DEVICE AND METHOD FOR SELECTING A VEHICLE IN FRONT

Abstract

Ahead vehicle selection device comprising: a curvature estimator (4, 41, S130) that estimates the curvature of a road on which a subject vehicle is traveling; an object position detection device (4, 42, S120) that detects an object ahead of the subject vehicle and determines a relative position with respect to the subject vehicle for each ahead object; an instantaneous probability calculation device (4, 43, S140) that repeatedly determines a subject vehicle lane probability instantaneous value for each ahead object based on the curvature estimated by the curvature estimator and the relative position determined by the object position detection device, wherein the subject vehicle lane probability instantaneous value is an instantaneous value of a probability of an ahead object being in the same lane as the subject vehicle;a filter calculation means (4, 44, S170) that determines a subject vehicle lane probability for performing a filter calculation on the subject vehicle lane probability instantaneous value calculated by the instantaneous probability calculation means; a lead vehicle selection means (4, 45, S180) that selects a lead vehicle on the basis of the subject vehicle lane probability determined by the filter calculation means;and a filter properties setting means (4, 46, S160) that adjusts properties of the filter calculation such that a forward object associated with a forward vehicle selected by the forward vehicle selection means in a previous processing cycle is relatively less affected by the subject vehicle lane probability instant value compared to a forward object associated with an object other than the forward vehicle, wherein the filter properties setting means modifies the properties of the filter calculation on the basis of a vehicle-to-vehicle time, which is the time required by the subject vehicle to reach a detection position of the forward object, such that the shorter the vehicle-to-vehicle time to the forward object, the more strongly it is affected by the subject vehicle lane probability instant value;the filter computation means determines an instantaneous value of the subject vehicle lane probability in an instantaneous processing cycle by performing the filter computation to calculate, using a weighting, a weighted mean (1) of the subject vehicle lane probability instantaneous value and (ii) of a previous value of the subject vehicle lane probability determined by the filter computation means in a previous processing cycle, and changes the weighting to calculate the weighted mean based on a vehicle-to-vehicle time, which is the time the subject vehicle takes to reach a detection position of the object ahead, and the subject vehicle lane probability is determined by P(t) = Pc × 100 + P(t − 1) × (1 − 100) where: P(t) is an instantaneous value of the subject vehicle lane probability determined in an instantaneous processing cycle;P(t-1) is the previous value of the subject vehicle lane probability, determined in a previous processing cycle; Pc is the instantaneous value of the subject vehicle lane probability; and Σ is a filter constant that is the weight used to calculate the weighted mean.

Description

The present invention relates to a technology for selecting a vehicle (preceding vehicle or "preceding vehicle" for short) that travels in front of an own vehicle (subject vehicle).As a technology for reducing operating forces to be applied from a driver who drives a vehicle, a vehicle-to-vehicle control device is known. The in-vehicle control device detects a vehicle (preceding vehicle) that travels in front of the own vehicle. The in-vehicle control device controls the vehicle speed and the like to maintain a certain distance between the subject vehicle and the preceding vehicle so that the subject vehicle can automatically follow the preceding vehicle.In this type of apparatus, an estimated radius of curvature of the road on which the subject vehicle travels is determined based on the yaw rate and the vehicle speed of the subject vehicle. Further, a radar device or the like is used to detect the position of the vehicle located in front of the subject vehicle. Based on the estimated radius of curvature and the position of the forward vehicle, a subject vehicle lane probability is determined for each detected forward vehicle. The subject vehicle lane probability is the probability that the vehicle is in the estimated lane of the subject vehicle. Based on the subject vehicle lane probability, selection of a preceding vehicle and cancellation of the selection are performed.Since the detection values of the yaw rate fluctuate, the calculation results of the estimated turning radius also fluctuate. Therefore, in order to prevent the effects caused by the variation, filter calculation of the subject vehicle lane probability is performed (see, e.g., JP 3427815 B2).As for a vehicle temporarily selected as a preceding vehicle, highly stable selection is required so that the selection is not erroneously canceled or deleted by noise or the like. On the other hand, with respect to a non-preceding vehicle (hereinafter, "non-preceding vehicle" for short) which is a vehicle other than the preceding vehicle, a delicate selection is required so that when a non-preceding vehicle cuts in front of the subject vehicle, this situation is immediately detected.Such stability and responsiveness can be realized by the characteristics of the filter calculation. However, stability and responsiveness are conflicting requirements. Therefore, there is a problem in that it is difficult to perform filter calculation suitable for both a preceding vehicle and a non-preceding vehicle.For a better understanding of the present invention, reference is also made to DE 196 14 061 A1, JP 2007-331608 A and JP 2002-36908 A.It is thus an object of the present invention to provide a technology in which filter calculation suitable for both a preceding vehicle and a non-preceding vehicle is realized.This object is achieved by the features of claim 1 or claim 3. Advantageous further developments are defined in the dependent claims.In the preceding vehicle selection apparatus according to the present invention, a filter calculation is performed for a preceding object associated with a preceding vehicle in which the subject vehicle lane probability instantaneous value cannot be easily considered. Therefore, when the preceding object is temporarily determined as the preceding vehicle, a process is performed which has very high stability and in which selection is not easily deleted or canceled by noise or the like. Further, a filter calculation is performed for a preceding object associated with a non-preceding vehicle in which a subject vehicle lane probability instantaneous value can be easily considered. Therefore, a process having very high responsiveness is performed in which a cut-in vehicle or the like is immediately selected as the preceding vehicle. In other words, a filter calculation suitable for both a preceding vehicle and a non-preceding vehicle can be realized.Numbers in brackets recited in the claims show the correlations with certain embodiments described below as an aspect of the invention. However, the technical scope of the present invention is not limited thereby.Further, the present invention can be realized by various embodiments in addition to the preceding vehicle selection apparatus described above. For example, the present invention can be realized by a system of which the preceding vehicle selection device is a component or a program that allows a computer to function as any means that is part of the preceding vehicle selection device.In the accompanying drawings are: FIG. 1 is a block diagram showing an overall configuration according to an embodiment of an in-vehicle control system including an in-vehicle controller applicable to a preceding-vehicle selection device; FIG. 2 is a flowchart of a preceding vehicle selection process performed by the vehicle-to-vehicle controller shown in FIG. 1 ; FIG. 3 is a flowchart showing details of a filter constant setting process shown in step S 160 in FIG. 2 ; FIG. 4 is a characteristic diagram showing details of a filter constant table used to set a filter constant in the filter constant setting process shown in FIG. 3; FIG. 5 is a graph showing changes in a subject vehicle lane probability before and after a filtering process performed by the vehicle-vehicle controller shown in FIG. 1 ; FIG. 6 is an explanatory diagram showing an example of a filter constant setting state; and FIG. 7 is a block diagram showing a functional configuration of the vehicle-to-vehicle controller shown in FIG. 1.An embodiment to which the present invention is applied will be explained below with reference to the drawings.A vehicle-to-vehicle control system 1 is installed in a car. The inter-vehicle control system 1 controls the vehicle speed of a vehicle (hereinafter, also referred to as a subject vehicle for short) to maintain a vehicle-to-vehicle distance between the subject vehicle and a vehicle traveling in front of the subject vehicle (hereinafter, also referred to as a "preceding vehicle" for short) at an appropriate distance.As shown in FIG. 1, the in-vehicle control system 1 basically includes an in-vehicle controller 4 that functions as a preceding-vehicle selection device according to the embodiment. The in-vehicle control system 1 further includes a sensor group 2, a switch group 3, and a group 5 including various electronic control / regulation units (ECU). The sensor group 2 includes various sensors used to detect the surroundings of the vehicle as well as the behavior and state of the vehicle. The switch group 3 includes various switches used to input instructions or instructions to the vehicle-to-vehicle controller 4.The ECU group 5 performs various controls based on commands from the in-vehicle controller 4.The sensor group 2 includes at least a radar sensor 21, a yaw rate sensor 22, a wheel speed sensor 23, and a steering sensor 24.The radar sensor 21 transmits laser light in a (preceding) area in front of the subject vehicle to thereby scan a predetermined angle range. The radar sensor 21 further detects a reflected light of the emitted laser light. The radar sensor 21 determines the distance between the subject vehicle and an object that has reflected the laser light based on the time period required for the laser light to travel to and from the object. Further, the radar sensor 21 determines the direction in which the object is located based on the direction in which the laser light is irradiated when the reflected light is detected. The sensor 21 is not limited to a radar sensor 21 using laser light, but may be a sensor using millimeter-band or micrometer-band radio or radio waves, ultrasonic waves, or the like. Further, a camera or the like may be used.The yaw rate sensor 22 outputs signals based on the yaw rate of the vehicle.Wheel speed sensors 23 are attached to the front left wheel, the front right wheel, the rear left wheel, and the rear right wheel. Each wheel speed sensor 23 outputs a pulse signal having a flank (pulse flank) generated at predetermined rotational angles of the wheel shaft. In other words, the wheel speed sensor 23 outputs a pulse signal having a pulse interval depending on the rotational speed of the wheel shaft.The steering sensor 24 outputs signals based on a relative steering angle of a steering wheel (change amount of the steering angle) or an absolute steering angle of the steering wheel (actual steering angle with respect to the steering position when the vehicle is traveling straight).The switch group 3 includes at least a control permission switch 31 and a control mode selection switch 32.The control permission switch 31 is used to input whether or not execution of adaptive cruise control (ACC) is permitted. The ACC is a known control that allows the subject vehicle to travel at a predetermined speed when no vehicle (preceding vehicle) is traveling ahead of it. The ACC performs a follow-up control in which a predetermined inter-vehicle distance is maintained when there is no preceding vehicle.The control mode selection switch 32 is used to select an ACC control mode.The ECU group 5 includes at least an engine ECU 51, a brake ECU 52, and a measurement ECU 53.The engine ECU 51 controls an engine start / stop, a fuel injection amount, an ignition sequence, and the like. The engine ECU 51 includes a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), and the like. Specifically, the engine ECU 51 controls a throttle ACT based on detection values from a sensor that detects the depression amount of an accelerator pedal. The throttle ACT is an actuator that opens and closes a throttle disposed in an air intake passage. Further, the engine ECU 51 controls the throttle ACT to increase and decrease the driving force of an internal combustion engine on the basis of instructions from the in-vehicle controller 4.The brake ECU 52 controls the braking operation of the subject vehicle. The brake ECU 52 includes a CPU, a ROM, a RAM, and the like. Specifically, the brake ECU 52 controls a brake ACT based on detection values from a sensor that detects the amount of depression of a brake pedal. The brake ACT is an actuator that opens and closes a pressure increase control valve and a pressure decrease control valve that are disposed in a hydraulic brake circuit. Further, the brake ECU 52 controls the brake ACT to increase and decrease a braking force of the subject vehicle based on instructions from the vehicle-vehicle controller 4.The measurement ECU 53 performs display control of a measurement display disposed in the vehicle on the basis of instructions of the units of the vehicle including the in-vehicle controller 4. Specifically, the measurement ECU 53 displays, on the measurement display, the vehicle speed, the engine speed, and the execution state and the control mode of the control performed by the in-vehicle controller 4.The in-vehicle controller 4 is basically configured as a known microcomputer including a CPU, a ROM, a RAM, and the like. Further, the vehicle-to-vehicle controller 4 includes a detection circuit, an analog / digital (A / D) conversion circuit, an input / output (I / O) interface, a communication circuit, and the like. The detection circuit and the A / D conversion circuit detect signals output from the sensor group 2 and convert these signals into digital values. The I / O interface receives inputs from the switch group 3. The communication circuit communicates with the ECU group 5. Therefore, detailed description is omitted here.In the in-vehicle controller 4, the CPU executes one or more programs stored in advance in the memory (e.g., the ROM) to perform a predetermined in-vehicle determination process, as described in detail below. Thus, as shown in FIG. 7, the in-vehicle controller 4 is capable of operating as the preceding vehicle selection device including a curvature estimation unit 41 (equivalent to the curvature estimation means), an object position detection unit 42 (equivalent to the object position detection means), an instantaneous probability calculation unit 43 (equivalent to the instantaneous probability calculation means), a filter calculation unit 44 (equivalent to the filter calculation means), a preceding vehicle selection unit 45 (equivalent to the preceding vehicle selection means), and a filter characteristic setting unit 46 (equivalent to the filter characteristic setting means).When the ACC is permitted by the control permission switch 31, the in-vehicle controller 4 periodically (about every 100 ms) performs a preceding-vehicle determination process. Further, the in-vehicle controller 4 performs an in-vehicle control process selected by the control mode selection switch 32 using the determination result of the preceding-vehicle determination process.Among the processes, in the in-vehicle control process, the in-vehicle controller 4 usually controls the vehicle speed by outputting instructions for increasing and decreasing the accelerator operation amount to the engine ECU 51. When control using the accelerator operation amount cannot be supported, the in-vehicle controller 4 limits the vehicle speed by outputting a brake command to the brake ECU 52. Further, the in-vehicle controller 4 outputs various ACC-related display information and commands for generating a warning signal to the measurement ECU 53 when predetermined conditions are satisfied.Next, details of the preceding vehicle determination process executed by the vehicle-to-vehicle controller 4 will be described with reference to the flowchart shown in FIG. 2. In the embodiment, a program that allows the CPU of the in-vehicle controller 4 to perform the in-vehicle determination process shown in FIG. 2 has been stored in the memory (e.g., the ROM) of the in-vehicle controller 4 in advance.When the preceding vehicle determination process is started, the vehicle-to-vehicle controller 4 firstly loads the distance and angle measurement data detected by the radar sensor 21 in step S 110.In subsequent step S 120, the inter-vehicle controller 4 converts the loaded distance and angle measurement data from the polar coordinate system expressed by the data into an orthogonal coordinate system. Based on the converted data, the vehicle-to-vehicle controller 4 performs an object recognition process to recognize a (ahead) object located in front of the subject vehicle. In the object recognition process, the inter-vehicle controller 4 groups (clusters) the distance and angle measurement data. The vehicle-to-vehicle controller 4 then determines the center coordinates of the object, the relative speed with respect to the subject vehicle, and the like for each group (each cluster). The object recognized here is referred to below as the "target". The in-vehicle controller 4 performs the processing in step S 120 and is then capable of functioning as the object position detection unit 42 in FIG. 7.In subsequent step S 130, an estimated turning radius R is calculated based on the following equation (1) based on the yaw rate γ detected by the yaw rate sensor 22 and the subject vehicle speed V calculated based on the detection results from the wheel speed sensors 23. The estimated curve radius R is the curve radius (inverse of the curvature) of a subject vehicle traveling curve. The in-vehicle controller 4 performs the processing in step S 130, and is then capable of functioning as the curvature estimation unit 41 in FIG. 7.In steps S 140 to S 170 described below, the same process is executed for all the targets acquired in step S 120 (object recognition process).In subsequent step S 140, the vehicle-to-vehicle controller 4 calculates a subject vehicle lane probability present value of the target. The subject vehicle lane probability is a parameter that is a measure of the probability of the target being a vehicle traveling in the same lane as the subject vehicle. The subject vehicle lane probability instantaneous value is an instantaneous value of the subject vehicle lane probability calculated based on detection data of the current processing cycle. The in-vehicle controller 4 performs the processing in step S 140 and is then capable of functioning as the present probability calculation unit 43 in FIG. 7.In this process, the in-vehicle controller 4 first determines the position of the target acquired in step S 120 as a converted position on the premise that the road on which the subject vehicle travels is a straight road. The in-vehicle controller 4 uses the estimated turning radius R calculated in step S 130 to determine the position of the target. Subsequently, the vehicle-to-vehicle controller 4 determines the subject vehicle lane probability present value from the converted position of the target using a subject vehicle lane probability map prepared in advance.The subject vehicle lane probability map is a known map in which the probability tends to be highest when the converted position is in the vicinity of a front-continuing longitudinal axis of the subject vehicle and at a short distance from the latter. Further, the likelihood tends to be less the farther away and laterally offset from the forward-continuing longitudinal axis of the subject vehicle the converted position is. A specific example and a specific use of the subject vehicle lane probability map are described in detail in JP-B-3427815.One reason for expressing whether or not the target is in the lane of the subject vehicle in the form of a probability is to assume a deviation between the curvature radius of the curvature (estimated curvature radius R) determined from the yaw rate and the actual curvature radius of the curvature.In subsequent step S 150, the in-vehicle controller 4 calculates an in-vehicle time from the subject vehicle to the destination. The vehicle-to-vehicle time may be determined by dividing the distance to the target by the subject vehicle speed.In subsequent step S 160, the in-vehicle controller 4 performs a process of setting a filter constant. The in-vehicle controller 4 performs the processing in step S 160 and is then capable of functioning as the filter characteristic setting unit 46 in FIG. 7.As shown in FIG. 3, in this process, the in-vehicle controller 4 first determines whether a target being sighted is associated with a target determined as a preceding vehicle in the previous processing cycle in step S 210. When it is determined "Yes" in step S 210, the in-vehicle controller 4 sets a filter constant α from the in-vehicle time in step S 220 based on a preceding-vehicle filter constant table set in advance. On the other hand, when it is determined "No" in step S 210, the in-vehicle controller 4 sets the filter constant α from the in-vehicle time in step S 230 on the basis of a previously set filter constant table for non-preceding vehicles.As shown in FIG. 4, both of the filter constant tables for preceding vehicles and for non-preceding vehicles are set such that the filter constant is an upper limit value when the vehicle-to-vehicle time is shorter than a close distance threshold Ta. The filter constant is a lower limit value when the inter-vehicle time is longer than a long distance threshold Tb. When the inter-vehicle time is the close distance threshold Ta or more and the close distance threshold Tb or less, the filter constant is set to become smaller from the upper limit value to the lower limit value in proportion to the increase in the inter-vehicle time. However, the filter constant table for preceding vehicles is set so that the filter constant α for all vehicle-vehicle times is a larger value than that in the filter constant table for non-preceding vehicles.As shown in FIG. 2, in a subsequent step S 170, the in-vehicle controller 4 uses the filter constant α set in step S 160 and performs a filtering process using the following equation (2) (calculation of the weighted average using the filter constant as the weighted value) for each processing cycle. The inter-vehicle controller 4 thus calculates the subject vehicle lane probability used for determining the preceding vehicle. Here, the subject vehicle lane probability to be determined in a current processing cycle is represented by P(t) (i.e., a current value of the subject vehicle lane probability). The subject vehicle lane probability determined in the previous processing cycle is represented by P(t-1) (i.e., a previous value of the subject vehicle lane probability). The subject vehicle lane probability present value determined in S 140 is represented by Pc.In other words, the filtering process operates as a so-called low-pass filter. The filter becomes weaker (the sensitivity becomes higher), the larger α, or in other words, the shorter the inter-vehicle time becomes. The subject vehicle lane probability instantaneous value Pc is more easily considered in the subject vehicle lane probability P(t). The filter becomes stronger (the stability becomes higher), the smaller α becomes, or in other words, the longer the inter-vehicle time becomes. The subject vehicle lane probability instantaneous value Pc is less easily considered in the subject vehicle lane probability P(t). Further, as shown in FIG. 6, it can also be said that a relatively strong filter with priority on stability is used to calculate the subject vehicle lane probability of a target associated with the preceding vehicle. A relatively weak filter with priority on responsiveness, on the other hand, is used to calculate the subject vehicle lane probability of a target associated with the non-preceding vehicle. The in-vehicle controller 4 performs the processing in step S 170 and is then adapted to function as the filter calculation unit 44 in FIG. 7.In subsequent step S 180, the vehicle-to-vehicle controller 4 determines the preceding vehicle on the basis of the subject vehicle lane probability P(t) calculated in step S 170. The vehicle-to-vehicle controller 4 then ends the process. Specifically, for example, the in-vehicle controller 4 determines a target of the targets whose subject vehicle lane probability P(t) is a threshold value (such as 50%) or above that is the shortest distance from the subject vehicle as the preceding vehicle. The in-vehicle controller 4 performs the processing in step S 180, and is then capable of functioning as the preceding-vehicle selection unit 45 in FIG. 7.The in-vehicle controller 4 performs the in-vehicle control process on the basis of the distance to the target determined as the forward vehicle by the forward vehicle determination process and the relative speed with respect to the target. The in-vehicle controller 4 outputs various commands to the ECU group 5.As described above, in the vehicle-to-vehicle control system 1, the filter calculation is performed on the subject vehicle lane probability instantaneous value Pc when the subject vehicle lane probability P(t) is calculated. The subject vehicle lane probability P(t) is used for preceding vehicle determination. Therefore, as shown in FIG. 5, fluctuations in the subject vehicle lane probability P(t) are prevented from being compared with the subject vehicle lane probability present value Pc. Consequently, frequent change of the determination results of the preceding vehicle can be prevented.Further, the filter characteristics are changed depending on the vehicle-vehicle time. The filter becomes stronger the longer the vehicle-to-vehicle time becomes. The filter becomes weaker as the vehicle-to-vehicle time becomes shorter. Therefore, the subject vehicle lane probability is stabilized for long distances at which the subject vehicle lane probability tends to become unstable. Accordingly, stable selection of the preceding vehicle can be realized. For small distances at which the subject vehicle lane probability is relatively stable, selection of the preceding vehicle can be realized, which is favorable in time in that the driver does not feel uncomfortable.Further, unlike when the filter characteristics are changed based on the inter-vehicle distance, the preceding vehicle may be selected at a timing or timings favorable for (not affecting) the driver for all vehicle speeds (in other words, independent of the vehicle speed). The inter-vehicle control using the preceding vehicle selected at the appropriate timings can be realized at favorable timings.Further, in the in-vehicle control system 1, the filter characteristics to be used differ between a target (hereinafter, simply referred to as "preceding vehicle") associated with a target determined to be the preceding vehicle in the previous processing cycle and a target associated with a non-preceding vehicle (hereinafter, simply referred to as "non-preceding vehicle") other than the preceding vehicle. A strong filter is applied to the forward vehicle. A weak filter is placed on the non-preceding vehicle.Therefore, with respect to the preceding vehicle, cancellation of the selection caused by noise, detection errors, and the like can be prevented. With respect to the non-preceding vehicle, it can be immediately determined whether or not the vehicle is a preceding vehicle when cut-in or the like occurs. In other words, a filter calculation suitable for both the preceding vehicle and the non-preceding vehicle can be realized.(Other Embodiments)An embodiment of the present invention is described above. However, the present invention is not limited to the above-described embodiment, but various other embodiments are possible as a matter of course. (1) According to the above-described embodiment, the estimated turning radius R is calculated from the yaw rate detected by the yaw rate sensor. However, the estimated turning radius R may be calculated from the steering angle detected by the steering sensor. (2) According to the above-described embodiment, an example in which the present invention is applied to an in-vehicle control system is given. However, the present invention is not limited thereto, but may be applied to any system as long as the system sets a preceding vehicle and performs any control based on the state of the preceding vehicle or the relative state between the preceding vehicle and the subject vehicle. (3) The elements of the present invention are conceptual and not limited to those according to the present embodiment. For example, functions provided by a single element may be divided into a plurality of elements. Alternatively, the functions of multiple elements may be combined into a single element. Further, at least some of the configurations according to the above-described embodiment may be replaced with a known configuration having similar functions. Further, at least some of the configurations according to the above-described embodiment may be added or replaced with other configurations according to the above-described embodiment, for example.

Claims

A preceding vehicle selection apparatus comprising: curvature estimating means (4, 41, S130) that estimates a curvature of a road on which a subject vehicle travels; object position detecting means (4, 42, S120) that detects an object in front of the subject vehicle and determines a relative position with respect to the subject vehicle for each preceding object; An instantaneous probability calculation means (4, 43, S140) that repeatedly determines a subject vehicle lane probability instantaneous value for each subject object on the basis of the curvature estimated by the curvature estimation means and the relative position determined by the object position detection means, wherein the subject vehicle lane probability instantaneous value is an instantaneous value of a probability of a subject object ahead to be in the same lane as the subject vehicle; a filter calculation means (4, 44, S170) that determines a subject vehicle lane probability for performing a filter calculation on the subject vehicle lane probability instantaneous value calculated by the instantaneous probability calculation means; a preceding vehicle selection means (4, 45, S180) that selects a preceding vehicle on the basis of the subject vehicle lane probability determined by said filter calculation means; and a filter characteristic setting means (4, 46, S160) that sets characteristics of the filter calculation so that a preceding object associated with a preceding vehicle selected by said preceding vehicle selection means in a previous processing cycle is relatively less affected by the subject vehicle lane probability current value from a preceding object associated with an object other than the preceding vehicle, wherein said filter characteristic setting means sets the characteristics of the filter calculation on the basis of a vehicle-vehicle time that is a time required by the subject vehicle, In order to reach a detection position of the present object, the shorter the vehicle-vehicle time to the present object, the more the subject-vehicle lane probability instantaneous value is affected; the filter calculating means determines an instantaneous value of the subject-vehicle lane probability in a present processing cycle by performing the filter calculation to calculate, using a weight, a weighted average value of (1) the subject-vehicle lane probability instantaneous value and (ii) a previous value of the subject-vehicle lane probability determined by the filter calculating means in a previous processing cycle, and the weight for calculating the weighted average value based on a vehicle-vehicle time that is a time required by the subject vehicle, In order to reach a detection position of the object ahead, the subject vehicle lane probability is changed, and the subject vehicle lane probability is determined by P (t)=P c×α 100+P (t-1)×(1-α 100) wherein: P(t) is a present value of the subject vehicle lane probability determined in a present processing cycle; P(t-1) is the previous value of the subject vehicle lane probability determined in a previous processing cycle; Pc is the subject vehicle lane probability present value; and α is a filter constant that is the weight for calculating the weighted average value.The preceding vehicle selection device according to claim 1, wherein the curvature of the road is estimated by: R = V γ wherein: R is a curvature radius which is the inverse of the curvature of the road; V is a speed of the subject vehicle; and γ is a yaw rate of the subject vehicle.A preceding vehicle selection method comprising the steps of: estimating, by a preceding vehicle selection device, a curvature of a road on which a subject vehicle travels; detecting, by the preceding vehicle selection device, an object located in front of the subject vehicle and determining a relative position of the subject vehicle with respect to each preceding object; repeatedly determining, by the preceding vehicle selection device, a subject vehicle lane probability instantaneous value for each preceding object based on the estimated curvature and the determined relative position, wherein the subject vehicle lane probability instantaneous value is an instantaneous value of a probability that a preceding object is located in the same lane as the subject vehicle; determining, by the preceding vehicle selection device, a subject vehicle lane probability by performing a filter calculation on the calculated subject vehicle lane probability current value; selecting, by the preceding vehicle selection device, a preceding vehicle based on the determined subject vehicle lane probability; setting, by the preceding vehicle selection device, characteristics of the filter calculation such that a preceding object associated with a preceding vehicle selected in a previous processing cycle is relatively less affected by the subject vehicle lane probability instantaneous value from an object associated with an object other than the preceding vehicle, wherein: the characteristics of the filter calculation are changed to be more affected by the subject vehicle lane probability instantaneous value the shorter the vehicle-vehicle time to the preceding object based on a vehicle-vehicle time that is a time required for the subject vehicle to reach a detection position of the preceding object; determining a current value of the subject vehicle lane probability in a current processing cycle by performing the filter calculation to calculate, using a weight, a weighted average value of (1) the subject vehicle lane probability current value and (ii) a previous value of the subject vehicle lane probability determined by the filter calculation means in a previous processing cycle, and changing the weight for calculating the weighted average value based on a vehicle-vehicle time that is a time required for the subject vehicle to reach a detection position of the present object; A subject vehicle lane probability determined by P (t)=P c×α 100+P (t-1)×(1-α 100) wherein: P(t) is a present value of the subject vehicle lane probability determined in a present processing cycle; P(t-1) is the previous value of the subject vehicle lane probability determined in a previous processing cycle; Pc is the subject vehicle lane probability present value; and α is a filter constant that is the weight for calculating the weighted average value.The preceding vehicle selection method according to claim 3, wherein the curvature of the traveling road is estimated by R = V γ wherein: R is a curvature radius which is the inverse of the curvature of the traveling road; V is a speed of the subject vehicle; and γ is a yaw rate of the subject vehicle.

Images