The vehicle roll acceleration (D_RATE) is determined based on the signal of the sensed vehicle roll rate value (CCU_4R); and The vehicle roll angle (R_ANGLE) is determined based on the signal of the vehicle roll rate value (CCU_4R) sensed by the indicated signal.

13. The system of claim 1, further comprising a pitch rate sensor for sensing a vehicle pitch rate value (CCU_5P) and providing a signal indicating the sensed vehicle pitch rate, wherein, The controller is configured to perform vehicle metric calculations to determine the vehicle pitch rate (P_RATE) based on the signal of the vehicle pitch rate sensed by the indication.

14. The system of claim 1, wherein, The actuable restraint device includes at least one of a seatbelt anchor pretensioner, a seatbelt retractor pretensioner, a curtain airbag, a chest airbag, a side airbag, an emergency notification, a door unlocking command, and a high-voltage power transmission disconnection command.

15. The system of claim 1, wherein, The controller is further configured to issue at least one of an emergency notification, a door unlocking command, and a high-voltage power transmission disconnection command in response to actuation of the actuable restraint device.

Images