JP3876358B2 - Fully closed output learning device for accelerator pedal sensor - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a fully closed output learning device for an accelerator pedal sensor that learns an output (fully closed output) at an accelerator fully closed position of an accelerator pedal sensor that outputs a signal corresponding to an accelerator opening.
[0002]
[Prior art]
Conventionally, a technique for improving the accelerator opening detection accuracy by learning the output of the accelerator pedal sensor at the accelerator fully closed position is known (see Japanese Patent Laid-Open No. 10-176555).
[0003]
[Problems to be solved by the invention]
With the accelerator pedal sensor that is placed directly on the accelerator pedal, a position that returns more than the normal accelerator fully closed position by applying force in the return direction to the accelerator pedal (an act of lifting the pedal with the instep) If the sensor output at the fully closed position is learned at this time, the accelerator opening is detected based on this position, and it is recognized that the accelerator pedal is slightly depressed even in the idle state. As a result, there is a problem that it becomes a high idle state.
[0004]
In addition, in the case of an accelerator pedal sensor of a type that is pulled apart by an accelerator cable separately from the accelerator pedal, even if the accelerator pedal is lifted, the accelerator cable only sags and there is almost no force in the sensor output decreasing direction. With the progress of simplification of parts configuration for the purpose of reducing the cost of parts, accelerator pedal sensors of the type that directly read the rotation angle of the accelerator pedal are becoming mainstream, especially in this type, sensor output by lifting the accelerator pedal Prone to decline.
[0005]
Therefore, for example, the output of the accelerator pedal sensor is monitored every 10 msec to calculate a moving average of a predetermined number of times, and the state where the moving average value is smaller than the learning value of the fully closed output until the previous time continues for a predetermined learning update period or longer. When the learning value of the fully closed output is updated by the moving average value, the learning update rate is slowed by setting the learning update period to be long, and erroneous learning by lifting the accelerator pedal However, if the setting is made in this way, a long learning time is required when learning for the first time at the time of shipment from the factory, resulting in a problem that productivity in the factory deteriorates.
[0006]
Therefore, an object of the present invention is to achieve both the prevention of mislearning due to temporary factors such as lifting of the accelerator pedal and the improvement of factory productivity by shortening the initial learning at the time of factory shipment. .
[0007]
[Means for Solving the Problems]
For this reason, the invention of claim 1 is a fully closed output learning device for an accelerator pedal sensor that learns an output at an accelerator fully closed position of an accelerator pedal sensor that outputs a signal corresponding to the accelerator opening, and the sensor output is If the learning value of the fully closed output is updated based on the sensor output when a state smaller than the learning value of the fully closed output continues for a predetermined learning update cycle or more, the learning update cycle can be switched to a plurality of different values. characterized in that it comprises a Do learning update cycle switching means.
Further, when updating the learning value of the fully-closed output, an update width limiting means for limiting the update width per time to a limit value or less is provided, and the limit value of the update width is changed simultaneously with the switching of the learning update cycle. An update width switching means is provided.
[0008]
In addition, two accelerator pedal sensors are provided in one accelerator pedal, and fault diagnosis means for diagnosing a fault when the output difference between the two accelerator pedal sensors exceeds a predetermined threshold value, The limit value is set to a value equal to or less than the predetermined threshold value.
[0009]
According to a second aspect of the present invention, the learning update cycle is switched according to the accumulated time of learning of the fully-closed output after the ignition is turned on, and when the accumulated time exceeds a predetermined time, the learning update cycle is lengthened. It is characterized by doing.
[0010]
【The invention's effect】
According to the first aspect of the present invention, the learning update cycle can be switched, so that the learning update cycle can be shortened and the learning update cycle can be shortened as in the factory shipment or the learning update cycle can be lengthened. Except for temporary factors such as lifting the accelerator pedal, it is possible to satisfy all the requirements for learning only the deterioration of the sensor system accurately.
Furthermore, by changing the limit value of the update width at the same time as switching the learning update cycle, the limit value of the update width is increased (the limit is reduced) to further increase the learning speed, or the limit value of the update width It is possible to reduce the learning speed (by increasing the limit).
[0011]
Furthermore, by setting the limit value of the update width according to the failure diagnosis threshold value of the dual system sensor, erroneous diagnosis in failure diagnosis can be prevented. In other words, when a fault is diagnosed when the output difference of the dual system sensor exceeds a predetermined threshold value, if one sensor is updated with an update width greater than the threshold value, In this case, it may be erroneously diagnosed as a sensor system failure, but this can be prevented.
[0012]
According to the invention of claim 2 , the learning update cycle is switched according to the accumulated time of learning of the fully closed output after the ignition is turned on, and when the accumulated time becomes a predetermined time or longer, the learning update cycle is lengthened. By doing this, it is preferable that a chance for rapid learning can always be created after the ignition is turned on when it is desired to complete learning early in parts replacement or the like at the time of factory shipment.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a schematic view of an electric throttle system showing an embodiment of the present invention.
The accelerator pedal 1 is biased to the accelerator fully closed position by a return spring 2.
[0014]
The electric throttle valve 3 is urged to an initial position (a neutral position on the relatively low opening side) by a return spring (torsion coil spring) 4 and is driven by a throttle control motor 5.
The engine control unit (ECU) 6 includes a CPU 7, an A / D converter 8, a motor drive circuit 9, and the like, and signals from various sensors are input to the CPU 7 via the A / D converter 8.
[0015]
The accelerator pedal sensors APS1 and APS2 are potentiometer-type sensors that generate a voltage signal corresponding to the opening degree (depression amount) of the accelerator pedal 1, and are independently provided with two systems, each of which is an A / D converter. 8 is input to the CPU 7.
The throttle position sensors TPS1 and TPS2 are potentiometer type sensors that generate a voltage signal corresponding to the opening degree of the electric throttle valve 3, and two independent systems are provided, each of which is an A / D converter. 8 is input to the CPU 7.
[0016]
In addition, signals such as engine speed, water temperature, vehicle speed, and gear position are also input from the other sensor group 100 as the operating states of the engine and the vehicle.
In the CPU 7, the target throttle opening is set according to the opening of the accelerator pedal 1 detected by the accelerator pedal sensors APS1 and APS2, and further the operating state of the engine and the vehicle, and the opening of the electric throttle valve 3 is set accordingly. Control. Specifically, the actual throttle opening detected by the throttle position sensors TPS1 and TPS2 matches the target throttle opening while controlling the supply of power (BATT) to the motor drive circuit 9 via the motor relay 10. Thus, by controlling the duty of the motor drive circuit 9, the current to the throttle control motor 5 is controlled and the opening degree of the electric throttle valve 3 is controlled.
[0017]
Here, when the accelerator opening is detected by the output voltage of the accelerator pedal sensors APS1 and APS2, the output at the accelerator fully closed position (fully closed output) is learned for each sensor, and the actual output is fully closed. The accelerator opening is detected based on a value obtained by subtracting the learning value of the output (full closing learning value).
Next, learning of the fully closed output of the accelerator pedal sensor (fully closed learning) will be described.
[0018]
FIG. 2 is a flowchart of an initialization routine started when the ignition is turned on (OFF → ON), and initializes both the fully closed learning accumulated time counter TAPS and the learning update period measurement counter CAPSCL (TAPS = 0, CAPSCL = 0).
FIG. 3 is a flowchart of a fully-closed learning routine executed every certain time (for example, 10 msec) after the ignition is turned on.
[0019]
In step 1 (denoted as S1 in the figure, the same applies hereinafter), the fully closed learning accumulated time counter TAPS (counted in steps 12 and 13 described later) is compared with a predetermined value APSINI, and the fully closed after the ignition is turned on. It is determined whether or not the learning accumulated time (TAPS) is within a predetermined time (APSINI).
If it is within the predetermined time, rapid learning is necessary. Therefore, the process proceeds to step 2, and a relatively short period TAPSCLH (for example, 1.5 sec) is set as the learning update period TAPSCL (TAPSCL = TAPSCLH) and once per time set to a relatively large 120mV within the update width of the limit value LIM fault diagnosis threshold.
[0020]
If the predetermined time has been reached, rapid learning has been completed, and thus the process proceeds to step 3 where a relatively long period TAPSCLL (for example, 60 sec) is set as the learning update period TAPSCL (TAPSCL = TAPSCLL) once. The per-update width limit value LIM is set to a relatively small 30 mV. Steps 1 to 3 correspond to learning update cycle switching means and update width switching means .
In step 4, the moving average value AVAPS of the accelerator pedal sensor output is compared with the current fully closed learning value APSCL.
[0021]
The moving average value AVAPS of the accelerator pedal sensor output is obtained as a simple average value of the sampling values of the sensor output within the learning update period TAPSCL, or the moving average value AVAPS up to the previous time and the current sensor output APS as in the following equation: It is calculated as a weighted average value.
AVAPS = (1-G) × AVAPS + G × APS
G is a weighting constant, and 0 <G <1.
[0022]
If the moving average value AVAPS <the fully closed learning value APSCL, it is necessary to learn, so the process proceeds to step 5 and the learning update period measurement counter CAPSCL is counted up (CAPSCL = CAPSCL + 1).
If the moving average value AVAPS> the fully closed learning value APSCL, there is no need for learning, so the process proceeds to step 6 and the learning update period measurement counter CAPSCL is reset (CAPSCL = 0). Therefore, the value of the counter CAPSCL represents the time during which the moving average value of the accelerator pedal sensor output is smaller than the fully closed learning value.
[0023]
In step 7, the learning update period measurement counter CAPSCL is compared with the learning update period TAPSCL determined in step 2 or 3, and whether or not the learning update period TAPSCL has been reached, in other words, the moving average of the accelerator pedal sensor output. It is determined whether or not the duration (CAPSCL) in a state where the value is smaller than the fully closed learning value has reached a predetermined time (learning update period TAPSCL).
[0024]
When the learning update period TAPSCL has been reached, since the update of the fully closed learning value is started, the process proceeds to step 8 where the moving average value AVAPS is lower than the update width limit value LIM with respect to the fully closed learning value APSCL. It is determined whether (AVAPS <APSCL− LIM ).
If YES in this determination, that is, if it is lower than the limit value LIM of the update width, the process proceeds to step 9 where the fully closed learning value APSCL is updated so as to decrease by LIM from the current value (APSCL = APSCL− LIM). ). Steps 8 and 9 correspond to update width limiting means.
[0025]
NO Conversely, i.e. if: LIM drop allowance is the limiting value of the update width, the process proceeds to step 10 to update replaces the fully closed learning value APSCL in this sensor output (moving average value) AVAPS (APSCL = AVAPS).
After the learning value is updated, the process proceeds to step 11 to reset the learning update period measurement counter CAPSCL (CAPSCL = 0).
[0026]
In step 12, it is determined whether or not the fully closed learning condition is satisfied and TAPSCL = TAPSCLH (short learning update cycle) in order to measure the fully closed learning accumulated time used for switching of the learning update cycle.
Here, the fully closed learning condition is (1) that there is no problem in diagnosis of the sensor itself, (2) that the battery voltage is a predetermined value (for example, 8V) or more, and (3) in the dual system sensor, both sensors are It shall be established when all of the idle determination is satisfied.
[0027]
If this determination is YES, that is, if the fully closed learning condition is satisfied and TAPSCL = TAPSCLH (short learning update cycle), the routine proceeds to step 13 where the fully closed learning accumulated time TAPS is counted up (TAPS = TAPS + 1). ). On the other hand, if NO, the routine is terminated as it is.
As described above, while the fully closed learning accumulated time TAPS after the ignition is ON is within the predetermined time (APSINI), the fully closed learning value is updated at a relatively short learning update period TAPSCLH (for example, 1.5 sec), and 1 The limit value LIM of the update width per time is set to a relatively large 120 mV.
[0028]
Further, the accelerator pedal sensor has two sensors for failsafe of the system, and constantly monitors the output difference between the two sensors, and diagnoses a failure when it exceeds a predetermined threshold value. In this case, since the fully closed learning values of the two sensors are not updated at the same time, the update width per sensor needs to be a small value that does not affect the fault diagnosis. It is set to 120 mv.
[0029]
In this case, in the initial learning at the time of factory shipment, if the initial value of the preset fully closed learning value is 600 mV, for example, when the limit value of the update width per one time is 120 mV, the fully closed learning is terminated. The fully closed learning value needs to be updated about 5 times. If the fully closed learning value is updated at a relatively short learning update period TAPSCLH (for example, 1.5 sec) at this time, 1.5 sec × 5 = 7. Fully closed learning can be completed in 5 seconds, and productivity can be improved.
[0030]
Thereafter , the fully closed learning value is updated at a relatively long learning update period TAPSCLL (for example, 60 sec), and the limit value LIM of the update width per time is set to a relatively small 30 mV.
Therefore, since the fully closed learning value is updated at a relatively long learning update period TAPSCLL (for example, 60 sec), even when a person accidentally applies a force in the pedal lifting direction, the fully closed output is increased. There will be no false learning. On the other hand, learning of the fully closed output is intended to guarantee very gradual deterioration such as pedal wear, so that the purpose can be sufficiently achieved even by learning at a gradual speed.
[0031]
Further, the limit value LIM of the update width per time at this time is set to a comparatively small 30 mV, the limit value LIM of the update width is reduced, and the update speed is made more gradual .
Since the purpose of the fully closed learning of the accelerator pedal sensor is to guarantee a very gradual deterioration such as pedal wear, an update width of about 30 mV is sufficient per time, and thus the update width is limited in this way. As a result, safety against mislearning caused by lifting the pedal or the like can be further improved.
[0032]
On the other hand, in the initial learning at the time of factory shipment, if the update width per one time is set to 30 mV, it is necessary to turn the ignition key on and off 10 to 20 times until the fully closed learning is completed, and production at the factory. The sex will be reduced. Accordingly, at this time, it is preferable to set the maximum value to about 120 mV within a range determined by the failure diagnosis threshold. As described above, it is possible to prevent erroneous learning of the fully closed output while satisfying the productivity in the factory.
[Brief description of the drawings]
FIG. 1 is a schematic diagram of an electric throttle system showing an embodiment of the present invention. FIG. 2 is a flowchart of an initialization routine when an ignition is turned on. FIG. 3 is a flowchart of a fully-closed learning routine after the ignition is turned on. ]
1 Accelerator pedal
2 Return spring
3 Electric throttle valve
5 Throttle control motor
6 Engine control unit (ECU)
7 CPU
APS1, APS2 Accelerator pedal sensor
TPS1, TPS2 Throttle position sensor

Claims (2)

A fully closed output learning device for an accelerator pedal sensor that learns an output at an accelerator fully closed position of an accelerator pedal sensor that outputs a signal corresponding to an accelerator opening,
When the sensor output is smaller than the learning value of the fully-closed output and continues for a predetermined learning update period or longer, the learning value of the fully-closed output is updated based on the sensor output.
Learning update cycle switching means capable of switching the learning update cycle to different values,
Further, when updating the learning value of the fully-closed output, an update width limiting means for limiting the update width per time to a limit value or less is provided, and the limit value of the update width is changed simultaneously with the switching of the learning update cycle. with an update width switching means for,
In addition, two accelerator pedal sensors are provided in one accelerator pedal, and a failure diagnosis means for diagnosing a failure when an output difference between the two accelerator pedal sensors exceeds a predetermined threshold value,
The fully closed output learning device for an accelerator pedal sensor , wherein the limit value of the update width is set to a value equal to or less than the predetermined threshold value . Switching of the learning update period is performed in accordance with the accumulated time of learning of full-closed position force after the ignition ON, if the accumulated time reaches a predetermined time, wherein, characterized in that lengthening the learning update period The fully closed output learning device for an accelerator pedal sensor according to Item 1 .

Images