JP4385064B2 - Control object recognition device - Google Patents

Description

  The present invention detects an object based on a reflected wave having a reflection level equal to or higher than a detection threshold among the reflected waves transmitted by the transmitting means and received by the receiving means, and controls the vehicle based on the detection result of the object. The present invention relates to a control object recognition apparatus that recognizes an object.

  When the obstacle detected by the radar device is temporarily not detected, the information of the obstacle after that is estimated for a predetermined time based on the information of the previous obstacle stored in the storage means. However, it is known from Patent Document 1 below to determine the possibility of the own vehicle contacting an obstacle based on the estimated information.

Further, in the second embodiment of Japanese Patent Application No. 2006-211839 (filed on August 3, 2006) relating to the application of the present applicant, there is a stop of a low gate and a preceding vehicle that can be passed through in front of the own vehicle. In the case of automatic deceleration or warning to avoid collision with an object, when a stopped object as a target is no longer detected, extrapolation processing (future position of the target that is no longer detected) is performed five times thereafter. Etc., and processing as if it was detected), so that even if the target becomes temporarily undetectable, automatic deceleration and warning to avoid collisions can be performed without hindrance Has been proposed.
JP-A-6-174847

  Next, based on FIGS. 7 to 11, problems of the one proposed in Japanese Patent Application No. 2006-211839 will be described. In FIGS. 7 to 11, ● represents actual data, and ◯ represents extrapolated data.

  FIG. 7 (A) shows the state of time 1, and it is assumed that there is a low gate that can pass through, which is a stopped object, ahead of the host vehicle, and a stopped vehicle that is a stopped object exists in front of the gate. In this case, it is desirable that only the collision avoidance control for the stopped vehicle is performed because the own vehicle is unlikely to collide with a gate that can pass through.

  Transmits electromagnetic waves forward from the radar device installed in the vehicle, compares the reflection level of the reflected wave from the gate and the stopped vehicle with the detection threshold, and automatically brakes or warns a stopped object whose reflection level is higher than the detection threshold Target of collision avoidance control such as. In the case of FIG. 7A, the reflection level of the gate near the host vehicle is equal to or greater than the detection threshold value, but the distance is outside the operation area of the collision avoidance control. Is not done. Further, since the stopped vehicle is far away, the reflection level of the reflected wave is less than the detection threshold, and naturally the collision avoidance control for the stopped vehicle is not performed.

  At time 4 in FIG. 7B, in addition to the reflection level of the reflected wave of the gate, the reflection level of the reflected wave of the stopped vehicle is equal to or greater than the detection threshold, but those distances are outside the operation area of the collision avoidance control. Therefore, collision avoidance control such as automatic braking and warning for the gate and the stopped vehicle is not performed.

  At time 5 in FIG. 8C, the reflection level of the reflected wave at the gate decreases because the vehicle approaches the gate further, and most of the electromagnetic waves of the radar device pass through the inside of the gate. Since the reflection level is still higher than the detection threshold, the state at time 4 in FIG. 7B is continued.

  At time 6 in FIG. 8D, the vehicle approaches the gate further, the reflection level of the reflected wave becomes less than the detection threshold value, and the gate is in a lost (undetectable) state. A first extrapolation process is performed. Also at this time, since the distance between the extrapolated gate and the stopped vehicle is outside the operation area of the collision avoidance control, the collision avoidance control such as automatic braking and warning for the gate and the stopped vehicle is not performed.

  At time 7 in FIG. 9E, since the gate is still in a lost state, a second extrapolation process for predicting the position of the gate is performed. Also at this time, since the distance between the extrapolated gate and the stopped vehicle is outside the operation area of the collision avoidance control, the collision avoidance control such as automatic braking and warning for the gate and the stopped vehicle is not performed.

  At time 8 in FIG. 9F, since the gate is still in the lost state, the third extrapolation process for predicting the position of the gate is performed. In this case, since the distance of the extrapolated gate falls within the operation area of the collision avoidance control, collision avoidance control such as automatic braking and warning is executed in order to avoid collision with the gate. However, the collision avoidance control is unnecessary because the gate is originally a stop object that can pass through, and unnecessary collision avoidance control is executed.

  At this time, the collision avoidance control is not an automatic braking but an alarm, and the driver ignores the alarm and advances his / her own vehicle. When time 11 in FIG. 10G is reached, the stopped vehicle is within the operation area of the collision avoidance control. Therefore, collision avoidance control such as automatic braking and warning for the stopped vehicle is executed.

  At time 12 in FIG. 10 (H), the stopped vehicle is still detected within the operation area of the collision avoidance control, so the collision avoidance control is continued as it is.

  At time 13 in FIG. 11 (I), since the own vehicle further approaches the stopped vehicle, the stopped vehicle is removed from the detection area of the radar device to the side, and the reflection level becomes less than the detection threshold value. Processing is performed. Then, the collision avoidance control is continued as it is based on the first extrapolation data.

  At time 14 in FIG. 11J, a second extrapolation process is performed, and the collision avoidance control is continued as it is based on the second extrapolation data.

  As described above, in the conventional apparatus, the extrapolation process is continuously performed up to five times while the obstacle is lost. Therefore, if extrapolation processing is continuously performed up to 5 times on a stationary object such as a low gate that can be dive-through, excessive collision avoidance control is executed at time 8 in FIG. 9F. This could cause the driver to feel uncomfortable.

  The present invention has been made in view of the above-described circumstances, and an object thereof is to prevent unnecessary vehicle control from being performed on a stationary object that does not become an obstacle to traveling of the host vehicle.

  In order to achieve the above object, according to the first aspect of the present invention, a transmission unit that transmits an electromagnetic wave toward a predetermined detection area at a predetermined time interval, and an electromagnetic wave transmitted by the transmission unit is reflected on an object. Receiving means for receiving the reflected wave, an object detecting means for detecting an object based on a reflected wave having a reflection level equal to or higher than a detection threshold among the reflected waves received by the receiving means, and detection by the object detecting means A relative relationship calculating means for calculating the relative relationship between the vehicle and the object based on the result, a relative relationship predicting means for predicting the current relative relationship based on the relative relationship previously calculated by the relative relationship calculating means, and the relative By comparing the relative relationship predicted this time by the relationship prediction unit with the relative relationship calculated this time by the relative relationship calculation unit, the object detected last time and the object detected this time Identity determining means for determining the same object, and control target object recognizing means for recognizing that the object that has been determined to be the same object by the identity determining means reaches a predetermined number of times as a control target object When the object that has been recognized as the control target by the control target object recognition unit is no longer recognized as the control target this time, the object predicted by the relative relationship prediction unit is actually detected. Extrapolating means for extrapolating a predetermined number of times as if the object is an object, stationary object determining means for determining whether or not the object recognized as the control object is a stationary object, and the control object Reflection level prediction means for predicting the reflection level at the next detection from the reflection level at the previous detection of the object recognized as the current detection time and the reflection level at the current detection, and the extrapolation means. Are those, and the predicted reflection level is of less than the detection threshold, the control target recognition apparatus characterized by reducing the limit of the number of times of performing the extrapolation is proposed.

  According to a second aspect of the present invention, in addition to the configuration of the first aspect, the extrapolation means reduces the limit of the number of extrapolations according to an increase in the vehicle speed of the host vehicle. A control object recognition device is proposed.

  According to the invention described in claim 3, the transmitting means for transmitting the electromagnetic wave toward the predetermined detection area at predetermined time intervals, and the reflected wave in which the electromagnetic wave transmitted by the transmitting means is reflected by the object is received. A receiving means; an object detecting means for detecting an object based on a reflected wave having a reflection level equal to or higher than a detection threshold value among the reflected waves received by the receiving means; and the vehicle based on a detection result of the object detecting means; A relative relationship calculating means for calculating a relative relationship between objects, a relative relationship predicting means for predicting a current relative relationship based on the relative relationship previously calculated by the relative relationship calculating means, and a current relationship predicted by the relative relationship predicting means. The relative relationship calculated this time is compared with the relative relationship calculated this time by the relative relationship calculation means, so that the object detected last time and the object detected this time are the same object. Identity determining means to be determined, control target object recognizing means for recognizing an object that has been determined to be the same object by the identity determining means as a control target object, and the control target object When an object that has been recognized as a control target until the previous time by the recognition unit is no longer recognized as a control target this time, the object predicted by the relative relationship prediction unit is regarded as actually detected and predetermined. Extrapolation means for extrapolating as many times as possible, and controlling the vehicle based on the relative relationship, and whether the object recognized as the control target object is actually detected or extrapolated Vehicle control means for changing the content of the control in accordance with the control object, stationary object determination means for determining whether or not the object recognized as the control object is a stationary object, and recognition as the control object. Reflection level prediction means for predicting the reflection level at the next detection from the reflection level at the previous detection of the detected object and the reflection level at the current detection, and the extrapolation means, the object is a stationary object, and When the predicted reflection level is less than the detection threshold, an object recognized as a control target this time is set as an extrapolated object, and the vehicle control means actually detects vehicle control on the extrapolated object. There is proposed a control object recognition apparatus characterized in that the control object is weaker than the vehicle control for the object that is being performed.

  According to the fourth aspect of the present invention, the transmitting unit that transmits the electromagnetic wave toward the predetermined detection area at predetermined time intervals, and the reflected wave in which the electromagnetic wave transmitted by the transmitting unit is reflected by the object is received. A receiving means; an object detecting means for detecting an object based on a reflected wave having a reflection level equal to or higher than a detection threshold value among the reflected waves received by the receiving means; and the vehicle based on a detection result of the object detecting means; A relative relationship calculating means for calculating a relative relationship between objects, a relative relationship predicting means for predicting a current relative relationship based on the relative relationship previously calculated by the relative relationship calculating means, and a current relationship predicted by the relative relationship predicting means. The relative relationship calculated this time is compared with the relative relationship calculated this time by the relative relationship calculation means, so that the object detected last time and the object detected this time are the same object. Identity determining means to be determined, control target object recognizing means for recognizing an object that has been determined to be the same object by the identity determining means as a control target object, and the control target object When an object that has been recognized as a control target until the previous time by the recognition unit is no longer recognized as a control target this time, the object predicted by the relative relationship prediction unit is regarded as actually detected and is predetermined. Extrapolation means for extrapolating as many times as possible, and controlling the vehicle based on the relative relationship, and whether the object recognized as the control target object is actually detected or extrapolated Vehicle control means for changing the timing for performing the control in accordance with the object, stationary object determination means for determining whether or not the object recognized as the control object is a stationary object, and the control object Reflection level predicting means for predicting the reflection level at the next detection from the reflection level at the previous detection of the object recognized as this time and the reflection level at the current detection, and the extrapolation means, the object is a stationary object When the predicted reflection level is less than the detection threshold, an object recognized as a control target this time is set as an extrapolated object, and the vehicle control means performs vehicle control on the extrapolated object. A control object recognition apparatus is proposed that is less likely to be performed than vehicle control for an object that is actually detected.

  According to the configuration of claim 1, when the object that has been recognized as the control target until the previous time by the control target object recognition unit is no longer recognized as the control target this time, the object predicted by the relative relationship prediction unit When the extrapolation means extrapolates a predetermined number of times, assuming that the object is actually detected, the object is a stationary object, and the reflection level at the previous detection of the object and the reflection level at the current detection When the predicted reflection level at the next detection is less than the detection threshold, the extrapolation means reduces the limit of the number of extrapolations, so the reflection level drops rapidly and is estimated not to be controlled The number of extrapolations can be reduced, unnecessary vehicle control for the object can be prevented, and the driver's discomfort can be eliminated.

  According to the second aspect of the present invention, the extrapolation means reduces the limit of the number of times of extrapolation according to the increase in the vehicle speed of the own vehicle, so that extrapolation data during extrapolation when the vehicle speed of the own vehicle is high. Even if the vehicle suddenly approaches the vehicle, unnecessary vehicle control can be prevented.

  According to the third aspect of the present invention, when the object that has been recognized as the control target until the previous time by the control target object recognition means is no longer recognized as the control target this time, it is predicted by the relative relationship prediction means. When the extrapolation means extrapolates the object up to a predetermined number of times, assuming that the object is actually detected, the object is a stationary object, and the reflection level at the previous detection of the object and the reflection level at the current detection If the reflection level predicted at the next detection is less than the detection threshold value, that is, if the reflection level of the object has fallen sharply and the object is not a control target, the extrapolation means is the control target this time. The object recognized as the extrapolated object is handled as extrapolated data, and the data of the object recognized as the current control target, which should be actual data, is handled as extrapolated data. From excessive vehicle control is performed also weaker than the vehicle control with respect to the object being sensed can be prevented.

  According to the fourth aspect of the present invention, when the object that has been recognized as the control target until the previous time by the control target object recognition unit is no longer recognized as the control target this time, it is predicted by the relative relationship prediction unit. When the extrapolation means extrapolates the object up to a predetermined number of times, assuming that the object is actually detected, the object is a stationary object, and the reflection level at the previous detection of the object and the reflection level at the current detection If the reflection level predicted at the next detection is less than the detection threshold value, that is, if the reflection level of the object has fallen sharply and the object is not a control target, the extrapolation means is the control target this time. The object recognized as the extrapolated object is handled as extrapolated data, and the data of the object recognized as the current control target, which should be actual data, is handled as extrapolated data. Is delayed than the vehicle control with respect to the object being sensed can be prevented from excessive vehicle control is performed.

  Embodiments of the present invention will be described below with reference to the accompanying drawings.

  1 to 11 show an embodiment of the present invention and a conventional example. FIG. 1 is a block diagram of an electronic control unit of a controlled object recognition apparatus. FIG. 2 is a time chart for explaining the operation of the conventional example. Is a time chart for explaining the operation of the embodiment, FIG. 4 is a diagram showing a main flowchart, FIG. 5 is a diagram showing a previous target and takeover sub-flowchart, FIG. 6 is a diagram showing a previous target processing subflowchart, and FIG. FIG. 8 is an operation explanatory diagram of the conventional example (time 5 and time 6), FIG. 9 is an operation explanatory diagram of the conventional example (time 7 and time 8), and FIG. FIG. 11 is an operation explanatory diagram (time 13 and time 14) of the conventional example, and FIG. 11 is an operation explanatory diagram (time 13 and time 14) of the conventional example.

  As shown in FIG. 1, the control target recognition apparatus according to the present embodiment includes a radar device R including a transmission unit Ra that transmits an electromagnetic wave and a reception unit Rb that receives a reflected wave in which the electromagnetic wave is reflected by an object. When the vehicle follows the preceding vehicle detected in step 2, the distance between the vehicle and the preceding vehicle is less than the predetermined value and the possibility of a collision increases. When a collision is unavoidable, a collision prevention system S that brakes the vehicle by automatic braking and an alarm unit W that prompts the driver to spontaneously brake are operated. The electronic control unit U connected to the radar device R, the collision prevention system S, and the alarm unit W includes an object detection unit M1, a relative relationship calculation unit M2, a relative relationship prediction unit M3, an identity determination unit M4, and a control. Target object recognition means M5, extrapolation means M6, stationary object determination means M7, reflection level prediction means M8, and vehicle control means M9 are provided.

  The object detection means M1 compares the reflection level of the reflected wave received by the reception means Rb of the radar device R with a predetermined detection threshold, and detects only an object whose reflection level is equal to or higher than the detection threshold. The relative relationship calculation means M2 calculates relative relationships such as the relative position and relative speed of the vehicle and the object based on the detection result of the object detection means M1. The relative relationship predicting means M3 predicts the next relative relationship from the current relative relationship calculated by the relative relationship calculating means M2. The identity determination means M4 determines the identity between the previously detected object and the currently detected object by comparing the actual relative relationship calculated this time with the estimated relative relationship predicted this time. The control target object recognition unit M5 finally controls only the objects that have been determined to be the same object by the identity determination unit M4 for a predetermined number of times in order to avoid misidentifying different objects as the same object. It is determined that the object is a target (automatic brake or alarm target).

  The extrapolation means M6, for example, when the object that the control target object recognition means M5 has recognized as the control target until the previous time is no longer recognized as the control target this time, for example, the object detection means M1 loses the control target. Assuming that the lost control object is continuously recognized for a predetermined number of times (5 in the embodiment), extrapolated data is used as the past control object data (relative position and relative speed). ) To extrapolate. That is, the data that is no longer recognized is assumed to be recognized and output as extrapolated data. Thereby, even if the control target cannot be temporarily recognized due to the influence of noise or the like, the control up to that point can be continued without interruption. Of course, if the control object that is no longer recognized is recognized again and the identity of the control object recognized again and the extrapolated data is confirmed, the extrapolated data is again taken over by the actual data.

  The normal maximum number of extrapolation processes by the extrapolation means M6 is 5 times. If the extrapolation process is not carried over to the actual data even after the extrapolation process has been performed 5 times, for example, the control object changes the course and the own vehicle Extraneous data disappears after 5 extrapolation processes. The maximum number of extrapolation processes is normally five. However, when the following first and second conditions are satisfied, the maximum number of extrapolation processes is substantially two times (temporary extrapolation data described later is used). Including 3 times).

  The first condition is that the controlled object is a stationary object. The stationary object determination means 6 can determine that the controlled object is a stationary object when the controlled object approaches the own vehicle at a speed close to the speed of the own vehicle. The second condition relates to the reflection level to be controlled. The reflection level prediction means M8 predicts the next reflection level from the previous reflection level to be controlled and the current reflection level. The extrapolation means M6 reduces the maximum extrapolation number from 5 to 2 when the controlled object is a stationary object and the predicted next reflection level is less than the detection threshold. At this time, when the predicted next reflection level is less than the detection threshold, even if the reflection level of the current control target is equal to or higher than the threshold, the data of the current control target is extrapolated data (temporary extrapolation data). Handled. Therefore, when the extrapolation data is regarded as extrapolation data, the maximum extrapolation count is reduced from five to three (of which one is provisional extrapolation data).

  When the actual data or extrapolated data to be controlled enters the system operation area in front of the own vehicle, the vehicle control means M9 operates the collision prevention system S and the alarm unit W so that the own vehicle does not collide with the controlled object. Control.

  In other words, in the conventional example, when a stationary object is lost, extrapolation data is inserted up to five times. However, in the present embodiment, when a stationary object such as a low gate that can be submerged is lost, The number of insertions of extrapolated data is limited to a maximum of three times (one of which is temporary extrapolated data) to prevent unnecessary automatic braking and warnings from being executed and causing the driver to feel uncomfortable.

  FIG. 2 is a time chart showing a conventional example with a maximum number of extrapolation times of 5 and corresponds to the case where the low gate and the stopped vehicle described in FIGS. 7 to 11 are detected. Here, ● represents actual data with a reflection level equal to or greater than the detection threshold, and ◯ represents extrapolated data with a reflection level less than the detection threshold. Each time corresponds to each time in FIGS.

  From time 1 to time 5, the reflection level of the gate is equal to or higher than the detection threshold value, so that the actual data is obtained. However, after time 6, the reflection level of the gate is less than the detection threshold value, and therefore extrapolated data. The process is continued up to 5 times before time 11. The reflection level of the extrapolation data is set to a magnitude immediately before the extrapolation. When the gate position gradually approaches as time passes and the extrapolated data of the gate enters the system operating area at time 8, automatic braking and warning for preventing collision with the gate which is not necessary is executed. There's a problem.

  The stop vehicle ahead of the host vehicle is continuously detected as actual data from time 4 to time 12, and is deviated to the side of the detection area after time 13 and becomes extrapolated data. (Only three times shown) Insert continuously. Since the actual data of the stopped vehicle enters the system operation area at time 11, automatic braking and alarms that are originally necessary for the stopped vehicle are started at that time.

  FIG. 3 is a time chart showing an embodiment in which the extrapolation frequency is a maximum of three.

  First, time 6 data (see □) is predicted from real data at time 4 and real data at time 5. The prediction is made based on the rate of change from the actual data at time 4 to the actual data at time 5. When the time 6 data predicted in this way (see □) is less than the detection threshold for the stationary object, the actual data at time 5 is forcibly set as temporary extrapolation data, and the extrapolation data subsequent thereto is Limit the maximum number of insertions to two. That is, the maximum number of extrapolated data including the temporary extrapolated data at time 5 is three. The data of time 5 is called temporary extrapolation data because it is extrapolation data even though its reflection level is equal to or higher than the detection threshold.

  Thus, the last extrapolation data is that of time 7, and the extrapolation data of time 7 is outside the area of operation of the anti-collision system S for stationary objects, so unnecessary operation of the anti-collision system S is avoided. This can eliminate the driver's uncomfortable feeling. Even if the gate enters the inside of the operation area of the collision prevention system S at time 8, extrapolation data does not exist at that time, so that the collision prevention system S performs an unnecessary operation and gives the driver a sense of incongruity. Absent.

  On the other hand, even if the data of time 13 (see □) is predicted from the actual data of time 11 and time 12 of the stopped vehicle, the predicted data is equal to or greater than the detection position, so the actual data of time 12 is used as temporary extrapolated data. No processing to reduce the number of extrapolations is performed. When the data falls below the detection threshold at time 13, extrapolated data is inserted up to five times after time 13. Then, as in the conventional example, the actual data of the stopped vehicle enters the system operation area at time 11, and at that time, the originally necessary automatic braking and warning for the stopped vehicle are started.

  Whether the maximum extrapolation count is 5 times normal or 3 times reduced is the reflection level of the next data predicted from the previous data and the current data is greater than or less than the detection threshold. It is decided by what. When an object such as a gate approaches the vehicle, the electromagnetic wave of the radar device R passes through and the reflection level rapidly decreases, and the reflection level of the next data predicted from the previous data and the current data is likely to be less than the detection threshold. On the other hand, an object such as a stopped vehicle does not allow the electromagnetic wave of the radar device R to pass through even if the host vehicle approaches, so that the reflection level does not drop rapidly, and the reflection level of the next data predicted from the previous data and the current data. Is likely to exceed the detection threshold.

  Also, when the next reflection level of the gate is predicted from the reflection level of the gate recognized as the control target last time and the reflection level of the gate recognized as the control target this time, and the reflection level is less than the detection threshold In other words, when the gate reflection level is drastically decreased and the gate is not a control target, the actual data of the gate recognized as the control target is forced to be extrapolated data, which will be described later. As described above, the vehicle control is not performed until the extrapolation data enters the system operation area of the narrow extrapolation data, and the strength of the vehicle control is weakened, so that unnecessary vehicle control is more reliably performed. Can be prevented.

  Next, the above operation will be described in more detail based on the flowcharts of FIGS.

  Vehicle data such as vehicle speed and yaw rate is obtained in step S1 of the main flow in FIG. 4, and the reflection level, distance, left and right position, and relative speed of the target detected by the radar device R are calculated in step S2, and this is used as the target memory. To remember. In the subsequent step S3, a previous target and takeover subflow, which will be described later, are executed, and further, a previous target processing subflow, which will be described later, is executed in step S4.

  In the subsequent step S5, the target of the output flag = 1 (the control target by the collision prevention system S) stored in the target memory this time is output to the electronic control unit U. At this time, an extrapolation counter value is also output. In subsequent step S6, the data in the current target memory is moved to the previous target memory, and in step S7, collision prediction is performed from the vehicle information and the radar detection target data. As a result, there is a target that may cause a collision in step S8, a target whose count value of the extrapolation counter is 1 or more exists in step S9, and in the “system operation area of extrapolation data” in step S10. If there is a target, 1/2 assist control is executed in step S11, and if there is no target whose count value of the extrapolation counter is 1 or more, full assist control is executed in step S12.

  In this way, extrapolation data is also subject to vehicle control in the process of the electronic control unit U. However, extrapolation that is less accurate than actual data is achieved by using 1/2 assist control in the case of extrapolation data. Vehicle control according to the data can be executed. 1/2 assist control means that the strength of vehicle control is weaker than that of full assist control. For example, when performing automatic braking, the deceleration of the braking force is lower than that during full assist. When setting or using both automatic braking and warning at the time of full assist, it corresponds to performing only warning without performing automatic braking.

  Further, the “system operation area of extrapolation data” is divided into a system operation area for a stationary object shown in FIGS. 7 to 11 into a portion close to the own vehicle and a portion far from the own vehicle, of which the vicinity is close to the own vehicle. Defined as part. In the case of actual data, full assist control is executed if it exists inside the system operation area. However, in the case of extrapolation data, the portion of the system operation area that is close to the host vehicle (that is, the system of extrapolation data). The 1/2 assist control is executed only when it exists in the “operation area”). As a result, in the case of extrapolation data that is less reliable than actual data, vehicle control is difficult to be performed until the vehicle approaches the vehicle, and the driver's uncomfortable feeling due to unnecessary vehicle control can be eliminated.

  FIG. 5 shows the contents of step S3 (previous target and takeover subflow) of the flowchart of FIG.

  First, the current target is called from the current target memory in step S21, and the previous target and takeover are confirmed in step S22. This takeover is based on the relative position of the previous target with respect to the vehicle position, and the relative position of the current target is estimated in consideration of the relative speed of the previous target, and if the current target is in the vicinity of the estimated position, When the target is estimated to be the same as the previous target, and this process is continued a predetermined number of times, both targets are assumed to be the same.

  If the current target is the same object that has succeeded the previous target in step S23, it is determined in step S24 whether the target is a stationary object or a moving object. Specifically, if "own vehicle speed" + "relative speed" is in the range of -10km to 10km, that is, if the target is approaching the own vehicle at a speed of "own vehicle speed" ± 10km, the target stops. It judges that it is a thing, and transfers to step S25. In step S25, the next target reflection level is predicted from the previous target reflection level and the current target reflection level. If the predicted next reflection level is greater than or equal to the detection threshold value in step S26, the extrapolation counter is set to 0 in step S27, and the output flag is set to "1" in step S29. On the other hand, if the predicted next reflection level in step S26 is less than the detection threshold, the extrapolation counter is not set to 0 in step S28, and the extrapolation counter is set to 3 which is larger than that, and then the step is performed. In S29, the output flag is set to “1”.

  If the answer to step S24 is NO and the target is a moving object, the steps S25 and S26 are skipped, and the extrapolation counter = 0 is set in the step S27. If the current target does not take over the previous target in step S23, that is, if it is a newly detected target, extrapolation counter = 0 is set in step S30, and the output flag is set to “0” in step S31. Set to.

  Steps S21 to S31 are repeated until all the current targets are called in step S32.

  The current target whose output flag is set to “1” is output to the electronic control unit U in step S5 of the flowchart of FIG. If the current target does not take over the previous target and the reliability is low (NO in step S23), the output flag is cleared to "0", so that vehicle control based on the current target is performed. Never executed.

  FIG. 6 shows the contents of step S4 (previous target processing subflow) in the flowchart of FIG.

  First, in step S41, the previous target is called from the previous target memory. In step S42, if the previous target that was called is not the target that has been taken over by the current target, that is, if it is an extrapolated target, the extrapolation counter in step S43. Is incremented by one. If the count value of the extrapolation counter is 5 or less in step S44, the current data is predicted from the previous data in step S45, that is, the current position is predicted from the previous position of the target. The output flag is set to “1” so as to be a control target. The current data predicted in the subsequent step S47 is stored in the current target data, and the count value of the extrapolation counter is also stored.

  In the case where the previous target called in step S42 is the target taken over by the current target, the case where the count value of the extrapolation counter exceeds 5 in step S44, and the case where the step S47 is passed, Steps S41 to S47 are repeated until all previous targets are called in step S48.

  By the way, when the reflection level estimated in the step S26 is equal to or higher than the detection threshold, the extrapolation counter = 0 is set in the step S27. Therefore, even if the extrapolation process is performed five times, the extrapolation counter> 5 is not satisfied. The maximum number of extrapolation processes is five. On the other hand, if the reflection level estimated in step S26 is less than the detection threshold, the extrapolation counter is set to = 3 in step S28, so that extrapolation counter> 5 when the extrapolation process is performed three times. Therefore, the number of extrapolation processes is limited to a maximum of two. If provisional extrapolation processing is added, the number of times is limited to a maximum of three.

  As a result, collisions are possible by reducing the number of extrapolation processes from normal 5 times to 2 (3 times if provisional extrapolation is applied) for stationary objects such as gates that can dive through. It is possible to prevent the collision prevention system S from reacting excessively to obstacles that do not have a sense of incongruity to the driver.

  The embodiments of the present invention have been described above, but various design changes can be made without departing from the scope of the present invention.

  For example, the normal extrapolation number and the reduced extrapolation number are not limited to the embodiment, and can be arbitrarily set. For example, as the vehicle speed of the host vehicle increases, the number of extrapolation processes when the reflection level is less than the detection threshold can be reduced from three times to two times. Accordingly, since the vehicle speed is high, it is possible to prevent extrapolation data from entering the system operation area before the extrapolation is stopped, and excessive vehicle control is performed to give the driver an uncomfortable feeling.

Block diagram of electronic control unit of control target recognition device Time chart explaining the operation of the conventional example Time chart explaining the operation of the embodiment Figure showing the main flowchart Figure showing the previous target and takeover sub-flowchart The figure which shows the last target processing sub flowchart Action explanatory diagram of conventional example (time 1 and time 4) Action explanatory diagram of conventional example (time 5 and time 6) Action explanatory diagram of conventional example (time 7 and time 8) Action explanatory diagram of conventional example (time 11 and time 12) Action explanatory diagram of conventional example (time 13 and time 14)

Explanation of symbols

M1 Object detection means M2 Relative relation calculation means M2 Relative relation prediction means M4 Identity determination means M5 Control target object recognition means M6 Extrapolation means M7 Stationary object determination means M8 Reflection level prediction means M9 Vehicle control means Ra Transmitting means Rb Receiving means

Claims (4)

Transmitting means (Ra) for transmitting electromagnetic waves toward a predetermined detection area at predetermined time intervals;
Receiving means (Rb) for receiving a reflected wave in which the electromagnetic wave transmitted by the transmitting means (Ra) is reflected by an object;
An object detection means (M1) for detecting an object based on a reflected wave having a reflection level equal to or higher than a detection threshold among the reflected waves received by the receiving means (Rb);
A relative relationship calculating means (M2) for calculating a relative relationship between the vehicle and the object based on the detection result of the object detecting means (M1);
A relative relationship predicting means (M3) for predicting the current relative relationship based on the relative relationship previously calculated by the relative relationship calculating means (M2);
By comparing the relative relationship predicted this time by the relative relationship prediction means (M3) with the relative relationship calculated this time by the relative relationship calculation means (M2), the object detected last time and the object detected this time are the same. Identity determining means (M4) for determining an object;
Control target object recognition means (M5) for recognizing that an object that has been determined to be the same object by the identity determination means (M4) reaches a predetermined number of times as a control target object;
When the object that has been recognized as the control target by the control target object recognition unit (M5) is no longer recognized as the control target this time, the object predicted by the relative relationship prediction unit (M3) Extrapolation means (M6) for extrapolating a predetermined number of times as if it was actually detected;
Stationary object determination means (M7) for determining whether or not the object recognized as the control target is a stationary object;
Reflection level prediction means (M8) for predicting the reflection level at the next detection from the reflection level at the previous detection and the reflection level at the current detection of the object recognized as the control target;
With
The extrapolation means (M6) reduces the limit of the number of extrapolations when the object is a stationary object and the predicted reflection level is less than the detection threshold. Recognition device.   The control object recognition apparatus according to claim 1, wherein the extrapolation means (M6) reduces the limit of the number of extrapolations according to an increase in the vehicle speed of the host vehicle. Transmitting means (Ra) for transmitting electromagnetic waves toward a predetermined detection area at predetermined time intervals;
Receiving means (Rb) for receiving a reflected wave in which the electromagnetic wave transmitted by the transmitting means (Ra) is reflected by an object;
An object detection means (M1) for detecting an object based on a reflected wave having a reflection level equal to or higher than a detection threshold among the reflected waves received by the receiving means (Rb);
A relative relationship calculating means (M2) for calculating a relative relationship between the vehicle and the object based on the detection result of the object detecting means (M1);
A relative relationship predicting means (M3) for predicting the current relative relationship based on the relative relationship previously calculated by the relative relationship calculating means (M2);
By comparing the relative relationship predicted this time by the relative relationship prediction means (M3) with the relative relationship calculated this time by the relative relationship calculation means (M2), the object detected last time and the object detected this time are the same. Identity determining means (M4) for determining an object;
Control target object recognition means (M5) for recognizing that an object that has been determined to be the same object by the identity determination means (M4) reaches a predetermined number of times as a control target object;
When the object that has been recognized as the control target by the control target object recognition unit (M5) is no longer recognized as the control target this time, the object predicted by the relative relationship prediction unit (M3) Extrapolation means (M6) for extrapolating a predetermined number of times as if it was actually detected;
Vehicle control that controls the vehicle based on the relative relationship and changes the content of the control depending on whether the object recognized as the control target object is actually detected or extrapolated Means (M9);
Stationary object determination means (M7) for determining whether or not the object recognized as the control target is a stationary object;
Reflection level prediction means (M8) for predicting the reflection level at the next detection from the reflection level at the previous detection and the reflection level at the current detection of the object recognized as the control target;
With
The extrapolation means (M6) sets an object recognized as a control target this time as an extrapolated object when the object is a stationary object and the predicted reflection level is less than the detection threshold. The vehicle control means (M9) makes vehicle control for the extrapolated object weaker than vehicle control for the actually detected object. Transmitting means (Ra) for transmitting electromagnetic waves toward a predetermined detection area at predetermined time intervals;
Receiving means (Rb) for receiving a reflected wave in which the electromagnetic wave transmitted by the transmitting means (Ra) is reflected by an object;
An object detection means (M1) for detecting an object based on a reflected wave having a reflection level equal to or higher than a detection threshold among the reflected waves received by the receiving means (Rb);
A relative relationship calculating means (M2) for calculating a relative relationship between the vehicle and the object based on the detection result of the object detecting means (M1);
A relative relationship predicting means (M3) for predicting the current relative relationship based on the relative relationship previously calculated by the relative relationship calculating means (M2);
By comparing the relative relationship predicted this time by the relative relationship prediction means (M3) with the relative relationship calculated this time by the relative relationship calculation means (M2), the object detected last time and the object detected this time are the same. Identity determining means (M4) for determining an object;
Control target object recognition means (M5) for recognizing that an object that has been determined to be the same object by the identity determination means (M4) reaches a predetermined number of times as a control target object;
When the object that has been recognized as the control target by the control target object recognition unit (M5) is no longer recognized as the control target this time, the object predicted by the relative relationship prediction unit (M3) Extrapolation means (M6) for extrapolating a predetermined number of times as if it was actually detected;
A vehicle that controls the vehicle based on the relative relationship and changes the timing for performing the control depending on whether an object recognized as a control target object is an actually detected object or an extrapolated object Control means (M9);
Stationary object determination means (M7) for determining whether or not the object recognized as the control target is a stationary object;
Reflection level prediction means (M8) for predicting the reflection level at the next detection from the reflection level at the previous detection and the reflection level at the current detection of the object recognized as the control target;
With
The extrapolation means (M6) sets an object recognized as a control target this time as an extrapolated object when the object is a stationary object and the predicted reflection level is less than the detection threshold. The vehicle control means (M9) makes the vehicle control for the extrapolated object less likely to be performed than the vehicle control for the actually detected object.

Images