Method and apparatus for tracking an object using LIDAR sensor, vehicle including the apparatus, and recording medium for recording program performing the method
The LiDAR sensor-based object tracking method addresses synchronization issues by selecting a main sensor and calculating time differences, ensuring accurate attribute information acquisition for enhanced vehicle performance.
Patent Information
- Application Number
- KR1020210030786
- Authority / Receiving Office
- KR · KR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-03-09
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2041-03-09
AI Technical Summary
Inaccurate information acquisition by multiple LiDAR sensors in autonomous driving systems due to timing discrepancies in object detection and point cloud data synchronization, leading to unreliable vehicle performance.
An object tracking method using LiDAR sensors that involves checking update times, selecting a main sensor based on point distribution, and calculating time differences to accurately obtain attribute information such as speed and shape.
Ensures accurate acquisition of object attributes like speed and shape even when synchronization is off, enhancing vehicle reliability.
Smart Images

Figure 112021027663157-PAT00003_ABST
Abstract
Description
Technology Field
[0001] The embodiments relate to an object tracking method and apparatus using a LiDAR sensor, a vehicle including the apparatus, and a recording medium recording a program for executing the method. Background Technology
[0002] For example, a vehicle's Highway Driving Pilot (HDP) system provides a function that maintains the vehicle's speed according to the conditions set by the driver, without the driver having to press the accelerator or brake pedals, once the driver sets the vehicle's speed and the distance to the preceding vehicle in their lane. Information about target vehicles can be acquired using LIDAR (Light Detection and Ranging) sensors, and the HDP function can be performed using this acquired information. However, if the information about the target vehicle acquired using the LIDAR sensor is inaccurate, the reliability of the vehicle may be compromised, such as by the HDP function being performed incorrectly.
[0003] In particular, when using multiple LiDAR sensors for autonomous driving, the timing of object detection and point cloud data acquisition and update by the multiple LiDAR sensors may not coincide, or there may be a lack of synchronization between the timing of point cloud data signal processing and the LiDAR sensor update timing. In this case, even though the same object is detected, the location of the point cloud data acquired by the multiple LiDAR sensors may differ from one another. The problem to be solved
[0004] The embodiment provides an object tracking method and apparatus using a LiDAR sensor capable of accurately obtaining attribute information of a track, a vehicle including the apparatus, and a recording medium recording a program for executing the method.
[0005] The technical problems to be solved in the embodiments are not limited to those mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art to which the present invention belongs from the description below. means of solving the problem
[0006] An object tracking method using a LiDAR sensor according to one embodiment may include: a step of checking whether the update times of a plurality of LiDAR sensors are different based on a recognition time point at which signal processing of point cloud data begins at each step; a step of selecting a main sensor among the plurality of LiDAR sensors at each step if the update times are different; a step of obtaining the time difference between the update time of the main sensor selected at the previous (t-1) step and the update time of the main sensor selected at the current (t) step as the step time difference at the current step; and a step of obtaining the attribute information of the track at the current step using the step time difference when signal processing the point cloud data acquired from the main sensor selected at the current step.
[0007] For example, the step of checking whether the update times are different may include: a step of recording the update times for each of the plurality of lidar sensors; a step of calculating the time difference between the recorded update times and the recognition times as an asynchronous time for each of the plurality of lidar sensors; and a step of checking whether the asynchronous time is different for each of the plurality of lidar sensors.
[0008] For example, the step of selecting the main sensor may include: counting the number of points distributed on the collision reference plane of the contour formed by the point cloud data for each of the plurality of lidar sensors; and selecting the sensor with the largest number of counted points among the plurality of lidar sensors as the main sensor.
[0009] For example, the step time difference in the current step above may correspond to the time difference between the latest update time of the main sensor selected in the previous (t-1) step and the latest update time of the main sensor selected in the current (t) step.
[0010] For example, the update cycle for sensing the point cloud data and the signal processing cycle for the point cloud data for each of the plurality of lidar sensors may not be multiples of each other.
[0011] For example, the step of obtaining the attribute information may use the step time difference to obtain at least one of the speed, heading information, or shape of the track as the attribute information.
[0012] An object tracking method using a LiDAR sensor according to another embodiment may include: a step of checking whether the number of updates for a plurality of LiDAR sensors is different at each step; a step of selecting a main sensor among the plurality of LiDAR sensors at each step if the number of updates is different; a step of obtaining the time difference between the update time of the main sensor selected at the previous (t-1) step and the update time of the main sensor selected at the current (t) step as the step time difference at the current step; and a step of obtaining the attribute information of the track at the current step using the step time difference when signal processing point cloud data acquired from the main sensor selected at the current step.
[0013] An object tracking device using a LiDAR sensor according to another embodiment may include: a sensing time point inspection unit that, at every step, checks whether the update times for a plurality of LiDAR sensors are different based on a recognition time point at which signal processing of point cloud data begins; a sensor selection unit that, in response to the result inspected by the sensing time point inspection unit, selects a main sensor among the plurality of LiDAR sensors at every step; a step time difference calculation unit that calculates the time difference between the update time of the main sensor selected at the previous (t-1) step and the update time of the main sensor selected at the current (t) step as the step time difference at the current step; and an attribute information generation unit that generates the attribute information of the track at the current step using the step time difference when signal processing point cloud data acquired from the main sensor selected at the current step.
[0014] For example, the sensing time inspection unit may include a time recording unit that records the update time for each of the plurality of lidar sensors; an asynchronous time calculation unit that calculates the time difference between the recorded update time and the recognition time as an asynchronous time for each of the plurality of lidar sensors; and an asynchronous time inspection unit that checks whether the asynchronous times are different for each of the plurality of lidar sensors.
[0015] For example, the sensor selection unit may include a counter that counts the number of points distributed on the collision reference plane of the contour formed by the point cloud data for each of the plurality of lidar sensors; and a count comparison unit that selects the sensor with the largest number of counts among the plurality of lidar sensors as the main sensor.
[0016] For example, the step time difference calculation unit can calculate the time difference between the latest update time of the main sensor selected in the previous (t-1) step and the latest update time of the main sensor selected in the current (t) step as the step time difference.
[0017] For example, the update cycle for sensing the point cloud data and the signal processing cycle for the point cloud data for each of the plurality of lidar sensors may not be multiples of each other.
[0018] For example, the attribute information generating unit can generate at least one of the speed, heading information, or shape of the track as attribute information using the step time difference.
[0019] A vehicle according to another embodiment may include a lidar sensor; and the aforementioned object tracking device utilizing the lidar sensor.
[0020] According to another embodiment, a recording medium storing a program for executing an object tracking method using a LiDAR sensor for tracking an object stores a program that implements, at each step, a function of checking whether the update times for a plurality of LiDAR sensors are different based on a recognition time point at which signal processing of point cloud data begins; a function of selecting a main sensor among the plurality of LiDAR sensors at each step if the update times are different; a function of calculating the time difference between the update time of the main sensor selected at the previous (t-1) step and the update time of the main sensor selected at the current (t) step as the step time difference at the current step; and a function of calculating the attribute information of the track at the current step using the step time difference when signal processing point cloud data acquired from the main sensor selected at the current step, and can be read by a computer.
[0021] For example, the function of checking whether the update times are different may include: a function of recording the update times for each of the plurality of lidar sensors; a function of obtaining the time difference between the recorded update times and the recognition times as an asynchronous time for each of the plurality of lidar sensors; and a function of checking whether the asynchronous time is different for each of the plurality of lidar sensors.
[0022] For example, the function of selecting the main sensor may include a function of counting the number of points distributed on the collision reference plane of the contour formed by the point cloud data for each of the plurality of lidar sensors; and a function of selecting the sensor with the largest number of counted points among the plurality of lidar sensors as the main sensor. Effects of the invention
[0023] The object tracking method and apparatus using a lidar sensor according to an embodiment, a vehicle including the apparatus, and a recording medium storing a program for executing the method can accurately obtain attribute information, such as speed, heading, and shape information, even when the synchronization of multiple lidar sensors is not correct, such as when the update times of multiple lidar sensors are different and the update period and the recognition time period are not multiples of each other, or when the update times are the same but the update period and the recognition time period are not multiples of each other, or when the update period and the recognition time period are multiples of each other but the update times are different.
[0024] In addition, the effects obtainable in this embodiment are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below. Brief explanation of the drawing
[0025] FIG. 1 is a flowchart for explaining an object tracking method using a LiDAR sensor according to one embodiment. FIG. 2 is a flowchart for explaining an object tracking method using a LiDAR sensor according to another embodiment. FIG. 3 is a block diagram illustrating an object tracking device using a LiDAR sensor according to an embodiment. Figure 4 shows a timing diagram to aid in understanding the object tracking method illustrated in Figures 1 and 2. FIG. 5 shows a flowchart according to an embodiment of step 110A shown in FIG. 1. FIG. 6 shows a flowchart according to an embodiment of the 120th step illustrated in FIG. 1 and FIG. 2. Figure 7 is a diagram to aid in understanding Step 120 or Step 120A. Figures 8 (a) to (c) are diagrams illustrating the process of selecting a main sensor for each step. Figure 9 is a timing diagram showing an example of the timing diagram illustrated in Figure 4. FIG. 10 shows a block diagram of an object tracking device using a LiDAR sensor according to an embodiment. FIG. 11 is a block diagram according to one embodiment of the sensing time point inspection unit illustrated in FIG. 10. FIG. 12 shows a block diagram according to an embodiment of the sensor selection unit illustrated in FIG. 10. Figures 13 (a) to (c) are drawings to explain the problems of the object tracking method according to the comparative example. FIGS. 14 (a) to (c) are drawings showing point cloud data acquired from the first and second lidar sensors in the step, respectively. Specific details for implementing the invention
[0026] Hereinafter, to specifically explain the present invention, embodiments will be described, and to aid in understanding the invention, the invention will be described in detail with reference to the accompanying drawings. However, embodiments according to the present invention may be modified in various different forms, and the scope of the present invention should not be interpreted as being limited to the embodiments described below. The embodiments of the present invention are provided to more completely explain the present invention to those with average knowledge in the art.
[0027] In the description of the present embodiment, where it is stated that each element is formed "on or under," the "on or under" includes both cases where two elements are in direct contact with each other and one or more other elements are formed indirectly between the two elements.
[0028] In addition, when expressed as "up" or "down" (on or under), it can include the meaning of a downward direction as well as an upward direction based on a single element.
[0029] Additionally, relational terms used below, such as "first" and "second," "upper / superior / above," and "lower / subordinate / below," may be used to distinguish one entity or element from another, without necessarily requiring or implying any physical or logical relationship or order between such entities or elements.
[0030] Hereinafter, an object tracking method (100A, 100B) and a device (600, 800) using a LiDAR sensor according to an embodiment, a vehicle (1000) including the device, and a recording medium recording a program for executing the method (100A, 100B) are described as follows with reference to the attached drawings.
[0031] FIG. 1 is a flowchart for explaining an object tracking method (100A) using a lidar sensor according to one embodiment, FIG. 2 is a flowchart for explaining an object tracking method (100B) using a lidar sensor according to another embodiment, and FIG. 3 is a block diagram for explaining an object tracking device (600) using a lidar sensor (500) according to an embodiment.
[0032] Before describing the object tracking method (100A, 100B) illustrated in FIG. 1 and 2, the configuration and operation of the object tracking device (600) illustrated in FIG. 3 are described as follows.
[0033] An object tracking device (600) using a LiDAR sensor (500) illustrated in FIG. 3 may include a preprocessing unit (610), a clustering unit (620), a shape analysis unit (or, a segment) unit (630), and an object tracking unit (or, a tracking unit, a tracking and classification unit, or an object detection unit) (640). Alternatively, the preprocessing unit (610) may be omitted in the object tracking device (600). Additionally, a vehicle (1000) according to an embodiment may include a LiDAR sensor (500), an object tracking device (600), and a vehicle device (700).
[0034] A lidar sensor (500) can sense information about an object, such as the distance from the lidar sensor (500) to the object, the direction of the object, speed, temperature, material distribution and concentration characteristics, etc., by irradiating an object with a single circular laser pulse having a wavelength of, for example, 905 nm to 1550 nm, and measuring the time it takes for a laser pulse reflected from an object within the measurement range to return. Here, the object may be another vehicle, person, object, etc. existing outside the vehicle (hereinafter referred to as 'self-vehicle') (1000) equipped with the lidar sensor (500), but the embodiment is not limited to a specific type of object.
[0035] The lidar sensor (500) may include a transmitter (not shown) that transmits laser pulses and a receiver (not shown) that receives lasers reflected back from the surface of an object within the sensor range. The receiver has a field of view (FOV), which is an area that the lidar sensor (500) can observe at once without moving or rotating.
[0036] The LiDAR sensor (500) has a higher detection accuracy in the longitudinal and lateral directions compared to the RaDAR (Radio Detecting And Ranging) sensor, so it can provide accurate longitudinal and lateral position information, making it easy to use for obstacle detection and vehicle position recognition. As for the LiDAR sensor (500), there are 2D LiDAR sensors and 3D LiDAR sensors. The 2D LiDAR sensor is configured to be capable of tilting or rotating, and can be used to obtain LiDAR data containing 3D information by tilting or rotating. The 3D LiDAR sensor can obtain multiple points in 3D, allowing for the prediction of obstacle height information, which helps in accurate and detailed object detection or tracking. The 3D LiDAR sensor may also generate LiDAR data containing 3D information by configuring the 2D LiDAR sensor into multiple layers.
[0037] The LiDAR sensor (500) outputs point cloud data consisting of multiple points (or points) for a single object.
[0038] The object tracking method (100A, 100B) and device (600) according to the embodiment are not limited to a specific form, location, and type of lidar sensor (500).
[0039] Additionally, according to the embodiment, the lidar sensor (500) may be a plurality of units located at least two locations among the front, side, rear, or top of the vehicle (1000), and the object tracking method (100A, 100B) and device (600) according to the embodiment and the vehicle (1000) including the same are not limited to a specific location where the lidar sensor (500) is mounted on the vehicle (1000).
[0040] Meanwhile, the object tracking device (600) receives point cloud data and uses it to check for the existence of an object, starts, maintains, or stops tracking the object, updates, stores, or deletes information about the object, and further classifies the type of object.
[0041] The preprocessing unit (610) can preprocess point cloud data. To do this, the preprocessing unit (610) can perform calibration to align the coordinates between the LiDAR sensor (500) and the vehicle (1000). That is, the point cloud data can be converted to fit a reference coordinate system according to the position angle at which the LiDAR sensor (500) is mounted on the vehicle (1000). In addition, the preprocessing unit (610) can remove points with low intensity or reflectivity through filtering using the intensity or confidence information of the point cloud data.
[0042] Additionally, the preprocessing unit (610) may remove data reflected by the vehicle body (1000). That is, since there is an area obscured by the vehicle body (1000) depending on the mounting position and viewing angle of the lidar sensor (500), the preprocessing unit (610) may remove data reflected from the vehicle body (1000) using a reference coordinate system.
[0043] The clustering unit (620) groups point cloud data, which consists of multiple points of an object acquired through the LiDAR sensor (500), into meaningful units according to a predetermined rule. If the preprocessing unit (610) is not omitted, the clustering unit (620) can group the point cloud data preprocessed by the preprocessing unit (610). For example, the clustering unit (620) can group the point cloud data by applying vehicle modeling or guardrail modeling, etc., to cluster the outer shape of the object. The result sensed by the LiDAR sensor (500) is multiple points, and each point has only information about its location. Therefore, the clustering unit (620) plays the role of grouping the multiple points sensed by the LiDAR sensor (500) into meaningful shape units.
[0044] For example, as types of clustering units (620), there are 2D clustering and 3D clustering. 2D clustering is performing clustering by projecting data onto an XY plane without considering height information, and clustering can be performed on a point-by-point basis or in a specific structured unit basis. 3D clustering means performing clustering on an XYZ plane while considering all height information (Z).
[0045] The shape analysis unit (630) generates information about a plurality of segment boxes per channel using the clustered results from the clustering unit (620). Here, a segment box may refer to a result obtained by converting the clustered results into a geometric box shape. Additionally, information about a segment box may refer to at least one of the width, length, position, or direction (or heading) of the segment box.
[0046] The description of the object tracking unit (640) according to the embodiment described below is not limited to the presence or absence of the preprocessing unit (610) or the specific operational forms performed in each of the preprocessing unit (610), the clustering unit (620), and the shape analysis unit (630). That is, the object tracking unit (640) according to the embodiment can be applied even if the preprocessing unit (610) is omitted, if the point cloud data is processed differently from what was previously described in the preprocessing unit (610), if the point cloud data is clustered differently from what was previously described in the clustering unit (620), or if the shape analysis unit (630) generates segment box information differently from what was previously described.
[0047] The object tracking unit (640) can select a segment box (or, final segment box, or 'associated segment box') that is currently associated with the object being tracked (hereinafter referred to as the 'target object') from among a plurality of segment boxes per channel. Here, 'association' refers to the process of selecting a box to be used to maintain tracking of the target object currently being tracked from among the information of the plurality of segment boxes, where multiple segment boxes may be acquired for the same object depending on the visibility of the LiDAR sensor (500) and the shape of the object. Such association can be performed every cycle.
[0048] The object tracking unit (640) can select a linked segment box from each of the multiple segment boxes provided by the shape analysis unit (630) for each channel by converting the information of each of the multiple segment boxes into a predetermined format and selecting a linked segment box from among the multiple segment boxes (or segment boxes of meta objects) having the converted format.
[0049] Additionally, the object tracking unit (640) can distinguish whether the target object is an obstacle, a vehicle, or a person by using information about the associated segment box.
[0050] For convenience of explanation, the object tracking method (100A, 100B) illustrated in FIGS. 1 and 2 is described as being performed in the object tracking unit (640) of the object tracking device (600) illustrated in FIG. 3, but the embodiment is not limited thereto. That is, according to other embodiments, the object tracking method (100A, 100B) illustrated in FIGS. 1 and 2 may also be performed in other components included in the object tracking device (600) in addition to the object tracking unit (640) illustrated in FIG. 3. Furthermore, the object tracking unit (640) illustrated in FIG. 3 is described as performing the object tracking method (100A, 100B) illustrated in FIG. 1 and 2, but the embodiment is not limited thereto.
[0051] According to an embodiment, the object tracking unit (640) can obtain attribute information of the track by signal processing point cloud data obtained from each of the plurality of LiDAR sensors (500). The object tracking unit (640) can perform signal processing to obtain such attribute information every cycle. Hereinafter, the cycle for obtaining attribute information is referred to as a 'step,' and the time at which signal processing for the point cloud data begins is referred to as a 'recognition time.'
[0052] According to the object tracking method (100A) of one embodiment, at each step, it is checked whether the update time is different for each of the plurality of LiDAR sensors (500) based on the recognition time (110A). Here, the 'update time' may mean the time when the LiDAR sensor (500) acquires point cloud data, the time when the acquired point cloud data is stored, the time when the LiDAR sensor (500) starts a sensing operation, or the time when the point cloud data acquired and stored by the sensing operation of the LiDAR sensor (500) is updated.
[0053] Alternatively, according to an object tracking method (100B) of another embodiment, at each step, it may be checked whether the number of updates for each of the multiple LiDAR sensors (500) is different based on the recognition time (step 110B). Here, the number of updates may mean the number of update times.
[0054] As such, except for the difference in step 110B, the object tracking method (100B) illustrated in FIG. 2 is identical to the object tracking method (100A) illustrated in FIG. 1; therefore, the same reference numerals are used for parts identical to FIG. 1, and the description of the object tracking method (100B) illustrated in FIG. 2 is omitted.
[0055] Hereinafter, to aid in understanding the object tracking method (100A, 100B) illustrated in FIGS. 1 and 2, a plurality of lidar sensors (500) are described as comprising two first and second lidar sensors, each of which is a sensor positioned at the front, side, rear, or top of the vehicle (1000) (for example, the first lidar sensor is a sensor positioned on the right side of the vehicle (1000), and the second lidar sensor is a sensor positioned on the left side of the vehicle (1000)), but the embodiment is not limited to a specific number of lidar sensors (500) or a specific location of each of the first and second lidar sensors.
[0056] FIG. 4 shows a timing diagram to aid in understanding the object tracking method (100A, 100B) illustrated in FIG. 1 and FIG. 2.
[0057] In FIG. 4, the arrow (A1) indicates the update time (t11, t12, t13, t14, t15, t16, t17) of the first lidar sensor (hereinafter referred to as the 'first update time'), the arrow (A2) indicates the update time (t21, t22, t23, t24, t25, t26) of the second lidar sensor (hereinafter referred to as the 'second update time'), and the arrow (A3) indicates the recognition time (t31, t32).
[0058] If the period of the first update time (hereinafter referred to as the 'first period') is 'T1', the time difference between adjacent time points at the first update time (t11, t12, t13, t14, t15, t16, t17) is T1, if the period of the second update time (hereinafter referred to as the 'second period') is 'T2', the time difference between adjacent time points at the second update time (t21, t22, t23, t24, t25, t26) is T2, and if the period of the recognition time (hereinafter referred to as the 'third period') is 'T3', the time difference between adjacent time points at the recognition time (t31, t32) is T3.
[0059] For better understanding, only three steps (S(t-1), St, S(t+1)) are shown in Fig. 4. S(t-1) represents the step at the previous time point (hereinafter referred to as the 'previous step'), S(t) represents the step at the current time point (hereinafter referred to as the 'current step'), and S(t+1) represents the step at the next time point (hereinafter referred to as the 'next step').
[0060] Each of steps 110A and 110B can be performed at each step, for example, the previous step (S(t-1)), the current step (St), and the subsequent step (S(t+1)).
[0061] According to one embodiment, in the previous step (S(t-1)), it is checked whether the first update time (t11, t12, t13) and the second update time (t21, t22) are different from each other based on the recognition time (t31) (step 110A). In the current step (St), it is checked whether the first update time (t14, t15) and the second update time (t23, t24, t25) are different from each other based on the recognition time (t32) (step 110A). Step 110A may be performed in the same way in the subsequent step (S(t+1)).
[0062] Hereinafter, in each step, the number of updates of the first lidar sensor is referred to as the 'first number' and the number of updates of the second lidar sensor is referred to as the 'second number'.
[0063] According to another embodiment, in the previous step (S(t-1)), it is checked whether the first count and the second count are different (step 110B). In the current step (St), it is checked whether the first count and the second count are different (step 110B). Step 110B may also be performed in the subsequent step (S(t+1)).
[0064] FIG. 5 shows a flowchart according to an embodiment of step 110A shown in FIG. 1.
[0065] Update times are recorded for each of the multiple LiDAR sensors (Step 112). For example, a first update time and a second update time are recorded for each step (S(t-1), St, S(t+1)). As exemplified in FIG. 4, the first update time (t11, t12, t13) and the second update time (t21, t22) are recorded in the previous step (S(t-1)), the first update time (t14, t15) and the second update time (t23, t24, t25) are recorded in the current step (St), and the first update time (t16, t17, …) and the second update time (t26, …) are recorded in the subsequent step (S(t+1)).
[0066] After step 112, for each step (S(t-1), St, S(t+1)), the time difference between the recorded update time and the recognition time of each of the multiple lidar sensors (hereinafter referred to as 'asynchronous time') is calculated for each of the multiple lidar sensors (step 114).
[0067] For example, referring to FIG. 4, in the previous step (S(t-1), the asynchronous time between the first update time (t13) and the recognition time (t31) (hereinafter referred to as the 'first asynchronous time') and the asynchronous time between the second update time (t22) and the recognition time (t31) (hereinafter referred to as the 'second asynchronous time') are calculated. In the current step (St), the first asynchronous time between the first update time (t15) and the recognition time (t32) and the second asynchronous time between the second update time (t25) and the recognition time (t32) are calculated. In the subsequent step (S(t+1), the first and second asynchronous times are also calculated. Based on the recognition time (t31), the first asynchronous time is as shown in the following Equation 1, and the second asynchronous time can be expressed as shown in the following Equation 2.
[0068]
[0069] Here, US1 represents the first asynchronous time.
[0070]
[0071] Here, US2 represents the second asynchronous time.
[0072] After step 114, at every step (S(t-1), St, S(t+1)), it is checked whether the asynchronous times for each of the multiple LiDAR sensors are different (step 116). For example, at the previous step (S(t-1), it is checked whether the first asynchronous time and the second asynchronous time are different, at the current step (St), it is checked whether the first asynchronous time and the second asynchronous time are different, and at the subsequent step (S(t+1), it is checked whether the first asynchronous time and the second asynchronous time are different.
[0073] Again, referring to FIG. 1, when step 110A is performed and the first and second update times are determined to be different, or when step 110B shown in FIG. 2 is performed and the first and second times are determined to be different, a main sensor is selected among a plurality of lidar sensors at every step (or per track) (step 120). In this way, step 120 can be performed for all lidar tracks.
[0074] In the case of FIG. 4, since the first asynchronous time and the second asynchronous time are different in each step (S(t-1), St, S(t+1)), step 120 can be performed. Alternatively, in the case of FIG. 4, since the first number and the second number are different in each step (S(t-1), St, S(t+1)), step 120 can be performed.
[0075] Among multiple LiDAR sensors, the sensor that acquires many points (or dots) related to the contour formed by point cloud data can be determined as the main sensor.
[0076] FIG. 6 shows a flowchart according to an embodiment (120A) of the 120th step shown in FIG. 1 and FIG. 2.
[0077] For each of the multiple lidar sensors, the number of points distributed on the collision reference plane of the contour is counted (Step 122). Here, the collision reference plane refers to a plane located on the side of the contour facing the vehicle (1000). For example, among the points distributed on the collision reference plane of the contour, the number of points acquired by the first lidar sensor is counted, and the number of points acquired by the second lidar sensor is counted.
[0078] After Step 122, among the plurality of LiDAR sensors, the sensor with the highest count is selected as the primary sensor (Step 124). For example, among the points distributed on the collision reference plane of the contour, if the number of points acquired by the first LiDAR sensor is greater than the number of points acquired by the second LiDAR sensor, the first LiDAR sensor is selected as the primary sensor. Alternatively, among the points distributed on the collision reference plane of the contour, if the number of points acquired by the second LiDAR sensor is greater than the number of points acquired by the first LiDAR sensor, the second LiDAR sensor is selected as the primary sensor.
[0079] FIG. 7 is a diagram to aid in understanding Step 120 or Step 120A, and is a diagram that exemplarily shows three tracking targets (B1, B2, B3) based on the vehicle (BO).
[0080] In FIG. 7, the first to third tracking targets (B1, B2, B3) are depicted in the form of rectangular boxes, and the first, second, and third tracking targets (B1, B2, B3) are shown proceeding in the first, second, and third directions (D1, D2, D3), respectively.
[0081] As described above, when a plurality of lidar sensors include two first and second lidar sensors, a point (PNT1) of point cloud data acquired by the first lidar sensor and a point (PNT2) of point cloud data acquired by the second lidar sensor at each of the tracking targets (B1, B2, B3) are shown.
[0082] Here, for the first tracking target (B1), the number of points (PNT1) acquired by the first lidar sensor among the points distributed on the collision reference plane within the circular dotted line is greater than the number of points (PNT2) acquired by the second lidar sensor. Therefore, among the first and second lidar sensors, the first lidar sensor that acquired the points (PNT1) distributed more on the collision reference plane is determined as the main sensor.
[0083] In the case of the second tracking target (B2), among the points distributed on the collision reference plane within the circular dotted line, the number of points (PNT2) acquired by the second lidar sensor is greater than the number of points (PNT1) acquired by the first lidar sensor. Therefore, among the first and second lidar sensors, the second lidar sensor that acquired the points (PNT2) distributed more on the collision reference plane is determined as the main sensor.
[0084] In the case of the third tracking target (B3), among the points distributed on the collision reference plane within the circular dotted line, the number of points (PNT1) acquired by the first lidar sensor is greater than the number of points (PNT2) acquired by the second lidar sensor. Therefore, among the first and second lidar sensors, the first lidar sensor that acquired the points (PNT1) distributed more on the collision reference plane is determined as the main sensor.
[0085] Figures 8 (a) to (c) are diagrams illustrating the process of selecting a main sensor for each step.
[0086] In the case of the previous step (S(t-1)) shown in Fig. 8 (a), the first lidar sensor is selected as the main sensor because the number of points (PNT1) acquired by the first lidar sensor among the points distributed on the collision reference plane is greater than the number of points (PNT2) acquired by the second lidar sensor.
[0087] In the case of the current step (St) shown in Fig. 8 (b), the second lidar sensor is selected as the main sensor because the number of points (PNT2) acquired by the second lidar sensor among the points distributed on the collision reference plane is greater than the number of points (PNT1) acquired by the first lidar sensor.
[0088] In the case of the subsequent step (S(t+1)) shown in Fig. 8 (c), the first lidar sensor is selected as the main sensor because the number of points (PNT1) acquired by the first lidar sensor among the points distributed on the collision reference plane is greater than the number of points (PNT2) acquired by the second lidar sensor.
[0089] Again, referring to FIGS. 1 and FIGS. 2, after step 120, the time difference between the update time of the main sensor selected in the previous (t-1) step and the update time of the main sensor selected in the current (t) step is calculated as the 'step time difference' in the current step (step 130).
[0090] FIG. 9 is a timing diagram showing an example of the timing diagram illustrated in FIG. 4, wherein the first period (T1) of the first update time is 40 mm sec, the second period (T2) of the second update time is 40 mm sec, and the third period (T3) of the recognition time is 100 mm sec. Accordingly, 10 mm sec, 50 mm sec, 90 mm sec, 130 mm sec, 170 mm sec, 210 mm sec, and 250 mm sec shown in FIG. 9 correspond to the first update time points (t11, t12, t13, t14, t15, t16, t17) shown in FIG. 4, respectively; 30 mm sec, 70 mm sec, 110 mm sec, 150 mm sec, 190 mm sec, and 230 mm sec shown in FIG. 9 correspond to the second update time points (t21, t22, t23, t24, t25, t26) shown in FIG. 4, respectively; and 100 mm sec and 200 mm sec shown in FIG. 9 correspond to the first and second recognition time points (t31, t32) shown in FIG. 4, respectively.
[0091] If multiple lidar sensors include two first and second lidar sensors, there may be four different step time differences (STD1, STD2, STD3, STD4) in the current step (St) as exemplified in FIG. 9, depending on the type of main sensor selected in the previous step (S(t-1)) and the main sensor selected in the current step (St).
[0092] In the first case (①) where the first lidar sensor is selected as the main sensor in the previous step (S(t-1)) and the first lidar sensor is selected as the main sensor in the current step (St), the time difference between the first update time of the first lidar sensor in the previous step (S(t-1)) and the first update time of the first lidar sensor in the current step (St) can be determined as the first step time difference (STD1).
[0093] In the second case (②) where the first lidar sensor is selected as the main sensor in the previous step (S(t-1)) and the second lidar sensor is selected as the main sensor in the current step (St), the time difference between the first update time of the first lidar sensor in the previous step (S(t-1)) and the second update time of the second lidar sensor in the current step (St) can be determined as the second step time difference (STD2).
[0094] In the third case (③) where the second lidar sensor is selected as the main sensor in the previous step (S(t-1)) and the first lidar sensor is selected as the main sensor in the current step (St), the time difference between the second update time of the second lidar sensor in the previous step (S(t-1)) and the first update time of the first lidar sensor in the current step (St) can be determined as the third step time difference (STD3).
[0095] In the fourth case (④) where the second lidar sensor is selected as the main sensor in the previous step (S(t-1)) and the second lidar sensor is selected as the main sensor in the current step (St), the time difference between the second update time of the second lidar sensor in the previous step (S(t-1)) and the second update time of the second lidar sensor in the current step (St) can be determined as the fourth step time difference (STD4).
[0096] In the current step (St), the step time difference (STD1, STD2, STD3, STD4) may correspond to the time difference between the latest update time of the main sensor selected in the previous step (S(t-1)) and the latest update time of the main sensor selected in the current step (St).
[0097] Referring to FIG. 4 and FIG. 9, in the first case (①), the step time difference (STD1) at the current step (St) may correspond to the time difference (t15-t13) between the latest time point (t13) among the first update times (t11, t12, t13) of the first lidar sensor, which is the main sensor selected in the previous step (S(t-1)), and the latest time point (t15) among the first update times (t14, t15) of the first lidar sensor, which is the main sensor selected in the current step (St). For example, referring to FIG. 9, in the first case (①), the step time difference (STD1) at the current step (St) may be 80 mm sec.
[0098] In the second case (②), the step time difference (STD2) at the current step (St) may correspond to the time difference (t25-t13) between the latest time point (t13) among the first update times (t11, t12, t13) of the first lidar sensor, which is the main sensor selected in the previous step (S(t-1)), and the latest time point (t25) among the second update times (t23, t24, t25) of the second lidar sensor, which is the main sensor selected in the current step (St). For example, in the second case (②), referring to FIG. 9, the step time difference (STD2) at the current step (St) may be 100 mm sec.
[0099] In the third case (③), the step time difference (STD3) at the current step (St) may correspond to the time difference (t15-t22) between the latest time point (t22) among the second update times (t21, t22) of the second lidar sensor, which is the main sensor selected in the previous step (S(t-1)), and the latest time point (t15) among the first update times (t14, t15) of the first lidar sensor, which is the main sensor selected in the current step (St). For example, in the third case (③), referring to FIG. 9, the step time difference (STD3) at the current step (St) may be 100 mm sec.
[0100] In the fourth case (④), the step time difference (STD4) at the current step (St) may correspond to the time difference (t25-t22) between the latest time point (t22) among the second update times (t21, t22) of the second LiDAR sensor, which is the main sensor selected in the previous step (S(t-1)), and the latest time point (t25) among the second update times (t23, t24, t25) of the second LiDAR sensor, which is the main sensor selected in the current step (St). For example, in the fourth case (④), referring to FIG. 9, the step time difference (STD4) at the current step (St) may be 120 mm sec.
[0101] Referring again to FIG. 1 and FIG. 2, after step 130, when signal processing of point cloud data acquired from the main sensor selected at the current step (St), attribute information of the track at the current step (St) is obtained using the step time difference obtained at the current step (St) (step 140). For example, at least one of the speed, heading information, or shape of the track can be obtained as attribute information using the step time difference. Speed is the result of dividing distance by time, and the time used to obtain speed corresponds to the step time difference obtained at step 130.
[0102] Hereinafter, an object tracking device using a LiDAR sensor according to an embodiment is described as follows with reference to the attached drawings.
[0103] FIG. 10 shows a block diagram of an object tracking device (800) using a lidar sensor according to an embodiment.
[0104] For example, the object tracking device (800) illustrated in FIG. 10 may be included in the object tracking unit (640) illustrated in FIG. 3, but the embodiment is not limited thereto.
[0105] For convenience of explanation, the object tracking method (100A) illustrated in FIG. 1 is described as being performed in the object tracking device (800) illustrated in FIG. 10, but the embodiments are not limited thereto. That is, according to other embodiments, the object tracking method (100A) illustrated in FIG. 1 may also be performed in an object tracking device having a different configuration from the object tracking device (800) illustrated in FIG. 10. Furthermore, the object tracking device (800) illustrated in FIG. 10 is described as performing the object tracking method (100A) illustrated in FIG. 1, but the embodiments are not limited thereto.
[0106] The object tracking device (800) illustrated in FIG. 10 may include a sensing time point inspection unit (72), a sensor selection unit (74), a step time difference calculation unit (76), and an attribute information generation unit (78).
[0107] The sensing time inspection unit (72) checks whether the update time for each of the multiple LiDAR sensors is different based on the recognition time when signal processing of the point cloud data begins at every step. That is, the sensing time inspection unit (72) performs the role of performing the 110A step illustrated in FIG. 1. For example, the sensing time inspection unit (72) receives the recognition time and the first and second update time through the input terminal IN1, checks whether the first update time and the second update time are different based on the recognition time, and can output the checked result to the sensor selection unit (74).
[0108] FIG. 11 is a block diagram according to one embodiment (72A) of the sensing time inspection unit (72) shown in FIG. 10.
[0109] For convenience of explanation, the step 110A illustrated in FIG. 5 is described as being performed in the sensing time inspection unit (72A) illustrated in FIG. 11, but the embodiment is not limited thereto. That is, according to other embodiments, the step 110A illustrated in FIG. 5 may also be performed in a sensing time inspection unit having a different configuration from the sensing time inspection unit (72A) illustrated in FIG. 11.
[0110] The sensing time point inspection unit (72A) illustrated in FIG. 11 may include a time point recording unit (82), an asynchronous time calculation unit (84), and an asynchronous time inspection unit (86).
[0111] The time recording unit (82) can record update times for each of the multiple LiDAR sensors and output the recorded time to the asynchronous time calculation unit (84). The time recording unit (82) can perform step 112 as illustrated in FIG. 5. To do this, the time recording unit (82) can receive and record at least one of the recognition time, the first update time, and the second update time through the input terminal IN2, or record it by storing it in advance. At this time, the time recording unit (82) can output the first and second update times to the step time difference calculation unit (76) through the output terminal OUT2.
[0112] The asynchronous time calculation unit (84) can calculate the asynchronous time, which is the time difference between the recorded first and second update times and the recognition time output from the time record unit (82), and output it to the asynchronous time inspection unit (86). The asynchronous time calculation unit (84) can perform step 114 as illustrated in FIG. 5. To this end, the asynchronous time calculation unit (84) can output the time difference between the recognition time and the first update time output from the time record unit (82) as the first asynchronous time, and output the time difference between the recognition time and the second update time output from the time record unit (82) as the second asynchronous time.
[0113] The asynchronous time inspection unit (86) can inspect whether the asynchronous time is different for each of the multiple LiDAR sensors and output the inspected result to the sensor selection unit (74) through the output terminal OUT3. The asynchronous time inspection unit (86) can perform the 116th step illustrated in FIG. 5. That is, the asynchronous time inspection unit (86) can compare and inspect whether the first asynchronous time and the second asynchronous time are different or identical to each other, and output the result to the sensor selection unit (74) through the output terminal OUT3.
[0114] Again, referring to FIG. 10, the sensor selection unit (74) can select a main sensor among a plurality of LiDAR sensors at each step in response to the result checked by the sensing time inspection unit (72), and output information about the selected main sensor to the step time difference calculation unit (76). The sensor selection unit (74) can perform the 120th step illustrated in FIG. 1.
[0115] FIG. 12 shows a block diagram according to an embodiment (74A) of the sensor selection unit (74) illustrated in FIG. 10.
[0116] For convenience of explanation, the 120A step illustrated in FIG. 6 is described as being performed in the sensor selection unit (74A) illustrated in FIG. 12, but the embodiment is not limited thereto. That is, according to other embodiments, the 120A step illustrated in FIG. 6 may also be performed in a sensor selection unit having a different configuration from the sensor selection unit (74A) illustrated in FIG. 12.
[0117] The sensor selection unit (74A) illustrated in FIG. 12 may include a counter (92) and a count comparison unit (94).
[0118] When the counter (92) recognizes that the first update time and the second update time are different based on the recognition time through the inspection result provided from the sensing time inspection unit (72) via the input terminal IN3, it counts the number of points for each of the multiple LiDAR sensors distributed on the collision reference plane of the contour formed by the point cloud data, and outputs the counted result to the count comparison unit (94). The counter (92) performs the role of performing the 122nd step illustrated in FIG. 6.
[0119] The count comparison unit (94) selects the sensor with the highest count among the multiple LiDAR sensors as the main sensor and outputs information about the selected main sensor to the step time difference calculation unit (76) through the output terminal OUT4. The count comparison unit (94) performs the role of performing the 124th step shown in FIG. 6.
[0120] Referring again to FIG. 10, the step time difference calculation unit (76) calculates the time difference between the update time of the main sensor selected in the previous step (S(t-1)) and the update time of the main sensor selected in the current step (St) as the step time difference in the current step, and outputs the calculated step time difference to the attribute information generation unit (78). The step time difference calculation unit (76) performs the role of performing the 130th step illustrated in FIG. 1. To this end, the step time difference calculation unit (76) may receive the update time of the main sensor selected in the previous step (S(t-1)) and the update time of the main sensor selected in the current step (St) from the sensing time point inspection unit (72).
[0121] When the attribute information generation unit (78) signals the point cloud data acquired from the main sensor selected at the current step (St), it generates attribute information of the track at the current step using the step time difference output from the step time difference calculation unit (76) and outputs the generated attribute information through the output terminal OUT1. The attribute information generation unit (78) performs the role of performing the 140th step illustrated in FIG. 1. Accordingly, the attribute information generation unit (78) can generate at least one of the speed, heading information, or shape of the track at the current step (St) as attribute information using the step time difference obtained at the current step (St) and output it through the output terminal OUT1.
[0122] Meanwhile, referring to FIG. 3, for example, when the aforementioned object tracking method (100A, 100B) and device (600) are used in a vehicle (1000), the vehicle (1000) may further include a vehicle device (700). The vehicle device (700) can control the vehicle (1000) according to information about the target object tracked by the object tracking unit (640) and the type of the target object.
[0123] The vehicle device (700) can control the vehicle (1000) by receiving information identified about an object from the object tracking device (600). For example, the vehicle device (700) may include a lane keeping system that maintains a lane while maintaining a distance from a preceding vehicle, an obstacle detection system that detects obstacles around the vehicle, a collision prevention system that detects a collision risk, an autonomous driving system that performs autonomous driving by detecting obstacles in all directions, and a safety driving system that performs warnings and vehicle control when an adjacent vehicle approaches.
[0124] Meanwhile, a recording medium that records a program for executing an object tracking method (100A, 100B) for tracking an object using a LiDAR sensor (500) records a program that implements a function to check whether the update time of a plurality of LiDAR sensors is different based on the recognition time of starting signal processing of point cloud data at each step, a function to select a main sensor among the plurality of LiDAR sensors at each step if the update time is different, a function to obtain the time difference between the update time of the main sensor selected at the previous step (S(t-1)) and the update time of the main sensor selected at the current step (St) as the step time difference at the current step, and a function to obtain attribute information of a track at the current step using the step time difference when signal processing of point cloud data obtained from the main sensor selected at the current step, and a computer can read the recording medium.
[0125] In addition, a program is recorded that implements a function for checking whether update times are different, including a function for recording update times for each of multiple lidar sensors, a function for calculating the time difference between the recorded update time and the recognition time as an asynchronous time for each of multiple lidar sensors, and a function for checking whether the asynchronous time is different for each of multiple lidar sensors, and the computer can read the recording medium.
[0126] In addition, a program is recorded that implements a function for selecting a main sensor, which includes a function for counting the number of points distributed on a collision reference plane of a contour formed by point cloud data for each of the multiple lidar sensors, and a function for selecting the sensor with the highest number of counts among the multiple lidar sensors as the main sensor, and the computer can read the recording medium.
[0127] Computer-readable recording media include all types of storage devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage devices. Additionally, computer-readable recording media may be distributed across networked computer systems, allowing computer-readable code to be stored and executed in a distributed manner. Furthermore, functional programs, codes, and code segments for implementing the variable lens driving method described above can be easily inferred by programmers skilled in the art to which the present invention belongs.
[0128] Hereinafter, an object tracking method according to comparative examples and embodiments is described as follows with reference to the attached drawings.
[0129] According to the object tracking method based on the comparative example, the update times of multiple LiDAR sensors may differ and not coincide, or the period between the update time and the recognition time may not be in a multiple relationship. Consequently, when processing point cloud data signals, problems may arise in integrating and processing point cloud data acquired from each of the multiple LiDAR sensors.
[0130] Figures 13 (a) to (c) are drawings to explain the problems of the object tracking method according to the comparative example.
[0131] Point cloud data acquired by two LiDAR sensors of the object indicated by the circular dotted line in Fig. 13 (a) is shown in Fig. 13 (b) and Fig. 13 (c), respectively.
[0132] If the first and second update times of the first and second lidar sensors are the same, and the first and second cycles (T1, T2) and the third cycle (T3) are multiples of each other, then the location of the point cloud data acquired from the first and second lidar sensors for the same object is the same, as indicated by the dotted line in FIG. 13 (c).
[0133] However, when the first and second lidar sensors sense information about the same object located in front of the vehicle shown in FIG. 13 (a), if the first update time and the second update time are different, or if the first and second cycles (T1, T2) are not in a multiple relationship with the third cycle (T3), according to the comparative example, a layering phenomenon occurs in the point cloud data as indicated by the circular dotted line in FIG. 13 (b), and the positions of the point cloud data acquired by the first and second lidar sensors become different from each other.
[0134] FIGS. 14 (a) to (c) are drawings showing point cloud data acquired from the first and second lidar sensors at each step (S(t-1), St, S(t+1)), and showing the appearance of the track box (TB) at each step (S(t-1), St, S(t+1)).
[0135] When the first and second update times of the first and second LiDAR sensors are different from each other, the first and second periods (T1, T2) are each 40 mm sec, and the third period (T3) is 100 mm sec and is not a multiple of the first and second periods (T1, T2), as shown in FIGS. 4 and 9, in the previous step (S(t-1)), the first count is 3 and the second count is 2, which are different from each other, and in the current step (St), the first count is 2 and the second count is 3, which are different from each other. Therefore, as shown in FIG. 14 (b), the positions of the point cloud data acquired from the first LiDAR sensor and the point cloud data acquired from the second LiDAR sensor become different from each other. In this case, if the attribute information of the track is obtained using a constant third period (T3), the attribute information may be obtained inaccurately.
[0136] On the other hand, if the first update time and the second update time are different and the first and second cycles (T1, T2) are not multiples of the third cycle (T3), or if the first update time and the second update time are the same but the first and second cycles (T1, T2) are not multiples of the third cycle (T3), or if the first and second cycles (T1, T2) are multiples of the third cycle (T3) but the first update time and the second update time are different, according to the embodiment, the attribute information of the track at the current step is obtained using only the step time difference obtained at the current step and the point cloud data obtained from the main sensor, without using point cloud data obtained from the LiDAR sensor that is not selected as the main sensor. Therefore, according to the embodiment, attribute information, such as speed, heading, and shape, can be accurately obtained.
[0137] The aforementioned embodiments may be combined with one another unless specifically stated otherwise.
[0138] In addition, any parts omitted in the description of any one of the embodiments may be applied to the description of other embodiments unless specifically mentioned otherwise.
[0139] Although the above description has focused on exemplary embodiments, this is merely illustrative and does not limit the invention. Those skilled in the art will understand that various modifications and applications not exemplified above are possible within the scope of the essential characteristics of the embodiments. For example, each component specifically shown in the embodiments may be modified. Furthermore, differences related to such modifications and applications should be interpreted as being included within the scope of the invention as defined in the appended claims.
Claims
Claim 1 A method for tracking an object using a LiDAR sensor, comprising: a step of checking whether the update times of a plurality of LiDAR sensors are different based on a recognition time point at which signal processing of point cloud data begins at each step; a step of selecting a main sensor among the plurality of LiDAR sensors at each step if the update times are different; a step of determining the time difference between the update time of the main sensor selected at the previous (t-1) step and the update time of the main sensor selected at the current (t) step as a step time difference at the current step, which is determined differently at each step; and a step of obtaining attribute information of a track at the current step using the step time difference, such that point cloud data obtained from a LiDAR sensor not selected as a main sensor at the current step is not used, and point cloud data obtained from the main sensor selected at the current step is signal processed, and attribute information of the track at the current step for an object is obtained through the main sensor corresponding to the result of the step of selecting the main sensor at each step. Claim 2 A method for tracking an object using a lidar sensor according to claim 1, wherein the step of checking whether the update times are different comprises: a step of recording the update times for each of the plurality of lidar sensors; a step of obtaining the time difference between the recorded update times and the recognition times as an asynchronous time for each of the plurality of lidar sensors; and a step of checking whether the asynchronous time is different for each of the plurality of lidar sensors. Claim 3 In claim 2, the step of selecting the main sensor comprises: counting the number of points distributed on the collision reference plane of the contour formed by the point cloud data for each of the plurality of lidar sensors; and selecting the sensor with the largest number of counted points among the plurality of lidar sensors as the main sensor. Claim 4 A method for tracking an object using a LiDAR sensor according to claim 1, wherein the step time difference in the current step corresponds to the time difference between the latest update time of the main sensor selected in the previous (t-1) step and the latest update time of the main sensor selected in the current (t) step. Claim 5 A method for tracking objects using a LiDAR sensor according to claim 1, wherein the update period for sensing the point cloud data and the period for signal processing the point cloud data for each of the plurality of LiDAR sensors are not multiples of each other. Claim 6 In claim 1, the step of obtaining the attribute information is an object tracking method using a LiDAR sensor that obtains at least one of the speed, heading information, or shape of the track as the attribute information using the step time difference. Claim 7 A method for tracking an object using a LiDAR sensor, comprising: a step of checking whether the number of updates for a plurality of LiDAR sensors is different at each step; a step of selecting a primary sensor among the plurality of LiDAR sensors at each step if the number of updates is different; a step of determining the time difference between the update time of the primary sensor selected at the previous (t-1) step and the update time of the primary sensor selected at the current (t) step as a step time difference at the current step, which is determined differently at each step; and a step of obtaining attribute information of a track at the current step using the step time difference, such that point cloud data obtained from a LiDAR sensor not selected as a primary sensor at the current step is not used, and point cloud data obtained from the primary sensor selected at the current step is signal processed. Claim 8 An object tracking device using a LiDAR sensor, comprising: a sensing time inspection unit that checks whether the update times of a plurality of LiDAR sensors are different based on a recognition time point at which signal processing of point cloud data begins at each step; a sensor selection unit that selects a main sensor among the plurality of LiDAR sensors at each step in response to the result checked by the sensing time inspection unit; a step time difference calculation unit that calculates the time difference between the update time of the main sensor selected at the previous (t-1) step and the update time of the main sensor selected at the current (t) step as a step time difference at the current step that is determined differently at each step; and an attribute information generation unit that generates attribute information of a track at the current step using the step time difference so as not to use point cloud data acquired from a LiDAR sensor not selected as a main sensor at the current step, and to signal process point cloud data acquired from the main sensor selected at the current step, and wherein the attribute information of the track at the current step for an object is obtained through the main sensor corresponding to the result selected by the sensor selection unit at each step. Claim 9 In claim 8, the object tracking device using a lidar sensor comprises: a time recording unit that records the update time for each of the plurality of lidar sensors; an asynchronous time calculation unit that calculates the time difference between the recorded update time and the recognition time as an asynchronous time for each of the plurality of lidar sensors; and an asynchronous time inspection unit that checks whether the asynchronous time is different for each of the plurality of lidar sensors. Claim 10 In claim 9, the sensor selection unit comprises: a counter for counting the number of points distributed on the collision reference plane of a contour formed by the point cloud data for each of the plurality of lidar sensors; and a count comparison unit for selecting the sensor with the largest number of counted points among the plurality of lidar sensors as the main sensor. Claim 11 In claim 8, the step time difference calculation unit calculates the time difference between the latest update time of the main sensor selected in the previous (t-1) step and the latest update time of the main sensor selected in the current (t) step as the step time difference, an object tracking device using a LiDAR sensor. Claim 12 In claim 8, an object tracking device using a LiDAR sensor in which the update cycle for sensing the point cloud data and the signal processing cycle for the point cloud data are not multiples of each other. Claim 13 In claim 8, the attribute information generating unit generates at least one of the speed, heading information, or shape of the track as attribute information using the step time difference, an object tracking device using a LiDAR sensor. Claim 14 A vehicle comprising a lidar sensor; and an object tracking device using the lidar sensor and described in any one of claims 8 to 13. Claim 15 A computer-readable recording medium having a program for executing an object tracking method using a LiDAR sensor, wherein, at each step, a function for checking whether the update times of a plurality of LiDAR sensors are different based on a recognition time point at which signal processing of point cloud data begins; if the update times are different, a function for selecting a primary sensor among the plurality of LiDAR sensors at each step; a function for determining the time difference between the update time of the primary sensor selected at the previous (t-1) step and the update time of the primary sensor selected at the current (t) step as a step time difference at the current step, which is determined differently at each step; and a function for obtaining attribute information of a track at the current step using the step time difference, such that point cloud data obtained from a LiDAR sensor not selected as a primary sensor at the current step is not used, and point cloud data obtained from the primary sensor selected at the current step is signal processed, and the program implements the function of obtaining attribute information of the track at the current step for an object through the primary sensor corresponding to the result selected in the function of selecting the primary sensor at each step. Claim 16 A computer-readable recording medium having a program that implements, wherein the function of checking whether the update times are different includes: a function of recording the update times for each of the plurality of lidar sensors; a function of obtaining the time difference between the recorded update times and the recognition times as an asynchronous time for each of the plurality of lidar sensors; and a function of checking whether the asynchronous times are different for each of the plurality of lidar sensors. Claim 17 A computer-readable recording medium having a program that implements, in claim 16, a function for selecting the main sensor, a function for counting the number of points distributed on the collision reference plane of a contour formed by the point cloud data for each of the plurality of lidar sensors; and a function for selecting the sensor with the largest number of counted points among the plurality of lidar sensors as the main sensor.
