JP7354949B2 - wall judgment device - Google Patents

Description

The present disclosure relates to a technique for determining the type of object detected on the side of a vehicle.

The relative velocity of an object detected by an on-vehicle radar device is not the relative velocity along the traveling direction of the vehicle, but the component in the range direction, which is the direction along the straight line connecting the on-vehicle radar device and the object. . Therefore, all objects that exist in the direction perpendicular to the direction of travel of the vehicle are detected as objects whose relative speed is zero. Therefore, it is not possible to distinguish between a wall and a vehicle running parallel to each other, and the wall is mistakenly recognized as a vehicle running parallel to each other, causing a false alarm to be generated in the BSM application, for example. Note that BSM is an abbreviation for blind spot monitor.

Patent Document 1 below describes that in a three-dimensional map expressed by direction, distance, and relative velocity, a stationary object area where stationary objects are detected is determined, and an object with a relative velocity of zero that is detected at a position outside of that area is , a technique has been proposed for determining that an object is moving in a cross-range direction.

JP2017-223461A

However, as a result of detailed study by the inventor, the following problems were found in the conventional technique described in Patent Document 1. In other words, in the conventional technology, the information used to determine whether an object is stationary or not is limited to three types: direction, distance, and relative speed. There was a problem in that it was not possible to determine whether it was a moving object with sufficient accuracy. For example, if radar waves are reflected from road surfaces or walls and are detected in a direction different from the actual direction, information based on reflected waves from a parallel vehicle may be detected in a stationary object area. Ta.

One aspect of the present disclosure provides a technique that improves the accuracy of determining whether an object detected on the side of a vehicle is a wall.

One aspect of the present disclosure is a wall determination device that includes an instantaneous value generation unit (4: S20), a feature value calculation unit (4: S240 to S260), a distribution storage unit (42), and a likelihood ratio calculation unit. A determination unit (4: S150) is provided. The instantaneous value generation unit includes at least an observation direction, an observation distance, and an observation speed using a signal obtained by transmitting and receiving continuous waves to a detection range including a direction orthogonal to the longitudinal direction of the vehicle. is configured to generate instantaneous values that are information. The observation direction is the direction in which there is a reflection point that reflected the continuous wave. The observation distance is the distance from the reflection point. The observation speed is the relative speed to the reflection point. The feature amount calculation section is configured to calculate a plurality of feature amounts using the plurality of instantaneous values generated by the instantaneous value generation section. The distribution storage unit is prepared for each feature, and stores a wall distribution, which is a distribution representing the relationship between the feature and the probability that the vehicle is a wall, and a vehicle distribution, which is a distribution that represents the relationship between the feature and the probability that the vehicle is a parallel vehicle. is memorized. The likelihood ratio calculation unit calculates a likelihood ratio, which is the ratio of the wall probability calculated using the wall distribution and the vehicle probability calculated using the vehicle distribution, for each feature calculated by the feature calculation unit. , and calculates an integrated likelihood ratio by integrating the likelihood ratios for all feature quantities. The determination unit is configured to use the integrated likelihood ratio to determine whether the object represented by the instantaneous value from which the feature value is calculated is a wall or a parallel vehicle.

According to such a configuration, it is possible to determine whether an object present on the side of a vehicle is a wall or a parallel vehicle by combining a plurality of feature quantities calculated from a plurality of instantaneous values. Therefore, the determination accuracy can be improved.

FIG. 2 is a block diagram showing the configuration of an on-vehicle radar device. FIG. 2 is an explanatory diagram showing how the on-board radar device detects the mounting state of the on-vehicle radar device and the relative speed and direction of a stationary object such as a wall. It is a flowchart of main processing. It is a stationary object graph showing the relationship between an observation direction and a theoretical relative velocity observed for a stationary object existing in that direction. It is a flowchart of wall distance estimation processing. FIG. 2 is an explanatory diagram regarding a stationary object area and a speed return range. FIG. 3 is an explanatory diagram showing a search area used for extracting instantaneous values for instantaneous wall distance calculation. It is a flowchart of wall likelihood calculation processing. FIG. 3 is an explanatory diagram showing a search area used for extracting instantaneous values for calculating a first feature amount. FIG. 7 is an explanatory diagram showing a search area used for extracting instantaneous values for second feature quantity calculation. FIG. 7 is an explanatory diagram showing a search area used for extracting instantaneous values for calculating third and fourth feature amounts. It is a flowchart of flag setting processing. These are a histogram and a normal distribution regarding the first feature amount. These are the histogram and normal distribution regarding the second feature amount. These are the histogram and normal distribution regarding the third feature amount. These are the histogram and normal distribution regarding the fourth feature amount. It is a flowchart of filter processing. It is a flowchart of switching determination processing. It is a flowchart of extrapolation processing. It is a flowchart of wall determination processing. FIG. 3 is a timing diagram showing an overview of wall distance estimation processing. FIG. 3 is a timing diagram showing an overview of extrapolation processing. FIG. 3 is a timing diagram showing an overview of switching determination processing. FIG. 3 is an explanatory diagram illustrating switching of walls.

Embodiments of the present disclosure will be described below with reference to the drawings.
[1. composition]
The in-vehicle radar device 1 shown in FIG. 1 is a radar-type device that detects at least speed using the phase of a signal, and includes, for example, a multi-frequency CW method, an FMCW method, and an FCM method that use continuous waves as radar waves. Either one is used. CW is an abbreviation for Continuous Wave, FMCW is an abbreviation for Frequency Modulated Continuous Wave, and FCM is an abbreviation for Fast-Chirp Modulation.

The in-vehicle radar device 1 includes an antenna section 2, a transmitting/receiving section 3, and a signal processing section 4, and is used by being installed in a bumper made of a material that transmits electromagnetic waves, for example. Here, as shown in Fig. 2, the bumper is installed near the right end of the bumper installed on the rear side of the vehicle in the direction of travel, and on the side of the vehicle (i.e., in the direction perpendicular to the longitudinal direction of the vehicle). ) is installed in the detection range. Further, the in-vehicle radar device 1 is communicably connected to other in-vehicle devices via an in-vehicle local area network (not shown).

Hereinafter, a coordinate system with a predetermined position of the vehicle as the origin, an X axis in the width direction of the vehicle, and a Y axis in the longitudinal direction of the vehicle will be referred to as a vehicle coordinate system. The origin of the vehicle coordinate system may be the center position of the vehicle, or may be the installation position of the GPS sensor in the vehicle. Further, a coordinate system having the installation position of the vehicle-mounted radar device 1 as its origin is referred to as a sensor coordinate system. In both the vehicle coordinate system and the sensor coordinate system, the angle in the XY plane is expressed as a plus sign when looking down on the vehicle from above, and a minus sign when turning the clockwise angle with respect to the direction of travel of the vehicle. Note that the mounting position (Xo, Yo) of the vehicle-mounted radar device 1 expressed using the vehicle coordinate system is a known value, and Xo is referred to as the horizontal mounting position and Yo is referred to as the vertical mounting position.

The antenna section 2 includes a plurality of antennas arranged at least horizontally, and transmits and receives electromagnetic waves as radar waves.
The transmitting/receiving section 3 periodically transmits and receives radar waves, which are continuous waves, via the antenna section 2 at regular time intervals. For each received signal received by each antenna belonging to the antenna section 2, a beat signal having a frequency component corresponding to the difference between the received signal and the transmitted signal is generated, and the received data obtained by A/D converting the beat signal is sent to the signal processing section 4. supply

The signal processing unit 4 includes a microcomputer having a CPU 41 and a semiconductor memory (hereinafter referred to as memory) 42 such as ROM or RAM. The signal processing unit 4 may include a coprocessor that performs fast Fourier transform processing and the like. The signal processing unit 4 analyzes the beat signal acquired from the transmitting/receiving unit 3 to perform at least main processing of generating information regarding an object that reflected radar waves. The signal processing unit 4 corresponds to a wall determination device.

[2. process]
[2-1. Main processing]
Next, the main processing executed by the CPU 41 of the signal processing section 4 will be explained using the flowchart of FIG. This process is started every measurement cycle in which radar waves are transmitted and received.

When this process is activated, the CPU 41 constituting the signal processing unit 4 acquires sampling data of the beat signal for one measurement cycle obtained by the transmitting/receiving unit 3 transmitting and receiving radar waves in S10.

In S20, the signal processing unit 4 generates an instantaneous value for each reflection point, which is a point where the radar wave is reflected, using a known method for radar, such as frequency analysis on the sampling data acquired in S110. The instantaneous value includes the distance to the reflection point (hereinafter referred to as observation distance) r, the direction in which the reflection point is located (hereinafter referred to as observation direction) θ, the relative velocity of the reflection point (hereinafter referred to as observation speed) v, and the reflection point. (hereinafter referred to as observed power) p. Note that, as shown in FIG. 2, the observed speed v is not the speed in the traveling direction of the vehicle (that is, the Y-axis direction), but is the speed component in the range direction from the reflection point toward the vehicle-mounted radar device 1.

In subsequent S30, the signal processing unit 4 uses the instantaneous value generated in S20 to execute wall distance estimation processing for estimating the distance to the wall located on the side of the vehicle.
In subsequent S40, the signal processing unit 4 detects a target object by performing tracking using the instantaneous value generated in S20.

In subsequent S50, the signal processing unit 4 determines whether the target is a target to be provided to subsequent processing or not, based on the wall distance Rf estimated in S30 and the tracking result in S40. It is determined whether the object is an unnecessary target. For example, a target located further away than the wall distance Rf in the direction in which a wall exists is determined to be an unnecessary target.

In subsequent S60, the signal processing unit 4 segments target objects to be provided that have the same speed and are within the same distance.
In subsequent S70, the signal processing unit 4 generates target information for each segment, selects and outputs target information to be output to subsequent processing, and ends the process. Note that when selecting target object information, for example, a predetermined number of target objects may be selected from those closest to the host vehicle.

In the main process, S20 corresponds to the instantaneous value generation section.
[2-2. Wall distance estimation process]
Before explaining the wall distance estimation process executed by the signal processing unit 4 in S30, the stationary object graph used in the process will be explained.

As shown in FIG. 2, when there is a stationary object such as a wall on the side of the vehicle, various locations on the wall serve as reflection points. The observation direction θ in which the reflection point on the wall exists and the observation speed v detected for the reflection point have the relationship shown in FIG. 4. That is, at a reflection point that is located directly lateral to the position where the vehicle-mounted radar device 1 is attached (ie, θ=−90 [deg]), v=0. Reflection points located on the side of the vehicle's traveling direction (i.e., in the target object approach area) from this directly lateral reflection point have a positive value (i.e., v>0) indicating that the reflection point is approaching the host vehicle. In addition, the observed speed v of a reflection point located on the opposite side of the vehicle's traveling direction (i.e., the target object separation area) from the reflection point in the lateral direction is a negative value indicating that the reflection point is moving away from the own vehicle (i.e., v<0). In either case, the absolute value of the observed velocity v increases as the distance from the lateral direction increases. However, the stationary object graph has a shape that is line symmetrical with respect to the direction directly behind the host vehicle. Further, the shape of the stationary object graph differs depending on the speed of the host vehicle, and the higher the speed, the greater the slope of the stationary object graph.

The wall distance estimation process executed by the signal processing unit 4 will be explained using the flowchart of FIG. 5.
In S110, the signal processing unit 4 calculates the observation horizontal position Xd and the observation vertical position Yd of the reflection point represented by the instantaneous values for all the instantaneous values generated in S20 according to equations (1) and (2).

Xd=r・sinθ (1)
Yd=r・cosθ (2)
That is, the observation horizontal position Xd and the observation vertical position Yd are positions expressed in a sensor coordinate system based on the mounting position (Xo, Yo) of the vehicle-mounted radar device 1.

In the following S120, the signal processing unit 4 sets the certified speed vd of the instantaneous value for all the instantaneous values generated in S20. That is, since the observed speed v is calculated using phase information of the frequency component extracted from the beat signal, it is not possible to distinguish between the phase β [rad] and the case where the phase is 2nπ + β [rad]. A loop occurs. Here, in addition to the observed velocity v (that is, when n=0), it is assumed that phase folding occurs once in each of the plus direction and the minus direction (that is, when n=±1). When these three provisional instantaneous values with different velocities are plotted on the v-θ coordinate as shown in Figure 6, if there is a provisional instantaneous value plotted within the stationary object area, the velocity of that provisional instantaneous value is adopted as the certified speed vd. If there is no provisional instantaneous value plotted within the stationary object area, the observed speed v is adopted as the certified speed vd. Note that the stationary object area refers to the range of permissible speed error that is set around the stationary object graph. The allowable speed error is set based on, for example, the magnitude of the error included in the detected value of the sensor.

In subsequent S130, the signal processing unit 4 calculates the instantaneous wall distance Rin. The instantaneous wall distance Rin is the distance from the vehicle-mounted radar device 1 to the wall estimated using only the instantaneous values obtained in the current measurement cycle. Specifically, first, the instantaneous value where the observation direction θ satisfies the equation (3) or (4), and the instantaneous value where the observation direction θ satisfies the equation (5) and the observation vertical position Yd satisfies the equation (6) are calculated. Extract.

-55[deg]≧θ≧-75[deg] (3)
-105[deg]≧θ≧-135[deg] (4)
-135[deg]>θ≧-175[deg] (5)
Yd≧-6 [m] (6)
That is, instantaneous values of reflection points existing within the search area A1 indicated by hatching in FIG. 7 are extracted. Note that the specific numerical values shown in formulas (3) to (6) are merely examples, and the present invention is not limited to these values. These numerical values are set so as to satisfy the following two conditions. One condition is that an area including the direction directly beside the vehicle, where it is difficult to determine whether the object is a moving object or a stationary object, is excluded from the search area A1. Another condition is that when the stationary object in the search area A1 is a guardrail, the support of the guardrail that provides a large reflection is included in the search area A1.

Then, the instantaneous values existing in the search area A1 and the stationary object area are extracted, and the observation lateral position Xd of the instantaneous value with the smallest observation distance r among the extracted instantaneous values is defined as the instantaneous wall distance Rin. do. That is, the instantaneous wall distance Rin is the distance in the X-axis direction from the mounting position of the vehicle-mounted radar device 1 to the wall. Note that if the number of instantaneous values extracted from search area A1 is less than the allowable value (for example, 3), the instantaneous wall distance Rin is set to an initial value indicating that the value has not been calculated, that is, it is invalid. Set to value. Hereinafter, such a value will be expressed as NULL.

In subsequent S140, the signal processing unit 4 executes wall likelihood calculation processing. In the wall likelihood calculation process, an integrated likelihood is calculated, which is the ratio of the likelihood that the instantaneous values detected on the right side of the own vehicle are a wall, and the likelihood that the group is a parallel vehicle. Calculate the degree ratio L. The integrated likelihood ratio L takes a larger positive value as the probability that it is a wall is higher than the probability that it is a vehicle, and conversely, the probability that it is a vehicle becomes larger. The higher the likelihood, the more negative the value. Note that details of the wall likelihood calculation process will be described later.

In subsequent S150, the signal processing unit 4 executes filter processing and ends the process. In the filtering process, a more reliable wall distance Rf is calculated using the instantaneous wall distance Rin calculated in S130, the integrated likelihood ratio L calculated in S140, and the calculation results from the previous measurement cycle. Details of the filter processing will be described later.

Note that S130 corresponds to an instantaneous wall distance calculation section, and S150 corresponds to a filter processing section.
[2-3. Wall likelihood calculation process]
The wall likelihood calculation process will be explained using the flowchart of FIG.

In S210, the signal processing unit 4 determines whether or not the instantaneous wall distance Rin is the initial value NULL. If the initial value is NULL, that is, the instantaneous wall distance Rin can be calculated in the current measurement cycle. If the initial value is not NULL, the process ends, and if the initial value is not NULL, the process moves to S220.

In S220, the signal processing unit 4 calculates the relative lateral position Xr for each of the instantaneous values calculated in S20. The relative lateral position Xr is calculated according to equation (7) using the observed lateral position Xd and observed vertical position Yd calculated in S110, and the curvature CVR of the road on which the vehicle is traveling.

Xr=Xd-((Yd+Yo)2/(-2・CVR)) (7)
The relative lateral position Xr is the result of correcting the observed lateral position Xd according to the degree of curve so that when the road is curved, the road can be considered to be a straight line. Note that the curvature CVR may be obtained from map information based on the current position, or may be calculated from information such as the steering angle and yaw rate obtained from other in-vehicle devices mounted on the host vehicle.

In subsequent S230, the signal processing unit 4 calculates the reference lateral position Xf according to equation (8).
Xf=-Rin+Xo (8)
That is, the reference lateral position Xf is the result of converting the lateral position up to the wall expressed in the radar coordinate system to the lateral position up to the wall expressed in the vehicle coordinate system.

In subsequent S240, the signal processing unit 4 calculates the first feature amount D1. Specifically, first, the signal processing unit 4 determines that the relative lateral position Extract the instantaneous value that satisfies.

Xf-1.8[m]≦Xr≦Xf+0.9[m] (9)
Yo-10[m]≦Yd≦Yo (10)
|vd|>0.5 [m/s] (11)
That is, an instantaneous value existing in the search area A2 shown in FIG. 9 and having a certified speed vd with a low possibility of being a parallel vehicle is extracted. The width of the search area A2 is set based on the reference lateral position Xf, and the vertical width is set based on the mounting position of the in-vehicle radar device 1. Equation (11) is a condition for removing instantaneous values that are unlikely to be objects moving at approximately the same speed as the host vehicle, that is, stationary objects such as walls. Note that the specific numerical values shown in formulas (9) to (11) are merely examples, and the present invention is not limited to these values.

Next, the signal processing unit 4 uses the instantaneous values extracted from the search area A2 to calculate the unbiased standard deviation σ of the observation direction θ according to equation (12). This unbiased standard deviation σ is the first feature amount D1.

θav is the average value of the observation direction θ of the instantaneous values extracted from the search area A2, θi is the observation direction of each instantaneous value extracted from the search area A2 and identified by i, and n is the observation direction θ of the instantaneous values extracted from the search area A2. is the number of instantaneous values.

Returning to FIG. 8, in S250, the signal processing unit 4 calculates the second feature amount D2. Specifically, first, the signal processing unit 4 extracts an instantaneous value where the relative lateral position Xr satisfies equation (9) and the observation direction θ satisfies equation (13). Equation (9) is the same as the equation shown above, but is reproduced to list the conditions. The same applies hereafter.

Xf-1.8[m]≦Xr≦Xf+0.9[m] (9)
80[deg]≦|θ|≦100[deg] (13)
That is, instantaneous values are extracted from the search area A3 shown in FIG.

Next, the signal processing unit 4 extracts the observation direction (hereinafter referred to as maximum power direction) θp of the instantaneous value in which the observed power p becomes maximum among the instantaneous values extracted from the search area A3. This maximum power direction θp is the second feature amount D2.
Returning to FIG. 8, in subsequent S260, the signal processing unit 4 calculates the third feature amount D3 and the fourth feature amount D4. Specifically, first, the signal processing unit 4 determines that the relative lateral position Extract the instantaneous value that satisfies.

Xf-1.8[m]≦Xr≦Xf+0.9[m] (9)
Yo-10[m]≦Yd≦Yo (10)
|θ|>105[deg] (14)
That is, instantaneous values are extracted from search area A4 shown in FIG. 11. The meaning of adding equation (14) to the condition is that in the vicinity of the true lateral direction (that is, θ = -90 [deg]), the observed velocity of both moving objects and stationary objects is close to v = 0. This is based on the fact that it is difficult to distinguish between the two based on the speed v (and thus the authorized speed vd).

Next, the signal processing unit 4 uses the instantaneous values extracted from the search area A4 to determine the number of instantaneous values (hereinafter referred to as the number of data within the area) C _IN that exist within the stationary object area and the number of instantaneous values that exist within the stationary object area. The number of existing instantaneous values (hereinafter referred to as the number of out-of-area data) _COU is counted. The number of data within the area _CIN counted here is the third feature amount D3.

Further, the signal processing unit 4 uses the observed speed Vd of the instantaneous value existing in the search area A4 and the value Vs of the stationary object graph at the observation direction θ of the instantaneous value to calculate the normalized speed V _NR according to equation (15). calculate.

V _NR =Vd/Vs (15)
Furthermore, the signal processing unit 4 calculates the average value (hereinafter referred to as normalized average speed) V _AV of the normalized speeds V _NR of the instantaneous values existing in the search area A4. This normalized average speed _VAV is the fourth feature amount D4.

Note that the feature quantities D1 to D4 are not calculated, that is, the feature quantities are invalid if the number of instantaneous values extracted from the search areas A2 to A4 according to the conditions does not reach a predetermined number (for example, 2). There may be one.

Returning to FIG. 8, in subsequent S270, the signal processing unit 4 executes flag setting processing.
Details of the flag setting process will be explained using the flowchart of FIG. 12.
In S271, the signal processing unit 4 resets the speed prohibition flag Fv to OFF.

In subsequent S272, the signal processing unit 4 determines whether the number of data within the area CIN calculated in S260 is larger than the number of data outside the area COU, and if C _IN > C _OU , the process is terminated, and C _IN ≦ If it is _COU , the process moves to S273.

In S273, the signal processing unit 4 calculates the average value _VOAV of the normalized velocity _VNR of the instantaneous value existing outside the stationary object area, and the calculated average value _VOAV is within the guard range (that is, GL≦VO It is determined whether the value is _AV ≦GU). The signal processing unit 4 ends the process if the average value VO _AV is within the guard range, and moves the process to S274 if it is outside the guard range.

Furthermore, assuming that the average speed in the vehicle distribution prepared for the average value VO _AV of the normalized speed V _NR of the instantaneous value existing outside the stationary object area is μc, and the standard deviation is σc, the lower limit guard value GL is calculated by the formula (16), The upper limit guard value GU is expressed by equation (17). The vehicle distribution will be described later.

GL=μc-3×σc (16)
GU=μc+3×σc (17)
Although 3 is used here as the coefficient by which σc is multiplied, it is not limited to 3, and may be smaller than 3 or larger than 3.

In S274, the signal processing unit 4 sets the speed prohibition flag Fv to ON, and ends the process.
When the speed prohibition flag Fv is OFF, it indicates that a significant speed can be calculated based on the instantaneous value, that is, the fourth feature amount D4 is valid. Further, when the speed impossibility flag Fv is ON, it indicates that it is impossible to calculate a significant speed based on the instantaneous value, that is, the fourth feature amount D4 is invalid.

That is, basically, if there are enough instantaneous values on the stationary object graph, it is determined that a significant velocity can be calculated. If there are many instantaneous values that deviate from the stationary object graph, there is a high possibility that the object represented by the instantaneous values is a car, and if it is a car, the calculated speed should be a value within the guard range. Therefore, if the calculated speed is within the guard range, it is determined that a significant speed can be calculated. Furthermore, if the calculated speed is outside the guard range, it is determined that a significant speed cannot be calculated. This is because instantaneous values based on parallel vehicles are not normally shown on the stationary object graph, but if the observation direction θ changes in a direction different from the actual direction due to reflection from the road surface, etc. In some cases, instantaneous values may occur. The flag setting process is intended to eliminate such instantaneous values.

Returning to FIG. 8, in S280, the signal processing unit 4 calculates the integrated likelihood ratio L and ends the process.
Here, the vehicle distribution and wall distribution used in the calculation of the integrated likelihood ratio L executed by the signal processing unit 4 in S280 will be described.

The memory 42 stores vehicle distribution and wall distribution for each of the plurality of feature quantities D1 to D4 extracted from the instantaneous values. The vehicle distribution is a distribution that represents the probability that when the feature amount Di is acquired, the instantaneous value from which the feature amount Di is generated is based on reflection from the vehicle. The wall distribution is a distribution that represents the probability that, when the feature amount Di is acquired, the instantaneous value from which the feature amount Di is generated is based on reflection from the wall. The vehicle distribution and the wall distribution are generated by learning in advance based on feature amounts collected in advance by driving a moving object in various environments where vehicles and walls exist.

Specifically, it is generated by generating a histogram for each feature amount Di and further determining a normal distribution. 13 to 16 illustrate vehicle distribution and wall distribution in the feature quantities D1 to D4 in the form of histograms and normal distributions.

The vehicle distribution and wall distribution generated for each of the four feature quantities D1 to D4 are stored in the memory 42. The information representing the vehicle distribution and the wall distribution may be a mathematical formula representing the distribution, or may be table data representing the distribution. In the memory 42, an area that stores vehicle distribution and wall distribution corresponds to a distribution storage section.

The first feature amount D1 is an unbiased standard deviation σ that represents the spread of the observation direction θ of the instantaneous values extracted in the search area A2. Here, we utilize the characteristic that the standard deviation σ of the observation direction θ tends to be larger at a wall than at a parallel vehicle.

The second feature amount D2 is the observation direction (that is, the maximum power direction) θp of the instantaneous value in which the observed power p extracted in the search area A3 is maximized. Here, we utilize the characteristic that the maximum power direction θp varies in the case of parallel vehicles, but tends to concentrate at almost one point in the lateral direction in the case of a wall.

The third feature amount D3 is the number of data in the area _CIN , which is the number of instantaneous values in the stationary object area extracted in the search area A4. Here, the characteristic that the number of data within an area C _IN tends to be large for walls and small for vehicles running parallel to each other is utilized.

The fourth feature quantity D3 is a normalized average speed _VAV obtained by normalizing the certified speed Vd of the instantaneous value outside the stationary object area in the search area A4 using the value of the stationary object graph, and is the deviation from the stationary object graph. represents the size of Here, the characteristic that the normalized average speed V _AV tends to be close to 1 for walls and close to 0 for vehicles running parallel to each other is utilized.

Then, the signal processing unit 4 first calculates the vehicle probability P (Di|C) and the wall probability P (Di|W) for each of the four feature quantities Di. i=1, 2, 3, 4. Specifically, using the vehicle distribution and wall distribution regarding the feature amount Di stored in the memory 42, the probability corresponding to the value of the feature amount Di is calculated. At this time, if the calculated feature amount Di is smaller than the lower limit guard value GLi, the calculated feature amount Di is rewritten to the lower limit guard value GLi. Similarly, when the calculated feature amount Di is larger than the upper limit guard value GUi, the calculated feature amount Di is rewritten to the upper limit guard value GUi. _Note that the guard value of _{each feature} Di is _determined by the following (18 ) to (21) are set.

GL1=μ _1c -3×σ _1c GU1=μ _1w +3×σ _1w (18)
GL2=μ _2c -3×σ _2c GU2=μ _2c +3×σ _2c (19)
GU3=μ _3w +3×σ _3w (20)
GL4=μ _4c -3×σ _4c GU4=μ _4w +3×σ _4w (21)
The signal processing unit 4 calculates feature scores α1 to α4, which are likelihood ratios for each of the feature quantities D1 to D4, according to equation (22) for each of the feature quantities D1 to D4.

Furthermore, the signal processing unit 4 calculates the integrated likelihood ratio L according to equation (23) when all the feature scores α1 to α4 are valid, and according to equation (24) when at least one feature is invalid. calculate.

wi is a weight, and in this embodiment, both are set to 1. M is the number of feature amounts, and in this embodiment, M=4.
Note that the calculation of the integrated likelihood ratio L uses a naive Bayes method that applies Bayesian estimation on the assumption that all vehicle probabilities P (Di | C) and wall probabilities P (Di | W) are independent events. is used.

Note that S220 corresponds to a lateral position correction section, S240 to S270 correspond to a feature quantity calculation section, and S280 corresponds to a likelihood ratio calculation section.
[2-4. Wall distance filter processing]
The wall distance filtering process executed by the signal processing unit 4 in S300 will be explained using the flowchart of FIG. 17.

In S310, the signal processing unit 4 determines whether the wall detection flag Fd is set to OFF, and if Fd=OFF, the process moves to S320, and if Fd is ON, the process moves to S390. do. Note that the wall detection flag Fd is initialized to OFF at the start of the main process.

In S320, the signal processing unit 4 sets the condition 1 that the instantaneous wall distance Rin is effectively calculated, and the condition 2 that the feature quantities D1 to D4 are all calculated effectively, and sets the condition 1 and the condition 2. It is determined whether both of the following are satisfied. If the signal processing unit 4 determines that both Condition 1 and Condition 2 are satisfied, the process proceeds to S330, and if it determines that at least one of Condition 1 and Condition 2 is not satisfied, the signal processing unit 4 proceeds to S330. The process moves to S360.

In S330, the signal processing unit 4 sets the wall detection flag Fd to ON.
In subsequent S340, the signal processing unit 4 registers the determination score S. Specifically, the integrated likelihood ratio L calculated in S200 is set as the determination score S.

In subsequent S350, the signal processing unit 4 registers the wall distance Rf, and advances the process to S470. Specifically, the instantaneous wall distance Rin is set as the wall distance Rf.
In S360, the signal processing unit 4 sets the wall detection flag Fd to OFF.

In subsequent S370, the signal processing unit 4 initializes the determination score S. For example, let 0 be the initial value of the determination score S.
In subsequent S380, the signal processing unit 4 sets the wall distance Rf to an initial value NULL indicating that the wall distance Rf has not been calculated, and advances the process to S470. For the initial value NULL, a value outside the range used as an effective wall distance Rf in processing is used.

In S390, the signal processing unit 4 determines whether condition 1 and condition 2 are both satisfied, as in S320, and if both are satisfied, the process proceeds to step S400, and if at least one is not satisfied, If so, the process moves to S430.

In S400, the signal processing unit 4 calculates the difference ΔRf between the wall distance Rf obtained in the previous measurement cycle and the instantaneous wall distance Rin according to equation (25), and calculates the absolute value of the difference ΔRf that is set in advance. It is determined whether or not it is smaller than a threshold value THf.

ΔRf=Rf−Rin (25)
When the signal processing unit 4 determines that |ΔRf|<THf, it assumes that the wall distance Rf and the instantaneous wall distance Rin represent the distance to the same object, and moves the process to S410. When the signal processing unit 4 determines that |ΔRf|≧THf, the wall distance Rf and the instantaneous wall distance Rin may represent different distances to objects, and the process proceeds to S430.

In S410, the signal processing unit 4 updates the determination score S using the integrated likelihood ratio L according to equation (26).
S←S+L (26)
However, if the updated determination score S exceeds the score upper limit (for example, 100), the determination score S is limited to the score upper limit. Further, if the updated judgment score S is less than the lower limit value of the score (for example, -100), the judgment score S is limited to the lower limit value of the score.

In subsequent S420, the signal processing unit 4 updates the wall distance Rf using equation (27), and advances the process to S470.
Rf←γ・Rin+(1−γ)・Rf (27)
γ is a real number satisfying 0<γ<1, and may be γ=0.3, for example. The smaller γ is, the higher the stability of the wall distance Rf is, and the larger γ is, the higher the followability of the wall distance Rf is.

In S430, the signal processing unit 4 executes switching determination processing. Details of the switching determination process will be described later, but in the switching determination process, a switching flag Fx and an extrapolation score Sg are set.
In S440, the signal processing unit 4 determines whether the switching flag Fx is set to ON, and if Fx=ON, it is assumed that a wall switching has occurred, and the process proceeds to S450, and Fx=ON. If it is OFF, the process moves to S480. For example, as shown in FIG. 24, a wall change means that at a branching point or merging point on an expressway, the detected wall changes from the wall W1 along the road you are driving on to the road on the branching/merging side. It means switching to the wall W2 along the wall W2 or in the opposite pattern.

In S450, the signal processing unit 4 registers the determination score S. Specifically, the extrapolation score Sg set in S430 is set as the determination score S.
In subsequent S460, the signal processing unit 4 registers the wall distance Rf, and advances the process to S470. Specifically, the instantaneous wall distance Rin is set as the wall distance Rf. In other words, the wall that exists at the instantaneous wall distance Rin (hereinafter referred to as the new wall) is different from the wall that existed at the position of the wall distance Rf (hereinafter referred to as the old wall), and the wall distance Rf is calculated from the old wall. Switch from wall to new wall.

In S470, the signal processing unit 4 initializes extrapolation parameters used in wall extrapolation processing in S480, which will be described later, and advances the process to S490. The extrapolation parameters include an extrapolation counter Cg, redetection distance Rr, extrapolation score Sg, and switching flag Fx, and specifically, Cg←0, Rr←NULL, Sg←0, Fx←OFF. Set.

In S480, the signal processing unit 4 performs extrapolation processing and advances the process to S490.
In S490, wall determination processing is performed and the processing ends.
Note that S430 to S460 correspond to a switching section, and S490 corresponds to a determining section.

[2-4-1. Switching judgment process]
The switching determination process executed by the signal processing unit 4 in S430 will be explained using the flowchart of FIG. 18.

In S510, the signal processing unit 4 determines whether condition 1 and condition 2 are both satisfied, and if both are satisfied, the process proceeds to S520, and if either condition is not satisfied, the process proceeds to S520. , the process moves to S550. In other words, if the process moves to this step due to a negative determination in the previous S390, the process moves to S550, and if the process moves to this step due to a negative determination in the previous S400, the process moves to S520. do.

In S520, the signal processing unit 4 determines whether the re-detected wall distance Rr is an initial value (ie, Rr=NULL). If Rr=NULL, the signal processing unit 4 moves the process to S610, and if Rr≠NULL, the signal processing unit 4 moves the process to S530.

In S530, the signal processing unit 4 calculates the difference ΔRr between the re-detected wall distance Rr and the instantaneous wall distance Rin according to equation (28).
ΔRr=Rr−Rin (28)
In subsequent S540, the signal processing unit 4 determines whether the absolute value of the difference ΔRr is smaller than a preset threshold THr. The threshold THr may be the same as the threshold THf. When the signal processing unit 4 determines that |ΔRr|<THr, it assumes that the re-detected wall distance Rr and the instantaneous wall distance Rin represent the distance to the same object, and moves the process to S570. When the signal processing unit 4 determines that |ΔRr|≧THr, it assumes that the re-detected wall distance Rr and the instantaneous wall distance Rin represent different distances to objects, and moves the process to S550. . Note that the range where |ΔRr|<THr corresponds to the allowable distance range.

In S550, the signal processing unit 4 initializes the extrapolation score Sg. The extrapolation score Sg is set to 0 by initialization.
In subsequent S560, the signal processing unit 4 initializes the re-detection wall distance Rr and ends the process. Upon initialization, the redetection distance Rr is set to NULL.

In S570, the signal processing unit 4 updates the extrapolation score Sg. Specifically, the extrapolation score Sg is updated using the integrated likelihood ratio L according to equation (29).
Sg←Sg+L (29)
In subsequent S580, the signal processing unit 4 updates the re-detection wall distance Rr. Specifically, the instantaneous wall distance Rin is set as the re-detected wall distance Rr. Note that the re-detection wall distance Rr may be updated by using an equation in which the wall distance Rf is replaced with the re-detection wall distance Rr in equation (27).

In subsequent S590, the signal processing unit 4 determines whether the extrapolation score Sg is greater than or equal to the switching threshold THg. The signal processing unit 4 assumes that wall switching has occurred if Sg≧THg, that is, the switching condition is satisfied, and moves the process to S600, and ends the process if Sg<THg. do. The switching threshold THg is set to a value smaller than at least the score upper limit, and may be set to THg=50, for example.

In S600, the signal processing unit 4 sets the switching flag Fx to ON, and ends the process.
In S610, the signal processing unit 4 registers the extrapolation score Sg. Specifically, the integrated likelihood ratio L is set as the extrapolation score Sg.

In S620, the signal processing unit 4 registers the re-detected wall distance Rr, and ends the process. Specifically, the instantaneous wall distance Rin is set as the re-detected wall distance Rr.
[2-4-2. Extrapolation processing]
The extrapolation process executed by the signal processing unit 4 in S480 will be explained using the flowchart of FIG. 19.

In S710, the signal processing unit 4 determines whether or not the speed prohibition flag Fv is ON. If Fv=ON, the process moves to S720, and if Fv=OFF, the process moves to S740.

In S720, the signal processing unit 4 holds the wall distance Rf obtained in the previous measurement cycle.
In subsequent S730, the signal processing unit 4 holds the extrapolated count value Cg and ends the process.
That is, when Fv=ON, the calculation accuracy of the wall distance Rf becomes low, and therefore only the current value is held without making a determination based on the instantaneous value.

In S740, the signal processing unit 4 determines whether the extrapolated count value Cg is smaller than the threshold TH1, and if Cg<TH1, the process moves to S750, and if Cg≧TH1, the process moves to S770. do.

In S750, the signal processing unit 4 holds the wall distance Rf obtained in the previous measurement cycle.
In subsequent S760, the signal processing unit 4 increments the extrapolation count value Cg and ends the process.

In S770, the signal processing unit 4 determines whether the extrapolated count value Cg is smaller than the threshold value TH2, and if Cg<TH2, the process moves to S780, and if Cg≧TH2, the wall is lost. It is assumed that this has occurred, and the process moves to S800. Note that the threshold value TH2 is set to a larger value than the threshold value TH1.

In S780, the signal processing unit 4 sets the wall distance Rf to a preset fixed value CNST. As the fixed value CNST, the maximum detection distance (for example, 15 [m]) of the wall distance Rf is used.
In subsequent S790, the signal processing unit 4 increments the extrapolation count value Cg and ends the process.

In S800, the signal processing unit 4 initializes the wall distance Rf. Upon initialization, the wall distance Rf is set to NULL.
In subsequent S810, the signal processing unit 4 increments the wall extrapolation count value Cg.

In subsequent S820, the signal processing unit 4 sets the wall detection flag Fd to OFF.
In subsequent S830, the signal processing unit 4 initializes the determination score S, the extrapolation score Sg, and the re-detection wall distance Rr, and ends the process. Through this initialization, specifically, S←0, Sg←0, and Rr←NULL are set.

[2-4-3. Wall judgment processing]
The wall determination process executed by the signal processing unit 4 in S490 will be explained using the flowchart shown in FIG. 20. The wall flag Fw is initialized to Fw=OFF at the start of the main process.

In S910, the signal processing unit 4 determines whether the wall flag Fw is ON, and if Fw=ON, the process moves to S920, and if Fw=OFF, the process moves to S950.
In S920, the signal processing unit 4 determines whether the determination score S is less than or equal to the off threshold THoff. The off threshold THoff may be set to 0, for example. If S≦THoff, the signal processing unit 4 moves the process to S930, and if S>THoff, the signal processing unit 4 moves the process to S940.

In S930, the signal processing unit 4 sets the wall flag Fw to OFF and performs processing because the instantaneous value group used to calculate the wall distance Rf is more likely to be a parallel vehicle than a wall. finish.

In S940, the signal processing unit 4 sets the wall distance Rf as the output wall distance, and ends the process.
In S950, the signal processing unit 4 determines whether the determination score S is greater than or equal to the on threshold value THon. The on threshold THon may be set to 50, for example, similarly to the switching threshold THg. If S≧THon, the signal processing unit 4 moves the process to S960, and if S<THon, ends the process.

In S960, the signal processing unit 4 sets the wall distance Rf as the output wall distance.
In subsequent S970, the signal processing unit 4 sets the wall flag Fw to ON and ends the process because the probability that the instantaneous value group used to calculate the wall distance Rf is a wall is sufficiently high.

[3. motion]
The basic operation of filter processing will be explained using FIG. 21.
In FIG. 21, for simplicity, it is assumed that the instantaneous distance Rin is calculated in every measurement cycle, and that all the feature quantities are calculated.

As shown in FIG. 21, when the instantaneous wall distance Rin is calculated in a state where the wall flag Fw=OFF and the wall determination score S=0, the instantaneous wall distance Rin is registered as the wall distance Rf, as shown at time t11. At the same time, the integrated likelihood ratio L is registered as the determination score S.

As shown from time t11 to t12, in the next measurement cycle, the instantaneous wall distance Rin is calculated, and if the distance to the wall distance Rf (that is, |ΔRf|) is sufficiently small, the wall distance Rf and the instantaneous wall distance Rin are calculated. The wall distance Rf is updated by the weighted addition result of Rin. Further, the judgment score S is updated by adding the integrated likelihood ratio L calculated in the current measurement cycle to the judgment score S.

By repeating the same process for each measurement cycle, when the determination score S exceeds the ON threshold THon as shown at time t12, the wall flag Fw is set to ON and the output of the wall distance Rf is started. .
Thereafter, by calculating the integrated likelihood ratio L having a negative value, the judgment score S decreases, and as shown at time t13, when it falls below the off threshold THoff, the wall flag Fw is set to OFF, and the wall flag Fw is set to OFF. The output of the distance Rf is also stopped.

Next, an outline of the extrapolation process will be explained using FIG. 22.
After the instantaneous wall distance Rin is once detected (Fd=ON), if the instantaneous wall distance Rin or any value of the feature quantities D1 to D4 becomes invalid, the extrapolated count value Cg is counted as shown at time t21. The upload will start.

Then, as shown from time t21 to t22, while the extrapolated count value Cg is Cg<TH1, the value of the wall distance Rf is held (that is, extrapolated). As shown from time t23 to t24, while the extrapolated count value Cg satisfies TH1≦Cg<TH2, a fixed value is extrapolated as the wall distance Rf. As shown at time t25, when the extrapolated count value Cg becomes Cg≧TH2, the wall distance Rf is invalidated and the wall detection flag Fd is turned OFF.

However, if the speed prohibition flag Fv is ON, only extrapolation is performed without counting up.
Next, an overview of the wall switching process will be explained. Furthermore, as shown in Fig. 24, wall switching means that when there is a merging lane or branching lane at an expressway interchange, etc., the instantaneous value that is the subject of calculation of the instantaneous wall distance Rin has been detected until then. A situation is assumed in which the user moves from wall W1 to a different wall W2.

As shown in FIG. 23, when the wall flag Fw is ON and the instantaneous wall distance Rin such that the difference ΔRf from the wall distance Rf becomes |ΔRf|>THr is calculated, the determination is made as shown at time t31. An extrapolation score Sg is registered separately from the score S, and a redetected wall distance Rr is registered separately from the wall distance Rf.

When instantaneous wall distances Rin such that the difference ΔRr from the re-detection distance Rr satisfies |ΔRr|<THr are continuously detected, the extrapolation score Sg changes to the integrated likelihood ratio L is updated using Then, as shown at time t32, when the extrapolation score Sg exceeds the switching threshold THg, the switching flag Fx indicating that the wall has switched is turned on, and the redetection distance Rr is registered as the wall distance Rf. Ru. Here, since THg=THon, interruption of the output of the wall distance Rf is suppressed when the wall is switched.

Note that when the instantaneous wall distance Rin is calculated such that the difference ΔRr from the re-detection distance Rr satisfies |ΔRr|≧THr, the extrapolation score Sg and the re-detection distance Rr are invalidated.
[4. effect]
According to the first embodiment described in detail above, the following effects are achieved.

(4a) The in-vehicle radar device 1 uses the integrated likelihood ratio L, which is the result of integrating the ratio of the vehicle probability and the wall probability calculated for each of the plurality of feature quantities D1 to D4 by applying Naive Bayes. , it is determined whether the object located at the wall distance Rf is a wall. In this way, since a plurality of feature amounts are used to determine whether or not it is a wall, the determination accuracy can be improved.

(4b) In the vehicle-mounted radar device 1, if the distance |ΔRf| between the instantaneous wall distance Rin and the wall distance Rf is |ΔRf|≧THf, the re-detection wall distance Rr is registered and the wall distance Rf is excluded. Insert. Then, in subsequent measurement cycles, it is determined whether or not the re-detected wall distance Rr is a wall using the integrated likelihood ratio L. Therefore, if it is determined that the object present at the re-detected wall distance Rr is a wall, it can be recognized that a wall has changed. Additionally, if it is determined that the object present at the re-detection wall distance Rf is not a wall, it may be a vehicle etc. entering from a side road, and it is possible to prevent such an object from being erroneously detected as a wall. .

(4c) The in-vehicle radar device 1 uses the relative lateral position Xr, which is a value obtained by correcting the observed lateral position Xd according to the road shape, to determine whether the instantaneous value is included in the search areas A1 to A4. Therefore, there is no need to change the search areas A1 to A4 depending on the road shape, and the processing load can be reduced.

[5. Other embodiments]
Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above-described embodiments and can be implemented with various modifications.

(5a) In the above embodiment, four feature amounts are used, but the types and number of feature amounts in the present disclosure are not limited to these. That is, the number of feature amounts may be two or three, or five or more.

(5b) The signal processing unit 4 and its method described in the present disclosure are provided by configuring a processor and memory programmed to execute one or more functions embodied by a computer program. It may also be realized by a dedicated computer. Alternatively, the signal processing unit 4 and its techniques described in this disclosure may be implemented by a dedicated computer provided by configuring the processor with one or more dedicated hardware logic circuits. Alternatively, the signal processing unit 4 and its method described in the present disclosure may include a processor configured with a processor and memory programmed to execute one or more functions, and one or more hardware logic circuits. It may also be realized by one or more dedicated computers configured in combination. The computer program may also be stored as instructions executed by a computer on a computer-readable non-transitory tangible storage medium. The method for realizing the functions of each section included in the signal processing section 4 does not necessarily need to include software, and all the functions may be realized using one or more pieces of hardware.

(5c) A plurality of functions of one component in the above embodiment may be realized by a plurality of components, and a function of one component may be realized by a plurality of components. . Further, a plurality of functions possessed by a plurality of constituent elements may be realized by one constituent element, or one function realized by a plurality of constituent elements may be realized by one constituent element. Further, a part of the configuration of the above embodiment may be omitted. Further, at least a part of the configuration of the above embodiment may be added to or replaced with the configuration of other embodiments.

(5d) In addition to the wall determination device realized by the signal processing unit 4 described above, a system including the wall determination device as a component, a program for causing a computer to function as the wall determination device, and a semiconductor memory in which this program is recorded. The present disclosure can also be realized in various forms, such as a non-transitional physical recording medium such as , a wall determination method, etc.

DESCRIPTION OF SYMBOLS 1... Vehicle radar device, 2... Antenna part, 3... Transmission/reception part, 4... Signal processing part, 41... CPU, 42... Memory, A1-A4... Search area.

Claims (10)

Using a signal obtained by transmitting and receiving a continuous wave to a detection range including a direction orthogonal to the longitudinal direction of the vehicle, an observation direction is determined, which is a direction in which a reflection point that reflects the continuous wave is present; an instantaneous value generation unit (4: S20) configured to generate an instantaneous value that is information including at least an observation distance that is a distance to a reflection point, and an observation speed that is a relative speed to the reflection point;
a feature quantity calculation unit (4: S240 to S260) configured to calculate a plurality of feature quantities using the plurality of instantaneous values generated by the instantaneous value generation unit;
A wall distribution, which is a distribution representing the relationship between the feature amount and the probability that the vehicle is a wall, and a vehicle distribution, which is a distribution that represents the relationship between the feature amount and the probability that the vehicle is a parallel vehicle, are stored for each of the feature amounts. a distribution storage unit (42);
For each of the feature quantities calculated by the feature quantity calculation unit, a likelihood ratio is calculated, which is a ratio between a wall probability calculated using the wall distribution and a vehicle probability calculated using the vehicle distribution. , a likelihood ratio calculation unit (4: S280) configured to calculate an integrated likelihood ratio by integrating the likelihood ratios for all the feature quantities;
A determining unit (4: S490) and
A wall determination device comprising: The wall determination device according to claim 1,
The wall determination device uses a degree of dispersion of the observation direction of the instantaneous value as the feature amount. The wall determination device according to claim 1 or claim 2,
The wall determination device uses the observation direction of the instantaneous value having the maximum power as the feature amount. The wall determination device according to any one of claims 1 to 3,
The wall determination device uses the number of instantaneous values in which the deviation of the observed speed from the theoretical relative speed of a stationary object existing in the observed direction is within a preset allowable speed error range as the feature quantity. . The wall determination device according to any one of claims 1 to 4,
The wall determination device uses a normalized average velocity, which is an average value of the observed velocities normalized by a theoretical relative velocity of a stationary object existing in the observation direction, as the feature amount. The wall determination device according to any one of claims 1 to 5,
The position along the vehicle width direction of the vehicle is defined as the lateral position, and the instantaneous value is used in which the deviation of the observed speed from the theoretical relative speed of a stationary object existing in the observation direction is within a range of allowable speed error. an instantaneous wall distance calculation unit (4: S130) configured to calculate an instantaneous wall distance that is an average value of the lateral position of the reflection point represented by the instantaneous value;
further comprising;
The feature amount calculation unit calculates the feature amount using the instantaneous value existing in a search area that is set based on the instantaneous wall distance. The wall determination device according to claim 6,
further comprising a lateral position correction unit (4: S220) configured to correct the lateral position of the reflection point calculated from the instantaneous value according to the shape of the road on which the vehicle is traveling;
The feature value calculation unit determines whether or not the wall exists in the search area using the horizontal position corrected by the horizontal position correction unit. The wall determination device according to claim 6 or claim 7,
The wall determination device further includes a filter processing unit (4: S150) configured to calculate a wall distance to be used for output using the instantaneous wall distance. The wall determination device according to claim 8,
The filter processing unit calculates the instantaneous wall distance and the wall distance when the difference between the instantaneous wall distance and the wall distance calculated in the previous measurement cycle is within a preset allowable distance range. A wall determination device that updates the wall distance based on a weighted addition result. The wall determination device according to claim 9,
If the difference between the instantaneous wall distance and the wall distance exceeds the allowable distance range, the wall distance is held and the instantaneous wall distance is registered as the re-detected wall distance, and in subsequent measurement cycles, the instantaneous wall distance is a switching unit configured to register the re-sensing wall distance as the wall distance when a switching condition including at least that a difference between the distance and the re-sensing wall distance is within the permissible distance range; The wall determination device further includes (4: S430 to S460).

Images