DEVICE FOR MONITORING ADJUSTING LANES

Abstract

Device (1) for monitoring an adjacent lane adjacent to a lane on which a own vehicle (71) is traveling, wherein the own vehicle (71) is a vehicle transporting the device (1), the device (1) comprising: a detector for another vehicle (3) that acquires position information of another vehicle (73) indicating a position of the other vehicle (73) around the own vehicle (71) in relation to the own vehicle (71); an odometry acquisition unit (101, S120) that acquires odometry information indicating a driving state of the own vehicle (71); a trajectory calculation unit (102, S130) that calculates a trajectory (31) of the own vehicle (71) using only the odometry information acquired by the odometry acquisition unit (101, S120); a neighboring lane estimation unit (103, S140) that, based on the Travel trajectory (31) of own vehicle (71),a neighboring lane area calculated by the trajectory calculation unit (102, S130), which estimates an area in which the neighboring lane is present; a unit for determining another vehicle (106, S160-S170), which, based on the position information of the other vehicle (73) acquired by the detector for another vehicle (3) and the neighboring lane area estimated by the neighboring lane estimation unit (103, S140), determines whether the other vehicle (73) is present in the neighboring lane; a probability distribution calculation unit (104, S140), which, based on error factors of the odometry information obtained by the odometry acquisition unit (101, S120), calculates a neighboring lane probability distribution (46, 47) that indicates a distribution of a probability of existence of the neighboring lane area; and an area estimation unit (105, S140),which, based on the neighboring lane probability distribution (46, 47) calculated by the probability distribution calculation unit (104, S140), estimates for each of at least one special probability value a range in which the probability of existence of the neighboring lane range is equal to the special probability value, as a special neighboring lane range corresponding to the special probability value, wherein the unit for determining another vehicle (106, S160-S170) determines in which range of at least one special neighboring lane range estimated by the range estimation unit (105, S140) and corresponding to the at least one special probability value, the position of the other vehicle (73), specified by the position information of the other vehicle (73) acquired by the detector for another vehicle (3),is present and, based on the determination result, determines whether the other vehicle (73) is present in the adjacent lane.

Description

Technical field The present invention relates to a device for monitoring the presence or absence of a vehicle on a lane other than a lane on which one's own vehicle is located. State of the art Several techniques have been proposed for determining the position of a vehicle along a driving path. JP 2013-159246A discloses a technique for estimating the position of a vehicle along a driving path that does not use sensors to observe the external environment, such as a GPS device and a camera, but instead uses odometry information from the vehicle itself regarding vehicle speed, yaw angle, steering angle, and other parameters. Estimating the vehicle's position along the driving path can be used in various systems to improve driving safety, such as lane departure warning systems and lane keeping systems. DE 10 2007 021 762 A1 discloses a device for monitoring an adjacent lane that is adjacent to a lane on which a vehicle is traveling, wherein the vehicle is a vehicle that carries the device, the device comprising: a detector for another vehicle that acquires position information of another vehicle indicating a position of the other vehicle around the vehicle in relation to the vehicle; an odometry acquisition unit that acquires odometry information indicating a driving state of the vehicle; a trajectory calculation unit that calculates a trajectory of the vehicle based on the odometry information acquired by the odometry acquisition unit;a neighboring lane estimation unit that, based on the trajectory of its own vehicle calculated by the trajectory calculation unit, estimates a neighboring lane area, which is an area in which the neighboring lane is present; and a other vehicle determination unit that, based on the position information of the other vehicle acquired by the other vehicle detector and the neighboring lane area estimated by the neighboring lane estimation unit, determines whether the other vehicle is present in the neighboring lane. EP 0 915 350 A2 discloses a device for determining data indicative of the course of a lane, comprising a lane detection sensor which provides lane detection measurement data, an object position sensor which detects at least the distance of an object located in front of the vehicle and its directional angle with respect to the vehicle's direction of travel, a vehicle motion sensor for detecting the vehicle's own motion, and an estimation device to which the lane detection measurement data from the lane detection sensor and the object position measurement data from the object position sensor as well as the vehicle motion measurement data from the vehicle motion sensor are supplied and which, depending on these measurement data, estimates the lane curvature and / or the lateral position of a respective object detected by the object position sensor relative to the lane by means of a predefinable,a dynamic vehicle motion model is used to determine the estimation algorithm. DE 10 2006 010 275 A1 discloses a method for avoiding a collision when a vehicle changes lanes onto a target lane, in which objects in the lateral rear area of ​​the vehicle are detected by means of an environmental monitoring device and measures to influence the vehicle are initiated. The method is characterized by the provision of graduated measures, wherein the measures, according to their level, include warning the driver and / or intervening in a steering system of the vehicle and / or influencing safety devices, and in that the measures of a selected level are initiated, the level being determined depending on the relative position of a detected object with respect to the vehicle and / or a relative speed between the vehicle and the detected object, as well as depending on the fulfillment of a lane-change condition. DE 10 2007 007 540 A1 discloses a device and a method for detecting objects in the vicinity of a vehicle on a roadway and for assigning the detected objects to the lanes of the roadway. The objects are detected and their position relative to the vehicle is determined. Taking into account the path traveled by the vehicle, the geometry of the optically detected lanes relative to the vehicle is determined. By combining the relative position of the detected objects and the determined lane geometry, an object assignment device spatially assigns the objects to the lanes in which the respective objects are currently located. DE 103 41 905 A1 discloses a method for determining the position of a motor vehicle on a road, wherein road geometry data and path data are acquired, and from the acquired road geometry data and path data, first position data of the motor vehicle relative to the road are calculated. From the road geometry data and the path data, road course data are calculated, second position data of a following vehicle relative to the vehicle are acquired, and from the road course data, the first position data of the vehicle, and the second position data of the following vehicle, third position data of the following vehicle relative to the road are calculated. Summary of the invention Technical problem To provide a driver with a high level of safety, it is advantageous to estimate not only the position of their own vehicle but also the position of other vehicles on the road. For example, it is beneficial if, when changing lanes to an adjacent lane, the driver's vehicle can detect whether another vehicle is reversing in that lane. For example, a technique has already been demonstrated in which a radar unit is mounted on the rear of one's own vehicle to detect the presence of another vehicle (for example, the distance and direction between one's own vehicle and the other vehicle). With such a radar-based technique for detecting the other vehicle, it is difficult to determine whether the other vehicle is traveling in a lane adjacent to the lane in which one's own vehicle is traveling. Taking the above into account, the object of the present invention is to provide a device for monitoring an adjacent lane, which is adjacent to a lane in which a vehicle is traveling, and which is capable of accurately estimating whether another vehicle is present in the adjacent lane. This object is achieved by a device for monitoring an adjacent lane having the features of claim 1. The dependent claims are directed to advantageous embodiments of the invention. Solution to the problem According to an exemplary embodiment of the present invention, a device for monitoring an adjacent lane is provided, which is adjacent to a lane on which a separate vehicle is traveling, namely a vehicle transporting the device. The device comprises a detector for acquiring another vehicle, an odometry acquisition unit, a trajectory calculation unit, an adjacent lane estimation unit, and a determination unit for identifying another vehicle. The other vehicle detection detector acquires positional information from another vehicle, indicating its position relative to the vehicle itself. The odometry acquisition unit acquires odometry information indicating the vehicle's driving state. Based on the odometry information acquired by the odometry acquisition unit, the trajectory calculation unit calculates the vehicle's trajectory. Based on the vehicle's trajectory calculated by the trajectory calculation unit, the neighboring lane estimator estimates the area where the adjacent lane is located.The determination unit for determining another vehicle determines, based on the position information of another vehicle acquired by the other vehicle detection detector and the adjacent lane area estimated by the adjacent lane estimation unit, whether the other vehicle is present in the adjacent lane. In the adjacent lane monitoring device, configured as described above, the vehicle's own trajectory is calculated based on odometry information, and the adjacent lane area is estimated based on the vehicle's own trajectory. The position information of another vehicle is acquired by the other vehicle detection device. The determination of whether the other vehicle is present in the adjacent lane is made based on the adjacent lane area and the other vehicle's position information. With the adjacent lane monitoring device, which is designed as described above, it is possible to accurately estimate whether another vehicle is present in the adjacent lane. The above and further tasks, features and advantages of the present invention will become clear with reference to the following detailed description of preferred embodiments in conjunction with the associated drawings. Brief description of the drawings Fig. 1A is a block diagram of a vehicle-mounted system according to an embodiment of the present invention; Fig. 1B is a functional block diagram of a control unit in a driver assistance ECU; Fig. 2 is a flowchart of adjacent lane monitoring processing; Fig. 3 is an explanatory view showing an estimation of a vehicle's own trajectory; Fig. 4 is an explanatory view showing a calculation of an adjacent lane probability characteristic field; Fig. 5 is an explanatory view showing a probability distribution of an adjacent lane; Fig. 6 is an example of an adjacent lane probability characteristic field; and Fig. 7 is an explanatory view showing a calculation of a wall history. Description of the embodiments In the following, embodiments of the present invention are described in more detail with reference to the accompanying drawings, which illustrate these embodiments. However, this invention can be implemented in many different ways and is not limited to the embodiments described below. Instead, these embodiments are such that this disclosure is complete and fully represents the scope of the invention to a person skilled in the art. Reference numerals denote identical elements. (1) Configuration of the system mounted on a vehicle A vehicle-mounted system 1 according to the present embodiment, shown in Fig. 1A, comprises an electronic driver assistance control unit (ECU) 2, a radar system 3, an alarm unit 4, and another control system 5. In the present embodiment, the vehicle-mounted system 1 is mounted in a vehicle 71 (see Fig. 7), which is a four-wheeled motor vehicle. The vehicle 71, which carries the vehicle-mounted system 1, is hereinafter referred to as the vehicle itself. The radar system 3 comprises several radar units 21, 22, .... In the present embodiment, the radar system 3 comprises at least one rear right radar unit 21, which is arranged on the rear right side of the vehicle 71, and one rear left radar unit 22, which is arranged on the rear left side of the vehicle 71. The rear right radar unit 21 obtains information regarding objects on the right side of the vehicle 71 (as well as objects on the rear right side of the vehicle 71). The rear left radar unit 22 obtains information regarding objects on the left side of the vehicle 71 (as well as objects on the rear left side of the vehicle 71).In cases where target information regarding objects present in the entire area on the right side of vehicle 71 (including predefined areas on the rear right and rear left sides of vehicle 71) can be obtained using a single radar unit, this single radar unit can be used instead of the two radar units 21, 22. The radar units 21, 22, ..., which are contained in radar system 3, have essentially the same configuration and function. Therefore, the following description refers to the rear right radar unit 21 as representative of all radar units. The rear right radar unit 21 emits radar waves and receives reflected waves. Based on the reflected waves, it calculates target information regarding a target that has reflected the radar waves and inputs this information into the driver assistance ECU 2. The target information to be calculated by the rear right radar unit 21 can include a distance D from the vehicle 71 to the target, a relative velocity V of the target with respect to the vehicle 71, and an azimuth angle φ of the target with respect to the vehicle 71. Target acquisition methods used in the radar units include, but are not limited to, FMCW and DFCW techniques. In the present embodiment, it is described as an example that the rear right radar unit 21 is a millimeter-wave radar that uses FMCW techniques. The FMCW millimeter-wave radar is well-known; therefore, only a brief overview will be given. In the rear right radar unit 21, a transmit signal exhibiting a triangular waveform is emitted as a radar wave from a transmitting antenna. More precisely, the transmit signal is designed such that each period of the transmit signal contains an ascending interval, in which the frequency increases linearly over time, and a descending interval, in which the frequency decreases linearly over time from the end of the ascending interval (at which point the frequency reaches a maximum frequency). The transmit signal is split into two parts: one part is transmitted to the transmitting antenna to be emitted as a radar wave, and the other part is used as a local signal to process a received signal. The reflected wave, emitted by the transmitting antenna and then reflected by a target, is received by multiple receiving antennas. A received signal at each receiving antenna is mixed with the local signal, generating a beat signal at each receiving antenna. A signal processing circuit calculates target information D, v, φ based on the beat signals using a known signal processing method. The rear right radar unit 21 outputs the calculated target information D, v, φ to the driver assistance ECU 2. The other control system 5 is installed in the vehicle 71 and contains several other ECUs that are not the driver assistance ECU 2. That is, in the present embodiment, the other ECUs are collectively referred to as the other control system 5. The driver assistance ECU 2 communicates data with the other control system 5 via the network 6. In the present embodiment, the network 6 can be a known CAN as a vehicle-specific network. Various odometry data are regularly output from the other control system 5 to the network 6. The driver assistance ECU 2 can obtain the odometry data that is regularly transmitted to the network 6, where the odometry data includes the speed v of the vehicle 71 (referred to as vehicle speed), the yaw rate ω of the vehicle 71, the steering angle of the vehicle 71, and the turning radius R of the vehicle 71, and so on. The other control system 5 calculates the vehicle speed v based on a signal from a vehicle speed sensor (not shown) and periodically outputs the result (vehicle speed data) to network 6. The other control system 5 calculates the yaw rate ω based on a signal from a yaw rate sensor (not shown) and periodically outputs the result (yaw rate data) to network 6. The other control system 5 calculates the steering angle based on a signal from a steering angle sensor (not shown) and periodically outputs the result (steering angle data) to network 6. The other control system 5 calculates the radius of rotation R using a predetermined R-calculation method based on the steering angle and periodically outputs the result (R-data) to network 6.The method for calculating R is one of several methods for calculating R. According to one example of the predetermined method for calculating R, the turning radius R can be calculated with reference to a table in which a relationship between a given steering angle and its corresponding turning radius R is stored in advance. According to an alternative embodiment, the driver assistance ECU 2 can use a table to calculate the turning radius R based on the steering angle obtained via network 6. The driver assistance ECU 2 includes a control unit 10, an input unit 16, and a network interface (hereinafter referred to as "network I / F") 17. The input unit 16 outputs the target information received from the respective radar units 21, 22, ..., which comprise the radar system 3, to the control unit 10. The network I / F 17 is a communication I / F for the control unit 10 to exchange data with the control system 5 via the network 6. The control unit 10 performs various processing operations based on the target information D, V, φ, which is received by the respective radar units 21, 22, which form the radar system 3, via the input unit 16. The control unit 10 has a neighboring lane monitoring function. This function determines whether another vehicle is traveling in a lane adjacent to (or adjacent to) the vehicle's own lane. If another vehicle is traveling in such an adjacent lane and a driving condition of the other vehicle meets a predetermined condition, the alarm unit 4 is activated to issue an alarm. The odometry information obtained from the other control system 5 via the network 6 and the target information obtained from the radar system 3 (primarily from the rear right radar unit 21 and the rear left radar unit 22) are also used to perform the neighboring lane monitoring function. The control unit 10 contains a CPU 11, which executes various programs to perform the corresponding processing, a ROM 12, which stores the various programs, and a RAM 13, which is used as working memory while the CPU 11 performs the processing. A rewritable non-volatile memory, such as flash memory, can be used as the ROM 12. The adjacent track monitoring processing program for performing the aforementioned adjacent track monitoring function is stored in the ROM 12. The adjacent track monitoring function can be performed by the CPU 11, which executes the adjacent track monitoring processing program. (2) Overview of the adjacent lane monitoring processing The adjacent lane monitoring processing to be performed in the control unit 10 is described below with reference to Fig. 2. The control unit 10 is configured to monitor adjacent lanes on the right and left sides of the vehicle. Here, only the adjacent lane monitoring processing for the adjacent lane on the right side of the vehicle is described. This processing can also be used for the adjacent lane on the left side of the vehicle. After starting up, the CPU 11 of the control unit 10 reads the program for the adjacent lane monitoring processing of Fig. 2 from the ROM 12 and performs this processing every predetermined control period T. After initiating the adjacent lane monitoring processing by the CPU 11 of the control unit 10, the control unit 10 obtains target information from the radar unit (i.e., the rear right radar unit 21) via the input unit 16 in step S110. The target information includes the position of the other vehicle (a relative position with respect to the own vehicle) and the position of a wall. The position of the wall is the position of a target if the wall is detected as a target on the right side of the own vehicle (a relative position with respect to the own vehicle, for example, coordinates with an origin at the position of the own vehicle).Determining the position of the wall allows for the calculation of a direction and a distance W from the vehicle to the wall. In step S120, the control unit 10 (more precisely, the CPU 11 of the control unit 10) obtains odometry information from the other control system 5 via the network 6. As described above, the odometry information includes at least a vehicle speed v, a yaw rate ω, a steering angle, and a turning radius R. In step S130, the control unit 10 estimates a driving trajectory of the own vehicle. More precisely, the control unit 10 estimates a driving trajectory of the own vehicle up to a predetermined number of previous cycles of the control period T (for example, up to the previous N cycles, where N is a positive integer greater than one). In the present embodiment, as described later, the control unit 10 calculates estimated positions of the own vehicle relative to its current position at the respective control times for N cycles prior to the current cycle, using the odometry information obtained in step S120 (values ​​obtained at the respective control times for up to N previous cycles).The control unit 10 estimates a line connecting the current position of the own vehicle and the estimated positions of the own vehicle as a driving trajectory of the own vehicle. Note that errors in the odometry information obtained in step S120, which includes the vehicle speed v and the yaw rate ω, may be present due to various factors, such as errors in the vehicle speed sensor and the yaw rate sensor, and noise. Therefore, in step S130, for each of the estimated positions (or position estimates) of the own vehicle at the respective time points for up to N previous cycles, the control unit 10 also calculates an estimated range of existence of the estimated position of the own vehicle, taking into account the errors in the odometry information. The estimated range of existence can be expressed as an error variance of the estimated position of the own vehicle.Furthermore, a projection of the error variance onto a lane width direction (that is, a direction perpendicular to the direction of travel) allows the probability of existence of the estimated position of one's own vehicle in the lane width direction to be expressed as a predefined probability distribution with the center at the estimated position of one's own vehicle. In the present embodiment, the error variance of the estimated position of the vehicle due to the error factors of the odometry information is modeled as a normal probability distribution (i.e., a Gaussian distribution). This means that the probability of the estimated position of the vehicle existing, calculated using the odometry information, results in a peak value with the highest probability in the normal probability distribution. Thus, the probability of the vehicle existing decreases with an increase in the track width distance from the estimated position of the vehicle, according to the normal probability distribution. In step S140, the control unit 10 calculates a neighboring lane probability characteristic field. More precisely, for each of the estimated positions of its own vehicle at the respective control times for the previous N cycles, the control unit 10 defines a position of the neighboring lane (more precisely, positions of both ends in the lane width direction) and projects the error variance of the estimated position of its own vehicle onto the position of the neighboring lane, thereby calculating a probability distribution for the neighboring lane. More precisely, the control unit 10 defines the positions of two lane divider lines that define the adjacent lane (lane divider lines on both sides of the adjacent lane). This means an inner estimated divider position, which is an estimate of the lane divider line's position close to the vehicle (inner divider position), and an outer estimated divider position, which is an estimate of the lane divider line's position on the far side of the vehicle (outer divider position). The control unit 10 then simply projects the error variance of the vehicle's estimated position onto the respective inner and outer estimated divider positions, thereby defining a lane width probability distribution for each of these inner and outer divider positions. This means that the probability of existence of the estimated inner separation point results in a peak value corresponding to the highest probability in the normal probability distribution. This probability decreases with an increase in the track width distance to the estimated inner separation point. Similarly, the probability of existence of the outer separation point also results in a peak value corresponding to the highest probability in the normal probability distribution. This probability decreases with an increase in the track width distance to the estimated outer separation point. The control unit 10 connects points of equal probability of existence (for example, points with a predetermined probability of existence P0) at the inner separation position at the respective control times for up to N previous cycles, and connects points of equal probability of existence (for example, points with a predetermined probability of existence P0) at the outer separation position at the respective control times for up to N previous cycles, thereby calculating the neighboring lane probability characteristic field. This neighboring lane probability characteristic field represents a region in which the probability of existence of the neighboring lane is equal to P0. That is, if a destination is located within such a region represented by the neighboring lane probability characteristic field, the probability of existence of the destination on the neighboring lane is determined to be P0. In step S150, the control unit 10 calculates a wall history. More precisely, based on the positions of a wall obtained in step S110 at the respective times for up to N previous cycles, the control unit 10 calculates the wall history, which is a trajectory of a wall's position. In step S160, the control unit 10 calculates the probability of a target existing on the adjacent lane. More precisely, the control unit 10 calculates a probability distribution for a target's existence based not only on the adjacent lane probability curve field calculated in step S140, but also on the wall history calculated in step S150. Essentially, the control unit 10 calculates the probability of a target's existence according to the adjacent lane probability curve field. However, if a wall is present, the probability of an adjacent lane existing in any area outside the wall is zero. Consequently, the probability of a target existing on the adjacent lane is zero. In step S170, the control unit 10 determines the presence or absence of a target on the adjacent lane. More precisely, the control unit 10 determines the presence or absence of a target on the adjacent lane based on the probability of a target's existence, calculated in step S160, and target position information acquired by the radar unit. For example, if a target is present in an area where the probability of a target's existence is equal to or greater than 70%, it is determined that the target is present on the adjacent lane. If a target is present in an area where the probability of a target's existence is less than 70%, it is determined that the target is not present on the adjacent lane. Furthermore, in an example from Fig. 5, which is described later, the probability of existence of the adjacent lane is P1, P2, or P3 (where P1 > P2 > P3), and there are three regions: one region with a probability of existence of P1, one region with a probability of existence of P2, and one region with a probability of existence of P3. For example, if a target exists in the region with a probability of existence of P1 or greater, it can be determined that the target is present on the adjacent lane. In step S180, the control unit 10 extracts an alarm object in a specific state for the vehicle. More precisely, if step S170 determines that a target is present in the adjacent lane, the control unit 10 checks for a positional relationship between the target and the vehicle. If the positional relationship between the target and the vehicle meets a predefined condition, the control unit 10 extracts the target as an alarm object. The predefined condition can be a first condition, that the relative distance of the target to the vehicle is equal to or less than a predetermined distance; a second condition, that the rate of decrease in the relative distance is equal to or greater than a decrease rate threshold (i.e., the target is rapidly approaching the vehicle); or a third condition, according to which both the first and second conditions are met. In step S190, control unit 10 performs alarm output processing. More precisely, when an alarm object is extracted in step S180, control unit 10 performs special processing to trigger alarm unit 4, causing it to output an alarm indicating the presence of the alarm object. (3) Details of the adjacent lane monitoring processing The adjacent lane monitoring processing is described in more detail below with reference to Fig. 2. The following describes various processing operations using different equations. A time (point in time) t, a yaw rate ω(t), a vehicle speed v(t), a pitch angle θ(t), a magnitude of a pitch angle change Δθ(t), and a turning radius R are defined according to the following equations (1)-(6). (Eq. 1) In the present embodiment, as described above, various processing operations are performed based on the estimated positions of the vehicle at the respective time points for up to N previous cycles (i.e., during a period from t=0 to t=-NT). The turning radius R, defined by equation (6), is used in the following description. Alternatively, a turning radius R obtained from the other control system 5 via the network 6 can be used. (3-1) Estimation of the vehicle's trajectory (S130) In step S130, the estimated positions x of the own vehicle at the respective time points for up to N previous cycles are calculated using the following equation (7) with respect to the current position of the own vehicle (origin). (Eq. 2) where x(0) = 0 That is, as shown in Fig. 3, the control unit 10 uses equation (7) at control time t1 to calculate the estimated positions of the own vehicle at the respective times for the N cycles before the current cycle with respect to the position of the own vehicle (origin) at time t1. A line connecting the estimated positions of the own vehicle yields the driving trajectory 31 of the own vehicle. Furthermore, taking into account errors in the odometry information, the control unit 10 calculates estimated existence ranges (error variances) 31a, 31b, 31c... for the respective estimated positions of its own vehicle, which were calculated as described above. More precisely, the control unit 10 defines an error distribution for the yaw rate ω and an error distribution for the vehicle speed v according to equations (8), (9). (Eq. 3) Then an estimate (a value that takes into account the error variance) of the magnitude of the change in the attitude angle Δθ can be expressed by the following equation (10). (Eq. 4) Thus, taking into account errors in the odometry information, the first and second terms on the right-hand side of equation (7) can be expressed according to equations (11) and (12). (Eq. 5) Taking into account errors in the odometry information, the estimated positions of the own vehicle at the respective time points for the N previous cycles up to the current cycle can be expressed according to equation (13). (Eq. 6) where Taking into account the error variances, the estimated positions of the own vehicle can be expressed as a probability distribution (in the present embodiment a normal probability distribution) given by equation (14). (Eq. 7) where In this way, at control time t1, as shown in Fig. 3, the control unit 10 calculates the estimated positions of the vehicle at the respective control times for the N cycles prior to the current time, their error variances (probability distributions), and the vehicle's trajectory 31. Then, at time t2, which is T seconds after time t1, the control unit 10 again calculates the estimated positions of the vehicle at the respective times for up to n previous cycles, their error variances (probability distributions), and the vehicle's trajectory 31. In this way, the estimated positions of the vehicle at the respective control times for up to N previous cycles, their error variances (probability distributions), and the vehicle's trajectory 31 are updated at each control time. (3-2) Calculation of the neighboring lane probability characteristic field (S140) In step S140, the control unit 10 calculates the adjacent lane probability characteristic curve field. First, a definition of the adjacent lane is described based on the estimated position of the vehicle at time t = -T, which is one cycle earlier than the determined control time (time t = 0), with reference to Fig. 4(a). In the present embodiment, the lane width is assumed to be 3.5 m. As shown in the upper part of Fig. 4(a), the control unit 10 defines a vehicle width line 36, which is a line passing through the estimated position of the own vehicle at time t = -T and is perpendicular to the own vehicle's trajectory 31 from the estimated position at time t = -T to the current position of the own vehicle. The control unit 10 defines an estimated inner separation position 41, which is a position on the vehicle width line 36 separated from the estimated position of the own vehicle at time t = -T by a predetermined distance (for example, half the vehicle width, i.e., 1.75 m), and subsequently defines an estimated outer separation position 42, which is a position on the vehicle width line 36 one vehicle width (3.5 m) further away from the estimated inner separation position 41. In this way, an adjacent lane can be estimated for up to the previous cycle as an area between the estimated inner dividing line 43, which is perpendicular to the vehicle width line 36 and passes through the estimated inner dividing position 41, and the estimated outer dividing line 44, which is perpendicular to the vehicle width line 36 and passes through the estimated outer dividing position 42. Subsequently, as shown in the central part of Fig. 4(a), the ECU 10 calculates a lane width probability distribution 38 of the vehicle for its estimated position, taking into account the odometry information. More precisely, the control unit 10 calculates the lane width probability distribution (estimated area of ​​existence or area of ​​existence) 38 of the vehicle by projecting the error variance 31a of the estimated position of the vehicle onto the lane width direction. That is, the distortion or uncertainty of the estimated position of the vehicle in the lane width is estimated. As shown in the lower part of Fig. 4(a), the control unit 10, assuming that the probability distribution 38 of the own vehicle is also a probability distribution of each of the estimated separation positions 41, 42, simply applies the probability distribution 38 of the own vehicle to each of the estimated separation positions 41, 42. More precisely, an inner probability distribution 46 is calculated by applying the probability distribution 38 of the own vehicle to the estimated separation position 41 such that the peak value of the probability distribution 38 results in a peak value of the probability distribution at the estimated inner separation position 41 (that is, the probability distribution of the estimated separation position 41 is calculated by a lane-width translation of the probability distribution 38 of the own vehicle).An outer probability distribution 47 is calculated by applying the probability distribution 38 of the vehicle to the estimated separation position 42 such that the peak value of the probability distribution 38 results in a peak value of the probability distribution at the estimated outer separation position 47. Note that the probability values ​​can change abruptly compared to those of the previous control time (control time of the previous cycle) due to measurement errors using sensors that acquire the odometry information. In the present embodiment, the probability value at each estimated position of the vehicle is smoothed using the probability values ​​at the estimated position of the vehicle calculated at a previous control time. More precisely, in the present embodiment, the probability value is smoothed using the probability value of the previous cycle according to equation (15), where α is a forgetfulness factor. (Eq. 8) Such smoothing of the probability values ​​can prevent the effects of abrupt changes in the probability values ​​due to measurement errors using the sensors. In the present embodiment, a truly significant distribution region of the inner probability distribution 46 (a region used in a subsequent calculation) is a region on the left side (on the side of the driving trajectory 31) of the peak value of the inner probability distribution 46. Therefore, in the following, a reference to the inner probability distribution 46 refers to the region on the left side of the peak value. A truly significant distribution region of the outer probability distribution 47 is a region on the right side (on the side of the driving trajectory 31 of the vehicle itself) of the peak value of the outer probability distribution 47. Therefore, in the following, a reference to the outer probability distribution 47 refers to the region on the right side of the peak value. The control unit 10 defines an inner variance-specific position 51, where the probability of existence of the estimated inner separation position 41 assumes a probability of P1, and an outer variance-specific position 56, where the probability of existence of the estimated outer separation position 42 assumes a probability of P1. In some embodiments, the control unit 10 can define several inner variance-specific positions 51, where the probability of existence of the estimated outer separation position 42 assumes different values, and several outer variance-specific positions 56, where the probability of existence of the estimated outer separation position 42 assumes different values. Such processing is performed in a similar manner at each control point for up to N cycles prior. As shown in Fig. 4(b), the control unit 10 calculates an inner lane separation curve array 61 by connecting the inner variance-specific positions 51 with the same probability of existence P1 at the respective control times for up to N cycles prior. Similarly, the control unit 10 calculates an outer lane separation curve array 66 by connecting the outer variance-specific positions 51 with the same probability of existence P1 at the respective control times for up to N cycles prior. The neighboring lane probability curve array is thus calculated. That is, the neighboring lane probability curve array represents a region between the inner lane separation curve array 61 and the outer lane separation curve array 66 and is a region in which the probability of existence of the neighboring lane is equal to P1. Assuming that the neighboring lane probability curve field of the existence probability P1 is calculated as shown in Fig. 4(b), if a target 73 is detected by the radar unit and is present within the neighboring lane probability curve field, it can be determined that the existence probability of the target on the neighboring lane is equal to or greater than P1. If a target 73 detected by the radar unit is present outside the neighboring lane probability curve field, it can be determined that the existence probability of the target on the neighboring lane is less than P1. Several existence probabilities can be defined for each of the estimated separation positions 41, 42. In such an embodiment, the inner and outer separation line characteristic curve fields can be calculated for each of the existence probabilities (thus, the neighboring lane probability characteristic curve field can be calculated for each of the existence probabilities), which allows a more precise calculation of the existence probability of a target on the neighboring lane. Fig. 5 shows the probability distribution of the adjacent lane for the estimated position of the vehicle at the control point several cycles earlier than the current cycle. As shown in Fig. 5, the inner probability distribution 46, which is a probability distribution of the estimated inner dividing line 43, is centered at the estimated inner dividing line position 41, and the outer probability distribution 47, which is a probability distribution of the estimated outer dividing line 44, is centered at the estimated outer dividing line position 42. The position with the highest probability of existence of the estimated inner dividing line 43 is, of course, the estimated inner dividing position 41. The probability of existence of the estimated inner dividing line 43 decreases with an increase in the distance between the estimated inner dividing line 43 and the estimated inner dividing position 41. In the present embodiment, three positions on the vehicle width line 36 are defined, corresponding to the three different probability values ​​of the probability distribution.More precisely, in the present embodiment, a position of a probability P1 corresponding to 1σ of the normal probability distribution is defined as an inner position of the first variance 51, a position of a probability P2 corresponding to 2σ of the normal probability distribution is defined as an inner position of the second variance 52, and a position of a probability P3 corresponding to 3σ of the normal probability distribution is defined as an inner position of the third variance 53, where P1 > P2 > P3. Similarly, the position with the highest probability of existence of the estimated outer dividing line 43 is, of course, the estimated outer dividing position 42. The probability of existence of the estimated outer dividing line 44 decreases with an increase in the distance between the estimated outer dividing line 44 and the estimated outer dividing position 42. In the present embodiment, three positions on the vehicle width line 36 are defined, corresponding to three different probability values ​​of the probability distribution (equal to the three probability values ​​of the inner probability distribution 46 above).More precisely, in the present embodiment, a position of a probability P1 corresponding to 1σ of the normal distribution function is defined as an outer position of first variance 56, a position of a probability P2 corresponding to 2σ of the normal probability function is defined as an outer position of second variance 57, and a position of a probability P3 corresponding to 3σ of the normal probability distribution is defined as an outer position of third variance 58, where P1 > P2 > P3. As shown in Fig. 6, the control unit 10 calculates a first inner dividing line characteristic field 61 by connecting inner positions of first variance 51 with the same probability of existence P1 at the respective control times up to N cycles prior. Similarly, the control unit 10 calculates a second inner dividing line characteristic field 62 by connecting inner positions of second variance 52 with the same probability of existence P2 at the respective control times up to N cycles prior. The control unit 10 also calculates a third inner dividing line characteristic field 63 by connecting inner positions of third variance 53 with the same probability of existence P3 at the respective control times up to N cycles prior. Similarly, the control unit 10 calculates a first outer dividing line characteristic field 66 by connecting outer positions of first variance 56 with the same probability of existence P1 at the respective control times for up to N cycles prior. Similarly, the control unit 10 calculates a second outer dividing line characteristic field 67 by connecting outer positions of second variance 57 with the same probability of existence P2 at the respective control times for up to N cycles prior. The control unit 10 also calculates a third outer dividing line characteristic field 68 by connecting outer positions of third variance 58 with the same probability of existence P3 at the respective control times for up to N cycles prior. The neighboring lane probability characteristic field of probability P1 represents an area between the inner dividing line characteristic field 61 and the outer dividing line characteristic field 66. If a target detected by the radar unit is present in this area, it can be said that the probability of existence of the target on the neighboring lane is equal to P1. The neighboring lane probability curve field of probability P2 represents an area between the inner dividing line characteristic curve field 62 and the outer dividing line characteristic curve field 67. If a target detected by the radar unit is present in this area but not in the area represented by the neighboring lane probability curve field of probability P1, it can be determined that the probability of existence of the target on the neighboring lane is equal to P2. The neighboring lane probability curve field of probability P3 represents an area between the inner dividing line characteristic curve field 63 and the outer dividing line characteristic curve field 68. If a target detected by the radar unit is present in this area, but is not present in either the area represented by the neighboring lane probability curve field of probability P1 or the area represented by the neighboring lane probability curve field of probability P2, it can be determined that the probability of existence of the target on the neighboring lane is equal to P3. In the example shown in Fig. 6, a target 73 exists outside the area represented by the neighboring lane probability curves of probability P2 and within the area represented by the neighboring lane probability curves of probability P3. Therefore, in this example, it can be determined that the probability of existence of target 73 on the neighboring lane is equal to P3. Although various methods are conceivable for determining the presence or absence of a target 73 in the area represented by the neighboring lane probability characteristic field, the following simplified method can be used. In the simplified method, a vehicle width line 36 that is closest to the target 73 is selected from the vehicle width lines 36 that were calculated at the respective time points up to N cycles prior, and, as shown in Fig. 5, a perpendicular is drawn from the target 73 to the selected vehicle width line 36 to calculate a corresponding position 60 of the target where the perpendicular and the selected vehicle width line 36 intersect. If the corresponding position 60 of the target is located between the positions of first variance 51, 56 on the vehicle width line 36 (that is, in the area represented by the neighboring lane probability characteristic field of probability P1), the probability of existence of the target 73 on the neighboring lane is determined as P1. If the corresponding position 60 of the target is located between the second variance positions 52 and 57 on the vehicle width line 36 (that is, in the area represented by the neighboring lane probability curve field of probability P2), but not between the first variance positions 51 and 56 on the vehicle width line 36, the probability of the existence of target 73 on the neighboring lane is determined as P2. If the corresponding position 60 of the target is located between the third variance positions 53 and 58 on the vehicle width line 36 (that is, in the area represented by the neighboring lane probability curve field of probability P3), but not between the second variance positions 52 and 57 on the vehicle width line 36, the probability of the existence of target 73 on the neighboring lane is determined as P3. (3-3) Calculation of wall history (S150) In the wall history calculation process in step S150, the control unit 10 calculates a distance between the vehicle (vehicle 71) and the wall immediately adjacent to the vehicle, based on the position of a wall obtained by the radar unit 21. More precisely, as shown in Fig. 7, when a wall is detected in a direction perpendicular to the vehicle's trajectory from the estimated position of the vehicle in the previous cycle (i.e., one cycle before the current cycle) to the vehicle's current position 80 (i.e., in the vehicle's width direction), the control unit 10 calculates a distance W(0) from the vehicle's current position 80 to the wall 81 (this distance is perpendicular to the vehicle's trajectory). The detection of wall 81 and the calculation of the distance W to wall 81 are performed each control period. Thus, information about the presence or absence of wall 81 and the distance W to wall 81, obtained at the respective control times for up to at least N cycles prior, is cumulatively stored in the control unit 10. Based on the positions of wall 81 obtained at the respective control times for the N previous cycles, the control unit 10 calculates the positions of wall 81 at the respective control times for the previous N cycles by offsetting the position 80 of its own vehicle, obtained at the current control time, and the estimated positions of its own vehicle 40, obtained at the respective control times for the N cycles prior to the current cycle, in a direction towards wall 81 by a distance W from wall 81 at the corresponding control time. The control unit 10 calculates a wall history 82, which is a line connecting the positions in wall 81 at the respective control times for the previous N cycles. That is, the wall history 82 is a trajectory of the wall for a period from the current cycle to the cycle that lies N cycles before the current cycle. If wall history 82 is calculated in this way, the probability of existence of the adjacent lane in an area outside wall history 82 is determined to be zero. If target 73 is located in such an area outside wall history 82, the probability of existence of target 73 on the adjacent lane is accordingly determined to be zero. Although various methods are conceivable for determining the presence or absence of a target 73 located in an area outside the wall history 82, a simplified method similar to the method described above with reference to Fig. 5 can be used to determine the presence or absence of a target 73 in the area represented by the neighboring lane probability characteristic field. In the simplified method, as shown in Fig. 7, a vehicle width line 36 that is closest to the target 73 is selected from the vehicle width lines 36 that were calculated at the respective times for up to N previous cycles, and a perpendicular is drawn from the target 73 to the selected vehicle width line 36 to calculate a corresponding position 60 of the target where the perpendicular and the selected vehicle width line 36 intersect. If the corresponding position 60 of the target is located on the selected vehicle width line 36 and outside the wall history 82, the probability of existence of the target 73 on the adjacent lane is set to zero. If the corresponding position 60 of the target is located on the selected vehicle width line 36 and within the wall history 82, the probability of existence of the target 73 on the adjacent lane is determined according to the adjacent lane probability characteristic curve field, as described above with reference to Figures 4 and 5. This means that the probability of existence of target 73 on the adjacent lane is essentially determined on the basis of the adjacent lane probability curve field, as described with reference to Figures 4 and 5. However, if a wall 81 is detected and its wall history 82 is then calculated, the probability of existence of a target 73 outside the wall history 82 is determined to be zero, irrespective of the adjacent lane probability curve field. In the present embodiment, the vehicle-mounted system 1 corresponds to a neighboring lane monitoring device. Fig. 1B shows a functional block diagram representing functions of the control unit 10, which are executed by a processor, software, or a combination of both. The control unit 10 includes an odometry acquisition unit 101, a trajectory calculation unit 102, a neighboring lane estimation unit 103, a unit for determining another vehicle 106, a special processing unit 107, and a unit for determining the state of another vehicle 108. The odometry acquisition unit 101 performs the operation in step S120. The trajectory calculation unit 102 performs the operation in step S130. The neighboring lane estimation unit 103 performs the operation in step S140. The unit for determining another vehicle 106 performs the operations in steps S160 to S170.The special processing unit 107 performs the operation in step S190. The unit for determining the state of another vehicle 108 performs the operation in step S180. The neighboring track estimation unit 103 contains a probability distribution calculation unit 104 and a range estimation unit 105, each of which performs part of the operation of step S140. (4) Advantages With the vehicle-mounted system 1 of the present embodiment, the trajectory calculation unit 102 calculates the trajectory of the vehicle itself based on the odometry information obtained by the odometry acquisition unit 101. The adjacent lane estimation unit 103 calculates an area where an adjacent lane exists, based on the trajectory of the vehicle itself. The position of a target (for example, another vehicle) is detected by the radar unit 3, which acts as a detector for another vehicle. The other vehicle determination unit 106 determines whether another vehicle is present in the adjacent lane, based on the position of the target and the area where the adjacent lane exists. With this configuration, it is possible to accurately estimate whether another vehicle is present in the adjacent lane. In the present embodiment, estimated positions of the vehicle at the respective control times are calculated for N cycles prior to the current cycle. Based on these estimated positions, a vehicle trajectory is calculated. With this configuration, the vehicle's trajectory can be correctly calculated for up to N cycles prior. Taking into account error factors in the odometry information, an error variance is calculated for each of the estimated positions of the vehicle at the respective previous time points. The probability distribution of each estimated position of the vehicle (probability distribution of the vehicle) in the lane width direction is calculated by projecting the error variance of the estimated position of the vehicle onto the lane width direction. The probability distribution calculation unit 104 calculates the positions of lane divider lines defining the adjacent lane (the inner divider position 41 and the outer divider position 42) by shifting the estimated positions of the own vehicle at the respective previous control times in the vehicle width direction. The probability distribution calculation unit 104 then simply applies the probability distribution of the own vehicle to each of the positions of the lane divider lines defining the adjacent lane to calculate the probability distribution of each of the lane divider lines (the inner probability distribution 46 and the outer probability distribution 47), thereby calculating the probability distribution of the adjacent lane at each of the respective previous control times.Based on the probability distribution of the neighboring track at the respective previous control times, the area estimation unit 105 calculates the neighboring track probability characteristic field (a characteristic field that represents an area in which the probability of existence of the neighboring track assumes a special probability). This means that the area where the adjacent lane exists is calculated as a field of characteristic curves corresponding to the probability distribution, taking into account errors in the odometry information. Based on this field of neighboring lane probability characteristic curves, the vehicle detection unit 106 calculates a probability value to determine whether a target is present in the adjacent lane. Despite the presence of errors in the odometry information, the final determination of whether a target is present in the adjacent lane can be correctly performed based on the probability of the target's existence in the adjacent lane. In the present embodiment, even if it is determined that a target exists in the adjacent lane, no alarm is issued, depending on the positional relationship between the vehicle and the target. That is, if it is determined that a target exists in the adjacent lane and the positional relationship between the vehicle and the target fulfills a predetermined condition, an alarm is issued. With this configuration, an alarm can be issued at the appropriate time. Other embodiments One embodiment of the present invention has been described above. However, the present invention is not limited to the embodiment described above, but can be implemented in various ways. (1) In the embodiment above, the probability of existence of a target on the adjacent lane is calculated using the simplified method. Alternatively, the probability of existence of a target on the adjacent lane can be calculated using a different method. That is, in the embodiment above, a perpendicular is drawn from the target to the vehicle width line 36 that is closest to the target in order to calculate a corresponding position 60 of the target where the perpendicular and the selected vehicle width line 36 intersect. A probability of the corresponding position 60 of the target in the adjacent lane probability distribution is calculated as the probability of existence of the target on the adjacent lane. Alternatively, the probability of existence of the target on the adjacent lane can be calculated using a method other than the simplified method described above. (2) Alternatively, only the neighboring lane probability curves can be calculated according to a specific probability (for example, P1). The probability of the target existing on the neighboring lane can be calculated using these neighboring lane probability curves. (3) In the embodiment above, the odometry information used to calculate the vehicle's own trajectory includes a vehicle speed v, a yaw rate ω, and a turning radius R. However, these are only examples, and it is possible to use other odometry information as required. (4) If it is determined that a target is present on an adjacent lane, an alarm may be issued regardless of the distance to the target or the rate of reduction of the distance to the target.

Claims

Device (1) for monitoring an adjacent lane adjacent to a lane on which a own vehicle (71) is traveling, wherein the own vehicle (71) is a vehicle transporting the device (1), the device (1) comprising: a detector for another vehicle (3) that acquires position information of another vehicle (73) indicating a position of the other vehicle (73) around the own vehicle (71) in relation to the own vehicle (71); an odometry acquisition unit (101, S120) that acquires odometry information indicating a driving state of the own vehicle (71); a trajectory calculation unit (102, S130) that calculates a trajectory (31) of the own vehicle (71) using only the odometry information acquired by the odometry acquisition unit (101, S120); a neighboring lane estimation unit (103, S140) that, based on the Travel trajectory (31) of own vehicle (71),a neighboring lane area calculated by the trajectory calculation unit (102, S130), which estimates an area in which the neighboring lane is present; a unit for determining another vehicle (106, S160-S170), which, based on the position information of the other vehicle (73) acquired by the detector for another vehicle (3) and the neighboring lane area estimated by the neighboring lane estimation unit (103, S140), determines whether the other vehicle (73) is present in the neighboring lane; a probability distribution calculation unit (104, S140), which, based on error factors of the odometry information obtained by the odometry acquisition unit (101, S120), calculates a neighboring lane probability distribution (46, 47) that indicates a distribution of a probability of existence of the neighboring lane area; and an area estimation unit (105, S140),which, based on the neighboring lane probability distribution (46, 47) calculated by the probability distribution calculation unit (104, S140), estimates for each of at least one special probability value a range in which the probability of existence of the neighboring lane range is equal to the special probability value, as a special neighboring lane range corresponding to the special probability value, wherein the unit for determining another vehicle (106, S160-S170) determines in which range of at least one special neighboring lane range estimated by the range estimation unit (105, S140) and corresponding to the at least one special probability value, the position of the other vehicle (73), specified by the position information of the other vehicle (73) acquired by the detector for another vehicle (3),is present and, based on the determination result, determines whether the other vehicle (73) is present in the adjacent lane. Device (1) according to claim 1, wherein the probability distribution calculation unit (104, S140) calculates the neighboring lane probability distribution (46, 47) in a lane width direction of the neighboring lane. Device (1) according to claim 2, wherein the probability distribution calculation unit (104, S140) calculates the neighboring lane probability distribution (46, 47) as a probability distribution that corresponds to a normal probability distribution. Device (1) according to one of claims 1 to 3, wherein the vehicle trajectory calculation unit (102, S130) calculates past positions of the own vehicle (71) in relation to the current position of the own vehicle (71) at respective past control times of a predetermined control period for a predetermined number of control periods before the current control time and calculates the vehicle trajectory (31) of the own vehicle (71) on the basis of the calculated positions of the own vehicle (71) at the respective past control times and the current position of the own vehicle (71). Device (1) according to one of claims 1 to 4, which further comprises a special processing unit (107, S190) which, when the unit for determining another vehicle (106, S160-S170) determines that the other vehicle (73) is present in the adjacent lane, performs a special processing operation. Device (1) according to claim 5, which further comprises a unit for determining a state of another vehicle (108, S180), which, when the unit for determining another vehicle (106, S160-S170) determines that the other vehicle (73) is present in the adjacent lane, determines whether the other vehicle (73) is in a special state with respect to its own vehicle (71), wherein the special processing unit (107, S190) performs the special processing when the unit for determining another vehicle (106, S160-S170) determines that the other vehicle (73) is present in the adjacent lane, and when the unit for determining a state of another vehicle (108, S180) determines that the other vehicle (73) is in the special state. Device (1) according to claim 6, wherein the special state comprises at least one state in which a relative distance, which is a distance between the other vehicle (73) and the own vehicle (71), is equal to or less than a predetermined distance, and / or a state in which a reduction rate of the relative distance is equal to or greater than a predetermined reduction rate.

Images