Object detection device and object detection method

Abstract

To provide an inexpensive and accurate object detection device that can operate with a light computational load.SOLUTION: An object detection device includes: a transmission / reception unit which is installed on a moving body; a calculation unit which calculates at least a position of an object on the basis of an IF signal obtained by mixing a transmitted wave and a received wave; a tracking processing unit which repeats, by a prescribed repetition frequency M, a series of processing expandingly allocating a frame average position obtained by averaging the position of the object calculated by the calculation unit with a frame that is an integer N multiple of a chirp period T as unit time as the position of the object to a corresponding cell and its surrounding cells among the plurality of cells defined on a prescribed two-dimensional plane on which a three-dimensional space is projected; and an object determination unit which determines that the object has been detected when there is a cell, among the cells to which the frame average position has been allocated by the tracking processing unit, in which a frequency of overlapping allocation in the plurality of frames becomes equal to or greater than a prescribed threshold.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present invention relates to an object detection device and an object detection method using a radar device. [Background technology]

[0002] Mobile objects used outdoors, such as construction machinery, agricultural machinery, and AMRs (Autonomous Mobile Robots), are equipped with object detection devices to prevent collisions. These mobile objects must detect obstacles, such as people, in the direction of travel and issue a warning light to the driver to slow down or stop the mobile object. In addition, if the driver is too late in operating the mobile object and there is a risk of a collision, the mobile object must be stopped by automatically applying the brakes using a cruise control device.

[0003] Sensors used for this purpose include ultrasonic sensors, image sensors, laser sensors, radar sensors, etc., but when used outdoors, high weather resistance is required.

[0004] Image sensors and laser sensors have excellent resolution and accuracy, but are more likely to malfunction under harsh weather conditions such as sunlight, rain, snow, and fog. Ultrasonic sensors are inexpensive, but their detection range is short, making them difficult to use as obstacle sensors for moving objects. Radar sensors are superior in terms of weather resistance, but their object detection stability is low and their accuracy is poor.

[0005] For example, a radar sensor using millimeter waves can determine the position and relative speed of an obstacle in the direction of movement of a moving object based on an IF signal that is generated by mixing the transmitted wave and the reflected wave. However, there are cases in which an obstacle may disappear due to interference between the obstacle and the reflected waves from surrounding reflectors, or an obstacle may appear in a direction where it does not exist due to an error in the direction calculation for the reflected waves from a strong reflector in an area outside the detection area.

[0006] Patent Document 1 discloses an object detection device that can detect an object with high accuracy even when time-series radar reflection data is unstable. The object detection device groups reflection data detected by a radar detection means into segments and detects objects based on the reflection data for each segment, and is equipped with a grouping means for grouping the reflection data into segments based on predetermined conditions, a segment history identification means for identifying the identification information of the segment to which each reflection data belonged in the past, and a segment identification means for determining the identity of the currently grouped segment and a previously grouped segment based on the past segment identification information of the reflection data belonging to the currently grouped segment. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2010-266225 A Summary of the Invention [Problem to be solved by the invention]

[0008] However, in the above-mentioned object detection device, the computational load required for the process of assigning identification information to segments and determining their identity based on the identification information is heavy, and a high-performance arithmetic circuit using an expensive CPU is required. Therefore, there was a demand for an inexpensive, versatile object detection device that could be mounted on a moving object traveling at a relatively slow speed.

[0009] In view of the above-mentioned problems, an object of the present invention is to provide an inexpensive object detection device and object detection method that can be handled with a light computation load and have high accuracy. [Means for solving the problem]

[0010] In order to achieve the above object, a first characteristic configuration of an object detection device according to the present invention includes a transmitting / receiving unit that is installed on a moving body and includes a transmitting antenna that transmits a chirp signal into a three-dimensional space and a receiving antenna that receives the chirp signal reflected from an object, a calculation unit that calculates at least the position of the object based on an IF signal that is a mixture of the chirp signal transmitted from the transmitting antenna and the chirp signal received by the receiving antenna, and a calculation unit that calculates the position of the object based on an IF signal that is a mixture of the chirp signal transmitted from the transmitting antenna and the chirp signal received by the receiving antenna, and a frame average position that is an average of frames that are an integer multiple of a chirp period T per unit time (T·N) of the object position calculated by the calculation unit, and a frame average position that is an average of frames that are an integer multiple of a chirp period T per unit time (T·N) of the object position, and a plurality of cells partitioned on a predetermined two-dimensional plane onto which the three-dimensional space is projected. The system includes a tracking processing unit that repeats a series of processes of expanding and allocating the frame average position to the corresponding cell and its surrounding cells a predetermined number of times M, and an object determination unit that determines that the object has been detected when a cell to which the frame average position has been allocated by the tracking processing unit has been allocated more than a predetermined threshold number of times in multiple frames, wherein the size of the cells partitioned on the two-dimensional plane is set based on the maximum value L of the relative movement distance of the object in units of the frame and the measurement error δ of the object, and the expansion range of the cell to which the frame average position is allocated is set based on the number of times M.

[0011] The calculation unit calculates information for identifying an object, including the object's position in three-dimensional space, based on the IF signal for each chirp period. The tracking processing unit calculates a frame average position by averaging the object's positions obtained during each unit frame, which is an integer N times the chirp period T. Furthermore, for a number of cells that are pre-divided into a predetermined two-dimensional plane projected onto a horizontal plane, for example, the three-dimensional space is expanded to a cell corresponding to the object's frame average position and its surrounding cells, and the process of allocating, i.e., marking, the position where the object exists is repeated M times.

[0012] The object determination unit determines that an object has been detected when there is a cell that is allocated to each frame by the tracking processing unit and the number of times the cell is allocated in duplicate across multiple frames is equal to or exceeds a predetermined threshold value. This utilizes the principle that even if the disappearance of an object due to interference or the occurrence of a ghost due to a calculation error occurs consecutively within a unit frame, the probability of the occurrence of such a case in multiple consecutive frames is low.

[0013] Here, the size of the cells partitioned on a two-dimensional plane is set based on the maximum value L of the relative movement distance of the object in units of frames and the measurement error δ of the object caused by the transmitting / receiving unit, the calculation unit, etc., and the range of the cells to be assigned by expanding the frame average position is set based on the number of repetitions M, so that the object can be captured reliably.

[0014] The second characteristic configuration, in addition to the first characteristic configuration described above, further includes a judgment area setting unit that sets a judgment area that overlaps at least partially with a plurality of cells partitioned in the two-dimensional plane and projects a three-dimensional detection area defined in the three-dimensional space onto the two-dimensional plane, and the object judgment unit judges that an object is present in the three-dimensional detection area when there is a cell among the cells included in the judgment area for which the number of times the frame average position is assigned in a duplicated manner is equal to or greater than a predetermined threshold, and outputs an object judgment signal.

[0015] The determination area set by the determination area setting unit is an area obtained by projecting the 3D detection area defined in the 3D space onto a 2D plane, and is set as a range that includes at least a part of the cells in the 2D plane to which the tracking processing unit repeatedly assigns the frame average position. If there is a cell in such a determination area to which the number of times the frame average position is repeatedly assigned is equal to or exceeds a predetermined threshold, it is determined that an object exists in the 3D detection area.

[0016] The third characteristic configuration of the present invention is, in addition to the first characteristic configuration described above, in that the size of the cell is defined as a rectangular area with the length of one side being (L×M+2×δ) / M.

[0017] The product of the maximum value L of the object's relative movement distance in a unit frame and the number M of tracking processes plus the error 2δ when detecting the object is the range of the object's relative movement to be grasped in M ​​tracking processes. Therefore, by dividing this value by M, the size of the cell required for one tracking process can be found.

[0018] The fourth characteristic configuration of the present invention is, in addition to the first characteristic configuration described above, that the range of the cell to which the frame average position is assigned is set to a rectangular area obtained by extending the cell vertically and horizontally by the number of repetitions M.

[0019] In the tracking process, not only the cell corresponding to the frame average position, but also the surrounding cells are expanded and assigned as cells in which the object exists. If the maximum number of cells that an object moves to in a unit frame is set to 1, it is expected that the object will move a maximum of the number of cells corresponding to the number of iterations M of the tracking process. Therefore, if the cells assigned as the frame average position are set to a rectangular area expanded vertically and horizontally by the number of iterations M, the object can be captured reliably.

[0020] A fifth characteristic configuration of the same is that, in addition to any one of the first to fourth characteristic configurations described above, the chirp signal transmitted and received by the transceiver unit is a millimeter wave that employs the FMCW method.

[0021] Millimeter wave band radio waves have good directionality and are well suited for use in outdoor object detection.

[0022] A first characteristic configuration of the object detection method according to the present invention includes a calculation step of calculating at least a position of the object based on an IF signal obtained by mixing a chirp signal transmitted into a three-dimensional space from a transmitting antenna of a transmitting / receiving unit installed in a moving object and a chirp signal reflected from an object and received by a receiving antenna of the transmitting / receiving unit, and a frame average position obtained by averaging the object position calculated in the calculation step with frames each of which is an integer multiple of a chirp period T as a unit time (T·N), and expanding the frame average position to a corresponding cell and its surrounding cells among a plurality of cells partitioned on a predetermined two-dimensional plane onto which the three-dimensional space is projected. The method includes a tracking processing step of repeating a series of processes of allocating the frame average position a predetermined number of times M, and an object determination step of determining that the object has been detected when a cell to which the frame average position has been allocated in the tracking processing step has been allocated more than a predetermined threshold number of times in multiple frames, wherein the size of the cells partitioned on the two-dimensional plane is set based on the maximum value L of the relative movement distance of the object in units of the frame and the measurement error δ of the object, and the range of the cells to which the frame average position is allocated is set based on the number of repetitions M.

[0023] The second characteristic configuration, in addition to the first characteristic configuration described above, further includes a judgment area setting step for setting a judgment area that overlaps at least partially with a plurality of cells partitioned in the two-dimensional plane and is a three-dimensional detection area defined in the three-dimensional space projected onto the two-dimensional plane, and the object judgment step determines that an object is present in the three-dimensional detection area when there is a cell among the cells included in the judgment area whose number of times the frame average position is assigned in a duplicated manner is equal to or greater than a predetermined threshold, and outputs an object judgment signal.

[0024] The third characteristic configuration of the present invention is, in addition to the first characteristic configuration described above, in that the size of the cell is defined as a rectangular area with the length of one side being (L×M+2×δ) / M.

[0025] The fourth characteristic configuration of the present invention is, in addition to the first characteristic configuration described above, that the range of the cell to which the frame average position is assigned is set to a rectangular area obtained by extending the cell vertically and horizontally by the number of repetitions M.

[0026] A fifth characteristic configuration of the same is that, in addition to any one of the first to fourth characteristic configurations described above, the chirp signal transmitted and received by the transceiver unit is a millimeter wave that employs the FMCW method. Effect of the Invention

[0027] As described above, according to the present invention, it is possible to provide an inexpensive object detection device and object detection method that can handle a light computation load and have high accuracy. [Brief description of the drawings]

[0028] [Figure 1] FIG. 1A is an explanatory diagram of a transmitting / receiving unit and a calculation unit of an object detection device, and FIG. 1B is an explanatory diagram of object identification information including the position of an object calculated by the calculation unit. [Diagram 2] (a) is an explanatory diagram of the phenomenon in which an object disappears, and (b) is an explanatory diagram of the phenomenon in which a ghost occurs. [Diagram 3] (a) is an explanatory diagram of a transmitted wave (chirp signal), (b) is an explanatory diagram showing the phase shift between the transmitted wave and the received wave, (c) is an explanatory diagram of an IF signal resulting from mixing of the transmitted wave and the received wave, and (d) is an explanatory diagram of the chirp period T and the unit frame. [Figure 4] (a) is an explanatory diagram of an object detection device and a three-dimensional detection area, and (b) is an explanatory diagram of a judgment area, which is a two-dimensional monitoring area obtained by projecting the three-dimensional detection area onto a two-dimensional plane and dividing it into cells. [Diagram 5] (a) is an explanatory diagram of cells in which frame average positions are assigned to cells partitioned on a two-dimensional plane projected onto a three-dimensional space. The left is an explanatory diagram of a 3x3 expanded cell, and the right is an explanatory diagram of a 5x5 expanded cell. (b) is an explanatory diagram of tracking processing with four iterations. (c) is an explanatory diagram of the operation when the object determination unit does not detect an object. (d) is an explanatory diagram of the operation when the object determination unit detects an object. [Figure 6] Functional block diagram of the object detection device [Figure 7] FIG. 13 is an explanatory diagram showing the procedure of processing when an object is detected. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0029] Hereinafter, an embodiment of an object detection device and an object detection method according to the present invention will be described. The object detection device 1 is a millimeter wave radar device using the FMCW method that is installed on mobile objects used outdoors, such as construction machinery, agricultural machinery, and AMRs (Autonomous Mobile Robots), and functions as a collision avoidance device that detects other mobile objects or people in the direction of travel of the mobile object, enabling it to brake or stop the object in advance.

[0030] 6 shows functional blocks of the object detection device 1. The object detection device 1 includes a transceiver unit 10 having a transmitting antenna TxA and a receiving antenna RxA, a calculation unit 20, a tracking processing unit 30, and an object determination unit 40.

[0031] The transceiver unit 10 includes at least one transmitting antenna TxA and multiple receiving antennas RxA arranged vertically and horizontally, and includes a signal generator that generates a chirp signal as shown in Fig. 3(a). The chirp signal generated by the signal generator is transmitted from the transmitting antenna TxA into three-dimensional space, and the chirp signal reflected by an object 2 existing in the three-dimensional space is received by the receiving antenna RxA.

[0032] The calculation unit 20 is a circuit block that calculates object identification information including the position of the object 2 from the chirp signal transmitted from the transmitting antenna TxA and the chirp signal received by the receiving antenna RxA, and specifically includes a mixing circuit (mixer) that mixes the transmitted chirp signal and the received chirp signal, an ADC circuit that converts an intermediate frequency signal (hereinafter referred to as "IF signal") generated by the mixing circuit into a digital signal, and an FFT circuit that performs FFT calculation on the digital signal output from the ADC circuit, as shown in Fig. 1(a). As shown in Fig. 1(b), the object identification information output from the calculation unit 20 includes the distance r from the object detection device 1 to the object 2, the angle θ with the X-axis in the XY plane, the angle φ with the Z-axis, the relative speed V between the object detection device 1 and the object 2, the reflection intensity I, etc.

[0033] In this embodiment, the 79 GHz band with a wavelength of approximately 4 mm is used, and a chirp signal that is frequency modulated so that the wavelength changes linearly between 77.0 and 81.0 GHz is repeatedly output by a signal generator (see FIG. 3(a)).

[0034] As shown in Figures 3(b) and (c), a signal Tx transmitted towards an object 2 at a distance r is delayed by a signal Rx reflected from the object 2 in a manner proportional to the distance. The IF frequency, which is the difference between the transmitted and received frequencies, is low for nearby objects and high for distant objects; performing an FFT on the IF signal converts it from the time domain to the frequency domain, and distance information r according to the frequency is obtained. Velocity information v is obtained by performing an FFT on the distance information for multiple repeated chirp signals. Angle information θ, φ is obtained by performing an FFT on the velocity information between multiple receiving antennas.

[0035] As shown in Fig. 2(a), if there is another object with the same reflection characteristics near the person to be detected, and the shape of the reflected wave from the person is the same as that of the other object, but the phase is 180° different, the signal returning to the receiving antenna RxA will interfere with and be lost. Objects with the same reflection characteristics as people include the ground, walls made of earth, and trees, and it is impossible to avoid them, so the received signal is often lost.

[0036] Also, as shown in Figure 2(b), when a reflected wave from an object with strong reflection characteristics, such as a car body or grating (ditch cover), enters from outside the 3D detection area, the strong signal entering the receiving antenna RxA causes the signal to be oversaturated, resulting in an error in the phase difference calculation with the adjacent receiving antenna RxA, making the value of the angle θ unstable, and it may be determined that an object with high reflection characteristics outside the 3D detection area is present within the 3D detection area. A similar phenomenon may also cause an error in the phase difference calculation when internal reflection occurs within the antenna case (radome). Such a signal is called a ghost signal, and there is a risk that it may be mistakenly determined that an object is present.

[0037] In the present invention, by providing the tracking processing unit 30 and the object determination unit 40, it is possible to suppress the influence of loss of received signals and ghost signals, and to accurately determine the presence of an object.

[0038] The tracking processing unit 30 repeatedly calculates the frame average position as the object position by averaging the object position calculated by the calculation unit 20 over a unit time (T·N) of frames that are an integer N times the chirp period T (see FIG. 3(d)), and repeats this series of processes a predetermined number of times M to expand and assign the average frame position to corresponding cells and their surrounding cells among a plurality of cells partitioned into a predetermined two-dimensional plane onto which three-dimensional space is projected, which in this embodiment is a horizontal plane.

[0039] For example, suppose that a rectangular parallelepiped region A defined in a part of the XYZ three-dimensional space shown in Figs. 4(a) and (b) is projected onto the XY plane, the projected planar region A' is divided into a number of regions, and each division is called a cell C. In this case, each object (three objects in this example) detected inside the region A is projected onto the planar region A' and assigned to the corresponding cell C (the cell C marked with a black circle in Fig. 4(b) is the assigned cell C). In this type of processing, when the frame average positions are different between adjacent frames, in order to determine whether or not the objects are the same, it is necessary to calculate the distance between them and determine that objects shorter than a predetermined distance are the same. When a large number of object points are detected in a unit frame, the computation load increases significantly, and a computation circuit equipped with an expensive CPU is required.

[0040] Therefore, the size of cell C is set based on the maximum value L of the relative movement distance of the object in units of a frame and the measurement error δ of the object, and the frame average position calculated in unit frame is expanded and assigned to the corresponding cell as well as to the surrounding cells, including the error, thereby making it possible to significantly reduce the computational load required to determine that they are the same object.

[0041] The example shown on the left of Figure 5(a) is an example in which the frame average position has been extended to eight surrounding cells (white circles in the figure) around the cell to which the frame average position has been assigned (black circle in the figure), and the example shown on the right of Figure 5(a) is an example in which the frame average position has been extended to 24 surrounding cells (white circles in the figure) around the cell to which the frame average position has been assigned (black circle in the figure).

[0042] FIG. 5(b) shows the state in which the frame average position assigned to each cell changes from left to right when the number of repetitions M is 4 in the example shown on the left of FIG. 5(a). The third allocation process shows that the object has disappeared. If the number of times that the frame average position has been assigned to each cell is added up over four iterations of the process, the result is as shown in FIG. 5(c). For example, if the threshold for the number of allocations to a cell is set to 3, it can be determined that an object exists if there is a cell with an allocation number of 3 or more.

[0043] In other words, the object determination unit 40 is configured to determine that an object has been detected when there is a cell among the cells to which the frame average position has been assigned by the tracking processing unit 30, the number of times that the cell has been assigned in duplicate across multiple frames being equal to or greater than a predetermined threshold value.

[0044] The object determination unit 40 is equipped with a determination area setting unit which sets a determination area JA (output ON area) which overlaps at least partially with a plurality of cells partitioned on a two-dimensional plane and which projects a three-dimensional detection area defined in three-dimensional space onto the two-dimensional plane. The object determination unit 40 is configured to determine that an object is present in the three-dimensional detection area when there is a cell among the cells included in the determination area JA (output ON area) whose frame average position is assigned in a duplicated manner the number of times equal to or greater than a predetermined threshold value (3 in this example), and to output an object determination signal (output ON) to a driving control device of the mobile body, etc.

[0045] In Figure 5(c), there are cells outside the judgment area (output OFF area) whose allocation count is greater than or equal to the threshold value of 3, but there are no cells in the judgment area JA (output ON area) whose allocation count is greater than or equal to the threshold value of 3, so the object judgment signal is turned off (output OFF).In Figure 5(d), there are cells in the judgment area JA (output ON area) whose allocation count is greater than or equal to the threshold value of 3, so the object judgment signal is turned on (output ON).

[0046] The size of cell C is set based on the maximum value L of the relative movement distance of the object in units of a frame and the measurement error δ of the object, and the expansion range of the cell to which the frame average position is assigned is preferably set based on the number of repetitions M. By setting the maximum value of the distance that can be moved in a unit frame as the size of cell C, it is possible to capture the object in units of a cell, and by setting the expansion range of the cell based on the number of repetitions M, it is possible to determine the identity of the objects without calculating the distance between the objects.

[0047] The size of cell C is preferably defined as a rectangular area with a side length of (L×M+2×δ) / M. For example, if the chirp period T is 100 μs and the value of N is 500, the unit frame (T·N) is 50 ms. If the moving object is an agricultural or construction machine, and the moving object speed is assumed to be 10 km / h, the walking speed of a person to be 1.6 m / s, and the measurement error δ to be ±100 mm, then the maximum value L of the relative distance between the moving object and a person in the unit frame is 218 mm. In other words, the value of L×M+2×δ is 1290 mm, and the cell size is 258 mm.

[0048] In the above explanation, the range of the cell to which the frame average position is assigned is set to a rectangular area expanded vertically and horizontally by the number of repetitions M of the cell, but it is sufficient that the size in at least the direction along the moving body's direction of travel is set to be the length of the cell to which the frame average position is assigned, and the size in the direction perpendicular to the moving body's direction of travel may be shorter than that value.

[0049] FIG. 7 shows an object detection method executed by the object detection device 1 described above. Transmission / reception processing and calculation processing are executed for each chirp period (SA1), and frame average position calculation processing is executed for each frame (SA2). Then, allocation processing (tracking processing) of the frame average positions to cells partitioned on a two-dimensional plane is repeated for M frames (SA4), and object determination processing is executed based on the results of the allocation processing for M frames (SA5). If it is determined that an object is present in the determination area (SA7, Y), it is determined that an object is present in the three-dimensional detection range (SA8), and the object detection signal is turned on (SA9). Note that if no object is detected in step SA6 and the object is outside the determination area in step SA7, the object detection signal is turned off.

[0050] As described above, the object detection method according to the present invention includes a calculation step of calculating at least the position of an object based on an IF signal obtained by mixing a chirp signal transmitted into three-dimensional space from a transmitting antenna of a transmitting / receiving unit installed in a moving object and a chirp signal reflected from an object and received by the receiving antenna of the transmitting / receiving unit; a tracking processing step of repeating a series of processes a predetermined number of times M of repetitions, in which a frame average position obtained by averaging the object position calculated in the calculation step with frames each of which is an integer N times the chirp period T as a unit time (T·N), is taken as the position of the object, and the process of extending and allocating the frame average position to corresponding cells and their surrounding cells among a plurality of cells partitioned on a predetermined two-dimensional plane onto which the three-dimensional space is projected, and an object determination step of determining that an object has been detected when there is a cell among the cells to which the frame average position has been assigned in the tracking processing step, the number of times that the cell has been assigned in multiple frames being equal to or greater than a predetermined threshold, the size of the cells partitioned on the two-dimensional plane being set based on the maximum value L of the relative movement distance of the object in units of frames and the measurement error δ of the object, and the range of the cells to which the frame average position is assigned is set based on the number of repetitions M.

[0051] The method further includes a judgment area setting step for setting a judgment area by projecting a three-dimensional detection area defined in three-dimensional space onto the two-dimensional plane, the three-dimensional detection area being at least partially overlapping with a plurality of cells partitioned on the two-dimensional plane, and the object judgment step judges that an object is present in the three-dimensional detection area when there is a cell among the cells included in the judgment area whose frame average position is assigned in a duplicated manner the number of times that is equal to or greater than a predetermined threshold, and outputs a judgment signal.

[0052] Moreover, the size of a cell is defined as a rectangular area with a side length of (L×M+2×δ) / M.

[0053] Furthermore, the range of the cell to which the frame average position is assigned is set to a rectangular area obtained by extending the cell vertically and horizontally by the number of repetitions M.

[0054] The object detection device and method described above have been described as an example in which the chirp signal transmitted and received by the transmitter / receiver is a millimeter wave employing the FMCW method, but it may also be possible to use bands other than millimeter waves, such as microwaves or submillimeter waves.

[0055] The above-described embodiment is merely one example of the present invention, and the specific configuration of each part can be appropriately modified and designed within the scope in which the effects of the present invention are achieved. [Explanation of symbols]

[0056] 1: Object detection device 2:Object 10: Transmitter / receiver 20: Arithmetic section 30: Tracking processing unit 40: Object determination section

Claims

1. A transceiver unit is provided on the mobile object, the transceiver unit having a transmitting antenna for transmitting a chirp signal into a three-dimensional space and a receiving antenna for receiving the chirp signal reflected from an object; a calculation unit that calculates at least a position of the object based on an IF signal obtained by mixing a chirp signal transmitted from the transmitting antenna and a chirp signal received by the receiving antenna; a tracking processing unit which repeats a series of processes for repeating a predetermined number M of times, the series of processes being the object position, in which the object position calculated by the calculation unit is averaged over a unit time (T·N) of frames each of which is an integer N times the chirp period T, and the series of processes being the object position, and which expands and assigns the object position to a corresponding cell and its surrounding cells among a plurality of cells partitioned on a predetermined two-dimensional plane onto which the three-dimensional space is projected; an object determination unit that determines that the object has been detected when a cell exists among the cells to which the frame average position has been assigned by the tracking processing unit, the cell being assigned in duplicate in a plurality of frames a number of times equal to or greater than a predetermined threshold; Equipped with An object detection device in which the size of the cells partitioned on the two-dimensional plane is set based on the maximum value L of the relative movement distance of the object in units of the frame and the measurement error δ of the object, and the expansion range of the cell to which the frame average position is assigned is set based on the number of repetitions M.

2. a determination area setting unit that sets a determination area by projecting a three-dimensional detection area defined in the three-dimensional space onto the two-dimensional plane, the determination area at least partially overlapping with a plurality of cells defined on the two-dimensional plane, 2. The object detection device according to claim 1, wherein the object determination unit determines that an object is present in the three-dimensional detection area when a cell among the cells included in the determination area is assigned the frame average position in a duplicated manner more than a predetermined threshold, and outputs an object determination signal.

3. The object detection device according to claim 1 , wherein the size of the cell is defined as a rectangular area with a side length of (L×M+2×δ) / M.

4. 2. The object detection device according to claim 1, wherein the range of the cell to which the frame average position is assigned is set to a rectangular area obtained by extending the cell vertically and horizontally by the number of repetitions M.

5. 5. The object detection device according to claim 1, wherein the chirp signal transmitted and received by the transmitting and receiving unit is a millimeter wave that employs an FMCW system.

6. a calculation step of calculating at least the position of the object based on an IF signal obtained by mixing a chirp signal transmitted into three-dimensional space from a transmitting antenna of a transmitting / receiving unit installed in a moving object and a chirp signal reflected from the object and received by a receiving antenna of the transmitting / receiving unit; a tracking processing step in which the object position calculated in the calculation step is averaged over a unit time (T·N) of frames that are an integer N times the chirp period T to obtain a frame average position, and the frame average position is set as the object position, and a series of processes is repeated a predetermined number of times M to expand and assign the object position to a corresponding cell and its surrounding cells among a plurality of cells partitioned on a predetermined two-dimensional plane onto which the three-dimensional space is projected; an object determination step of determining that the object has been detected when a cell exists among the cells to which the frame average position has been assigned in the tracking processing step, the number of times the cell has been assigned in a plurality of frames being equal to or greater than a predetermined threshold; Equipped with The size of the cells partitioned on the two-dimensional plane is set based on the maximum value L of the relative movement distance of the object in units of the frame and the measurement error δ of the object, and the range of the cell to which the frame average position is assigned is set based on the number of repetitions M.

7. a determination area setting step of setting a determination area by projecting a three-dimensional detection area defined in the three-dimensional space onto the two-dimensional plane, the determination area at least partially overlapping with a plurality of cells defined on the two-dimensional plane, 7. The object detection method according to claim 6, wherein the object determination step determines that an object is present in the three-dimensional detection area when a cell among the cells included in the determination area is assigned the frame average position multiple times more than a predetermined threshold, and outputs an object determination signal.

8. The object detection method according to claim 6 , wherein the size of the cell is defined as a rectangular region with a side length of (L×M+2×δ) / M.

9. 2. The object detection method according to claim 1, wherein the range of the cell to which the frame average position is assigned is set to a rectangular area obtained by extending the cell vertically and horizontally by the number of repetitions M.

10. 10. The object detection method according to claim 6, wherein the chirp signal transmitted and received by the transmitting and receiving unit is a millimeter wave that employs an FMCW system.

Images