阅读说明：本技术 一种目标跟踪数据滤波处理方法及装置 (Target tracking data filtering processing method and device ) 是由 李坤乾 朱飞亚 张志豪 顾翔 于 2020-12-29 设计创作，主要内容包括：本发明提供了一种目标跟踪数据滤波处理方法及装置,该方案根据自车与当前追踪目标的不同运动状态设置不同的初始滤波比。而且,基于当前追踪目标与自车之间的相对空间位置信息确定滤波比修正因子,并利用该滤波比修正因子对初始滤波比进行修正,使得最终得到的动态滤波比与滤波比修正因子正相关,而且,随着空间距离增加,滤波比修正因子逐渐增大并趋近于某个收敛值,即,动态滤波比随空间距离增加而增大,并趋于一个收敛值。随着当前追踪目标的距离增大,测量噪声也会增加,设置较大的滤波比以抑制测量噪声的影响。利用该方案最终实现目标跟踪的实时性和连续性,并且跟踪处理得到的追踪目标的航迹平滑、稳定。(The invention provides a target tracking data filtering processing method and device, and different initial filtering ratios are set according to different motion states of a self vehicle and a current tracking target. And, the filter ratio correction factor is determined based on the relative spatial position information between the current tracked target and the own vehicle, and the initial filter ratio is corrected by using the filter ratio correction factor, so that the finally obtained dynamic filter ratio is positively correlated with the filter ratio correction factor, and the filter ratio correction factor gradually increases and approaches a certain convergence value as the spatial distance increases, that is, the dynamic filter ratio increases with the increase of the spatial distance and approaches a convergence value. As the distance of the currently tracked target increases, the measurement noise also increases, and a larger filter ratio is set to suppress the influence of the measurement noise. By utilizing the scheme, the real-time performance and continuity of target tracking are finally realized, and the track of the tracked target obtained by tracking processing is smooth and stable.)

1. A target tracking data filtering processing method is characterized by comprising the following steps:

determining an initial filtering ratio of an alpha-beta filtering algorithm according to the motion state of the self vehicle and the track tracking stable stage information of the current tracking target;

determining a filter ratio correction factor according to the relative spatial position information of the self-vehicle and the current tracking target; the relative spatial position information comprises an spatial distance and an azimuth angle, the filter ratio correction factor is gradually increased and converged along with the increase of the spatial distance under the same azimuth angle, the larger the azimuth angle is, the smaller the filter ratio correction factor is, the azimuth angle is the angle of the current tracking target detected by the vehicle-mounted radar, and the spatial distance is the radial distance between the current tracking target and the vehicle-mounted radar;

generating a dynamic filtering ratio of the current tracking target according to the initial filtering ratio and the filtering ratio correction factor, wherein the dynamic filtering ratio is positively correlated with the filtering ratio correction factor;

and comparing the current motion state predicted value and the current motion state measured value of the current tracking target by using the dynamic filtering to carry out filtering processing, so as to obtain target state data of the current tracking target.

2. The method according to claim 1, wherein the determining an initial filter ratio of an alpha-beta filter algorithm according to the motion state of the self vehicle and the track tracking stabilization phase information of the current tracking target comprises:

when the motion state of the self-vehicle is a turning state, determining an initial filter ratio of an alpha-beta filter algorithm as a first filter ratio, wherein the first filter ratio enables the target state data to be heavier than the current motion state measured value;

when the motion state of the self-vehicle is a turning transition state or a lane change state, determining the motion stage of the current tracking target according to the track tracking stable stage information, and determining the initial filter ratio of an alpha-beta filter algorithm according to the motion stage of the current tracking target; when the motion state of the self-vehicle is a turning transition state or a lane changing state, dividing the motion phase of the current tracking target into a first initial phase and a first stable phase;

when the motion state of the self-vehicle is a straight-going state, determining the motion stage of the current tracking target according to the track tracking stable stage information, and determining the initial filtering ratio of an alpha-beta filtering algorithm according to the motion stage of the current tracking target; when the motion state of the self-vehicle is a straight-going state, the motion phase of the current tracking target is divided into a second starting phase, a relative stable phase and a second stable phase.

3. The method according to claim 2, wherein when the motion state of the host vehicle is an overturning state or a lane changing state, determining an initial filter ratio of an alpha-beta filter algorithm according to the motion phase of the current tracking target comprises:

determining the initial filter ratio to be a second filter ratio that emphasizes the target state data over the current motion state measurement when the current tracked target is at the first initial stage;

when the current tracking target is in the first stable stage, determining the initial filtering ratio as a third filtering ratio, wherein the third filtering ratio is larger than the second filtering ratio.

4. The method according to claim 2 or 3, wherein the track following stable stage information comprises a target continuous tracking frame number of the current tracking target;

the determining the motion stage of the current tracking target according to the track tracking stable stage information of the current tracking target includes:

when the target continuous tracking frame number of the current tracking target is smaller than a first frame number threshold, determining that the current tracking target is currently in the first initial stage;

and when the target continuous tracking frame number of the current tracking target is greater than or equal to a first frame number threshold, determining that the current tracking target is in the first stable stage.

5. The method according to claim 2, wherein when the motion state of the self-vehicle is a straight-going state, determining an initial filter ratio of an alpha-beta filter algorithm according to the motion stage of the current tracking target comprises:

when the motion phase of the current tracked target is in a second initial phase, determining the initial filter ratio as a fourth filter ratio, wherein the fourth filter ratio enables the target state data to focus on the current motion state measured value;

when the motion phase of the current tracking target is in the relatively stable phase, determining that the initial filtering ratio is a fifth filtering ratio, wherein the fifth filtering ratio is greater than the fourth filtering ratio;

when the motion phase of the current tracking target is in the second stable phase, determining that the initial filtering ratio is a sixth filtering ratio, and the sixth filtering ratio is greater than the fifth filtering ratio.

6. The method according to claim 2 or 5, wherein the track following stable stage information comprises a target continuous tracking frame number of the current tracking target;

the determining the motion stage of the current tracking target according to the track tracking stable stage information comprises:

when the target continuous tracking frame number of the current tracking target is smaller than a second frame number threshold value, determining that the current tracking target is currently in the second initial stage;

when the target continuous tracking frame number of the current tracking target is greater than or equal to a second frame number threshold and smaller than a third frame number threshold, determining that the current tracking target is currently in the relatively stable stage;

and when the target continuous tracking frame number of the current tracking target is greater than or equal to a third frame number threshold value, determining that the current tracking target is in the second stable stage.

7. The method of claim 1, wherein determining a filter ratio correction factor based on the relative spatial location information of the host vehicle and the current tracking target comprises:

calculating the filter ratio correction factor according to the following formula:

wherein range is the spatial distance, a and b are adjustment coefficients respectively, and θ is the azimuth angle.

8. The method of claim 1, wherein generating the dynamic filter ratio for the current tracked target based on the initial filter ratio and the filter ratio correction factor comprises:

and calculating the dynamic filtering ratio according to the following formula:

wherein k represents the dynamic filter ratio, factor represents the filter ratio correction factor,is the initial filter ratio of the current filtering process.

9. The method according to claim 1, wherein the obtaining target state data of the current tracking target by filtering the current motion state predicted value and the current motion state measured value of the current tracking target by using the dynamic filtering ratio comprises:

and calculating the product of the dynamic filtering ratio and the current motion state predicted value, and calculating the sum of the product and the current motion state measured value to obtain the target state data.

10. An object tracking data filter processing apparatus, comprising: a memory and a processor;

the memory has stored therein program instructions;

the processor is configured to call program instructions in the memory to perform the target tracking data filtering processing method of any one of claims 1 to 9.

Technical Field

The invention belongs to the technical field of computers, and particularly relates to a target tracking data filtering processing method and device.

Background

With the rapid development of artificial intelligence technology, many fields need to detect and track targets, for example, in an intelligent driving scene, target detection and tracking are indispensable components of an intelligent driving assistance system. In an intelligent driving scene, the radar has good speed measurement capability on a target and good penetration capability on environmental interference such as rain, fog and the like, and becomes an irreplaceable sensor in an intelligent driving system.

The basic principle of radar-based target detection is that point cloud data detected by a radar is further divided into moving target point cloud data and static target point cloud data according to discrimination conditions, wherein static targets comprise absolute static targets (such as guardrails, cement piers and the like) and low-speed moving targets (such as pedestrians, transversely moving bicycles and the like). In consideration of the characteristics of the static target, a simple tracking algorithm can be selected for the point cloud data of the static target in the target tracking process, and a good tracking effect can be obtained. For example, in alpha-beta filtering, in conventional alpha-beta filtering algorithm processing, a fixed filtering ratio is usually set in the whole tracking process of a target, and a phenomenon that the tracking effect is different due to different values of the fixed filtering ratio is easy to occur.

Therefore, a method capable of adaptively setting a filter ratio according to a motion state of a target is needed to improve a target tracking effect.

Disclosure of Invention

In view of the above, the present invention aims to provide a target tracking data filtering processing method and device, so as to solve the technical problem of poor tracking effect caused by adopting a fixed filtering ratio for a static target in a target tracking process, and the disclosed technical solution is as follows:

in one aspect, the present invention provides a target tracking data filtering processing method, including:

determining an initial filtering ratio of an alpha-beta filtering algorithm according to the motion state of the self vehicle and the track tracking stable stage information of the current tracking target;

determining a filter ratio correction factor according to the relative spatial position information of the self-vehicle and the current tracking target; the relative spatial position information comprises an spatial distance and an azimuth angle, the filter ratio correction factor is gradually increased and converged along with the increase of the spatial distance under the same azimuth angle, the larger the azimuth angle is, the smaller the filter ratio correction factor is, the azimuth angle is the angle of the current tracking target detected by the vehicle-mounted radar, and the spatial distance is the radial distance between the current tracking target and the vehicle-mounted radar;

generating a dynamic filtering ratio of the current tracking target according to the initial filtering ratio and the filtering ratio correction factor, wherein the dynamic filtering ratio is positively correlated with the filtering ratio correction factor;

and comparing the current motion state predicted value and the current motion state measured value of the current tracking target by using the dynamic filtering to carry out filtering processing, so as to obtain target state data of the current tracking target.

In a possible implementation manner, the determining an initial filter ratio of an α - β filter algorithm according to a motion state of the vehicle and track tracking stabilization phase information of a current tracking target includes:

when the motion state of the self-vehicle is a turning state, determining an initial filter ratio of an alpha-beta filter algorithm as a first filter ratio, wherein the first filter ratio enables the target state data to be heavier than the current motion state measured value;

when the motion state of the self-vehicle is a turning transition state or a lane change state, determining the motion stage of the current tracking target according to the track tracking stable stage information, and determining the initial filter ratio of an alpha-beta filter algorithm according to the motion stage of the current tracking target; when the motion state of the self-vehicle is a turning transition state or a lane changing state, dividing the motion phase of the current tracking target into a first initial phase and a first stable phase;

when the motion state of the self-vehicle is a straight-going state, determining the motion stage of the current tracking target according to the track tracking stable stage information, and determining the initial filtering ratio of an alpha-beta filtering algorithm according to the motion stage of the current tracking target; when the motion state of the self-vehicle is a straight-going state, the motion phase of the current tracking target is divided into a second starting phase, a relative stable phase and a second stable phase.

In another possible implementation manner, when the motion state of the host vehicle is a turning transition state or a lane change state, determining an initial filter ratio of an α - β filter algorithm according to a motion phase in which the current tracking target is located includes:

determining the initial filter ratio to be a second filter ratio that emphasizes the target state data over the current motion state measurement when the current tracked target is at the first initial stage;

when the current tracking target is in the first stable stage, determining the initial filtering ratio as a third filtering ratio, wherein the third filtering ratio is larger than the second filtering ratio.

In yet another possible implementation manner, the track following stabilization phase information includes a target continuous tracking frame number of the current tracking target;

the determining the motion stage of the current tracking target according to the track tracking stable stage information of the current tracking target includes:

when the target continuous tracking frame number of the current tracking target is smaller than a first frame number threshold, determining that the current tracking target is currently in the first initial stage;

and when the target continuous tracking frame number of the current tracking target is greater than or equal to a first frame number threshold, determining that the current tracking target is in the first stable stage.

In another possible implementation manner, when the motion state of the self-vehicle is a straight-ahead state, determining an initial filter ratio of an α - β filter algorithm according to a motion stage where the current tracking target is located includes:

when the motion phase of the current tracked target is in a second initial phase, determining the initial filter ratio as a fourth filter ratio, wherein the fourth filter ratio enables the target state data to focus on the current motion state measured value;

when the motion phase of the current tracking target is in the relatively stable phase, determining that the initial filtering ratio is a fifth filtering ratio, wherein the fifth filtering ratio is greater than the fourth filtering ratio;

when the motion phase of the current tracking target is in the second stable phase, determining that the initial filtering ratio is a sixth filtering ratio, and the sixth filtering ratio is greater than the fifth filtering ratio.

In yet another possible implementation manner, the track following stabilization phase information includes a target continuous tracking frame number of the current tracking target;

the determining the motion stage of the current tracking target according to the track tracking stable stage information comprises:

when the target continuous tracking frame number of the current tracking target is smaller than a second frame number threshold value, determining that the current tracking target is currently in the second initial stage;

when the target continuous tracking frame number of the current tracking target is greater than or equal to a second frame number threshold and smaller than a third frame number threshold, determining that the current tracking target is currently in the relatively stable stage;

and when the target continuous tracking frame number of the current tracking target is greater than or equal to a third frame number threshold value, determining that the current tracking target is in the second stable stage.

In another possible implementation manner, the determining a filter ratio correction factor according to the relative spatial position information of the own vehicle and the current tracking target includes:

calculating the filter ratio correction factor according to the following formula:

wherein range is the spatial distance, a and b are adjustment coefficients respectively, and θ is the azimuth angle.

In another possible implementation manner, the generating a dynamic filter ratio of the current tracking target according to the initial filter ratio and the filter ratio modification factor includes:

and calculating the dynamic filtering ratio according to the following formula:

wherein k represents the dynamic filter ratio, factor represents the filter ratio correction factor,is the initial filter ratio of the current filtering process.

In another possible implementation manner, the performing filtering processing on the current motion state predicted value and the current motion state measured value of the current tracking target by using the dynamic filtering ratio to obtain target state data of the current tracking target includes:

and calculating the product of the dynamic filtering ratio and the current motion state predicted value, and calculating the sum of the product and the current motion state measured value to obtain the target state data.

On the other hand, the invention also provides a target tracking data filtering processing device, which comprises: a memory and a processor;

the memory has stored therein program instructions;

the processor is configured to call the program instructions in the memory to execute the target tracking data filtering processing method according to any one of the possible implementation manners of the first aspect.

According to the target tracking data filtering processing method provided by the invention, different initial filtering ratios are set according to different motion states of the self-vehicle and the current tracking target. And, the filter ratio correction factor is determined based on the relative spatial position information between the current tracked target and the own vehicle, and the initial filter ratio is corrected by using the filter ratio correction factor, so that the finally obtained dynamic filter ratio is positively correlated with the filter ratio correction factor, and the filter ratio correction factor gradually increases and approaches a certain convergence value as the spatial distance increases, that is, the dynamic filter ratio increases with the increase of the spatial distance and approaches a convergence value. As the distance of the currently tracked target increases, the measurement noise also increases, and a larger filter ratio is set to suppress the influence of the measurement noise. Moreover, the filter ratio correction factor with the convergence characteristic can avoid the phenomenon that the actual position of the tracking target is really changed, and the tracking real-time performance is deteriorated due to a larger filter ratio, so that the real-time performance and continuity of target tracking are finally realized, and the track of the tracking target obtained by tracking processing is smooth and stable.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

Fig. 1 is a flowchart of a target tracking data filtering processing method according to an embodiment of the present invention;

fig. 2 is a schematic diagram of a relative position relationship between a radar and a tracking target according to an embodiment of the present invention;

FIG. 3 is a schematic diagram illustrating a variation curve of a filter ratio correction factor according to a radial distance of a tracked target according to an embodiment of the present invention;

FIG. 4 is a graph illustrating a dynamic filter ratio versus a target spatial distance according to an embodiment of the present invention;

fig. 5 is a schematic structural diagram of a target tracking data filtering processing apparatus according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of an initial filter ratio determining module according to an embodiment of the present invention.

Detailed Description

For the defects of the fixed filter ratio algorithm, an improved alpha-beta filter algorithm also exists in the related art, however, the improvement direction mainly focuses on the adaptive setting of the filter ratio parameters based on the covariance of the motion state measured value and the predicted value of the tracked target in the statistical sense, the alpha-beta filter algorithm has large calculation amount, and the finally obtained adaptive filter ratio parameters are only optimal in the statistical sense and are not the filter ratio parameters suitable for the actual application scene. Therefore, the invention provides a new target tracking data filtering processing method, the scheme sets different initial filtering ratios according to the motion state of the self-vehicle and the track tracking stabilization stage information of the current tracking target, so the initial filtering ratios are more suitable for the actual scene, a filtering ratio correction factor is determined based on the relative spatial position information between the current tracking target and the self-vehicle, the initial filtering ratios are corrected by using the filtering ratio correction factor, the filtering ratios in the target tracking process can be adjusted adaptively along with the track tracking stabilization stage information of the current tracking target and the motion state and stability of the self-vehicle, and the target tracking effect is improved.

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Referring to fig. 1, a flowchart of a target tracking data filtering processing method provided by an embodiment of the present invention is shown, where the method is applied to a vehicle controller, and achieves stable tracking of a static target (a completely static target or a low-speed moving target) in a vehicle-mounted radar target tracking data processing process, as shown in fig. 1, the method includes the following steps:

and S110, determining an initial filtering ratio of an alpha-beta filtering algorithm according to the motion state of the self vehicle and the track tracking stable stage information of the current tracking target.

The filtering ratio in the alpha-beta filtering algorithm refers to alpha/beta, and in the application scenario of the present invention, the filtering ratio represents the proportion of the current motion state prediction value and the current motion state measurement value contained in the final motion state of the current tracking target.

In this embodiment, the current motion state prediction value is prediction data of the current motion state of the current tracking target, which is obtained by prediction according to the motion state transition model of the current tracking target. The current motion state measured value refers to measured data of the current motion state of the current tracking target obtained by using vehicle-mounted radar measurement.

The smaller the filtering ratio is, the smaller the proportion of the predicted value of the current motion state is, and the filtering result has faster dynamic property and real-time property, but the noise influence is also increased; the larger the filtering ratio is, the larger the proportion of the current motion state prediction value is, and the smoother the filtered value is, but the dynamic property and the real-time property are correspondingly deteriorated. Therefore, a reasonable filtering ratio needs to be set, and balance is made between dynamic performance and filtering noise.

According to the embodiment of the invention, the initial filtering ratio can be dynamically adjusted according to the motion state of the self-vehicle and the track tracking stable stage information of the current tracking target. Namely, a certain preset corresponding relation exists between the motion state of the self vehicle and the track tracking stable stage information of the current tracking target and the initial filtering ratio; and determining an initial filtering ratio matched with the preset corresponding relation, the self-vehicle motion state and the track tracking stable stage information of the current tracking target.

In addition, the preset corresponding relationship may be obtained by fitting according to experimental data, and the preset corresponding relationship may be expressed in various ways such as a function and a table.

The track tracking stabilization stage information represents a track tracking stage/motion stage of the current tracking target, and is a stability index of the current tracking target, generally, the number of frames for tracking the current tracking target is small, namely, the current tracking target is considered to be in an initial stage when the current tracking target starts to be tracked, the stability is poor, and when the number of frames for tracking the current tracking target reaches a certain number, the track tracking is in a stabilization stage, and the stability is good. Of course, in practical applications, the initial stage and the stable stage may be further divided.

In one embodiment of the present invention, the motion state of the host vehicle is mainly used for determining the motion relationship between the host vehicle and the current tracking target. The different complexity of the motion relationship will result in the accuracy of the predicted value of the current motion state of the tracked target. The more complex the motion relationship, the more difficult it is for the tracked target to satisfy a single motion state transition model, and the accuracy of the current motion state prediction value obtained by using the motion state transition model to predict is difficult to guarantee. Conversely, the simpler the motion relationship, the easier it is to find a motion state transition model that matches the motion state of the tracked target, in which case the more accurate the current motion state prediction value is.

The filtering process is related to attributes of different motion phases in which the tracking target is located, in addition to the motion state of the own vehicle, and for example, the motion phase in which the tracking target is located may be divided into an initial phase and a stable phase, or may be divided into an initial phase, a relatively stable phase, and a stable phase. The filtering ratio can be set appropriately in connection with tracking different motion phases of the target.

In one embodiment of the present invention, the motion state of the host vehicle mainly includes a turning state, an over-turning or lane-changing state, and a straight-going state (or a nearly straight-going state).

When the self-vehicle is in a turning state, the motion relation between the self-vehicle and the current tracking target is complex, and a single motion state transition model is difficult to meet.

When the self-vehicle is in a non-turning state, the motion stage of the current tracking target can be determined according to the track tracking stable stage information of the current tracking target, and then the initial filtering ratio is determined according to the motion stage of the current tracking target. The initial filtering ratio is positively correlated with the stability of the current tracking target in the motion phase, that is, the more stable the track of the tracking target is, the larger the initial filtering ratio is.

When the self-vehicle is in a turning transition state or a lane changing state, the motion relation between the self-vehicle and the current tracking target is not very complex, and when the initial filtering ratio is set, the matched initial filtering ratio can be determined by properly combining the track stability of the current tracking target.

In this case, the motion phase of the current tracking target may be divided into a first initial phase and a first stable phase. The initial filter ratio determined when the current tracking target is in the first stable stage is larger than the initial filter ratio determined when the current tracking target is in the first initial stage. Herein, the initial filter ratio determined when the current tracking target is in the first initial stage is referred to as a second filter ratio, and the initial filter ratio determined when the current tracking target is in the first stable stage is referred to as a third filter ratio. Typically, the second filter ratio places the target state data in focus on the current motion state measurement, but typically the second filter ratio is greater than the first filter ratio.

When the self-vehicle is in a straight running state or an approximately straight running state, the relative motion relation between the self-vehicle and the current tracking target is simple, and the motion state transfer model is single.

Under the motion condition, the motion phase of the current tracking target can be divided into a second initial phase, a relatively stable phase and a second stable phase, wherein the initial filtering ratio corresponding to the relatively stable phase is greater than the initial filtering ratio corresponding to the first initial phase, and meanwhile, the initial filtering ratio corresponding to the second stable phase is greater than the initial filtering ratio corresponding to the relatively stable phase. Herein, the initial filtering ratio determined when the current tracking target is in the second initial stage is referred to as a fourth filtering ratio, the initial filtering ratio determined when the current tracking target is in the relatively stable stage is referred to as a fifth filtering ratio, and the initial filtering ratio determined when the current tracking target is in the second stable stage is referred to as a sixth filtering ratio. Generally, the fourth filter ratio emphasizes the target state data over the current motion state measurements, but typically the fourth filter ratio is greater than the first filter ratio.

In practical application, the track tracking stabilization stage information may include a target continuous tracking frame number of the current tracking target, and the current motion stage of the current tracking target may be determined according to the target continuous tracking frame number of the current tracking target.

Optionally, in a possible implementation manner of the present invention, when the motion state of the host vehicle is an excessive turning state or a lane change state, determining the motion stage of the current tracking target according to the track tracking stable stage information of the current tracking target, includes:

when the target continuous tracking frame number of the current tracking target is smaller than a first frame number threshold, determining that the current tracking target is at a first initial stage;

when the target continuous tracking frame number of the current tracking target is greater than or equal to the first frame number threshold value, determining that the current tracking target is in a first stable stage.

Optionally, in a possible implementation manner of the present invention, when the motion state of the self-vehicle is a straight-ahead state, determining the motion stage where the current tracking target is located according to the track tracking stable stage information of the current tracking target, includes:

when the target continuous tracking frame number of the current tracking target is smaller than a second frame number threshold value, determining that the current tracking target is at a second initial stage;

when the target continuous tracking frame number of the current tracking target is greater than or equal to the second frame number threshold and smaller than the third frame number threshold, determining that the current tracking target is in a relatively stable stage;

and when the target continuous tracking frame number of the current tracking target is greater than or equal to the third frame number threshold, determining that the current tracking target is in a second stable stage.

The first filtering ratio, the second filtering ratio, the third filtering ratio, the fourth filtering ratio, the fifth filtering ratio and the sixth filtering ratio are empirical values or calibrated values, and the first frame number threshold, the second frame number threshold and the third frame number threshold are empirical values or calibrated values.

It is understood that, in other embodiments, the track following stable phase information may include tracking duration information of the current tracking target, and the motion phase of the current tracking target is determined according to the tracking duration information.

In practical application, when the initial filter ratio is adjusted, parameters α and β are adjusted, wherein specific values of α and β are not limited, but a proportional relationship between α and β satisfies the adjustment requirement of the initial filter ratio.

And S120, determining a filter ratio correction factor according to the relative spatial position information of the self-vehicle and the current tracking target.

The relative spatial position information of the self-vehicle and the current tracking target comprises a spatial distance and an azimuth angle, wherein the spatial distance is a radial distance between the current tracking target and the vehicle-mounted radar, and the azimuth angle is an angle of the current tracking target detected by the vehicle-mounted radar, and the angle is an angle of the current tracking target detected by the vehicle-mounted radar under a radar coordinate system.

Referring to fig. 2, a schematic diagram of a relative position relationship of a tracking target detected by a vehicle-mounted radar is shown, as shown in fig. 2, assuming that a motion state transition model of the tracking target is established in a rectangular coordinate system, and taking a vehicle-mounted radar as a central point, relative spatial position information between the tracking target and the vehicle-mounted radar may be represented by (X, Y, θ) or (r, θ), where X represents a distance component of a linear distance between the tracking target and the vehicle-mounted radar on an X axis, Y represents a distance component on a Y axis, θ represents an azimuth angle of the tracking target relative to the vehicle-mounted radar, and r represents a radial distance between the current tracking target and the vehicle-mounted radar, that is, a spatial distance between the current tracking target and the vehicle-. Obviously, the spatial distance can also be expressed in x and y.

The relative spatial position information between the vehicle-mounted radar and the tracking target and the filter ratio correction factor have a preset corresponding relationship, and the preset corresponding relationship has the following characteristics: under the condition that the azimuth angle is fixed, the filter ratio correction factor is gradually increased and converged along with the increase of the spatial distance; and, with fixed spatial distance, the filter ratio correction factor is inversely related to the azimuth.

Determining a filter ratio correction factor according to relative spatial position information of a self-vehicle and a current tracking target and a preset corresponding relation between the relative spatial position information and the filter ratio correction factor; the preset corresponding relation comprises positive correlation between the filter ratio correction factor and the spatial distance, and negative correlation between the filter ratio correction factor and the azimuth angle, and can be obtained according to experimental data fitting, in other words, the relation between the filter ratio correction factor and the relative spatial position information can be represented by a certain function.

And determining a filter ratio correction factor of the radar data of the current frame by combining the relative spatial position information (such as radial distance, azimuth angle and the like) of the self-vehicle and the tracking target, wherein the factor is f (x, y, theta) or f (r, theta).

The specific functional expression of the filter ratio correction factor may be in many ways, and the embodiment of the present invention provides an example shown in the following formula, but is not limited to this expression:

in formula 1, range is the radial distance between the current position of the current tracking target and the vehicle-mounted radar, that is to saya. b is respectively an adjusting coefficient, theta is an azimuth angle of the current tracking target relative to the vehicle-mounted radar, obviously, the functional relation among the radial distance, the azimuth angle and the filter ratio correction factor is not limited to formula 1, other functional relations capable of presenting the function characteristics of formula 1 can be applied to the embodiment of the invention, even formula 1 and the subsequent formula 2 can be changed in a linkage manner, as long as the finally obtained filter ratio can reach the filter ratio of the inventionThe technical effect of the embodiment is as follows.

Fig. 3 is a schematic diagram illustrating a variation curve of a filter ratio correction factor according to a radial distance of a tracked target according to an embodiment of the present invention.

As shown in fig. 3, the curve reflects that when the spatial distance of the tracking target (i.e., the distance between the tracking target and the vehicle-mounted radar) is relatively short, the filter ratio correction factor changes rapidly, and as the spatial distance of the tracking target increases, the change of the filter ratio correction factor gradually decreases until the tracking target converges to a stable value, and the filter ratio obtained based on the filter ratio correction factor changes in a certain positive correlation with the spatial distance of the tracking target, but does not increase in a positive correlation indefinitely, but as the distance increases, the corrected filter ratio also tends to a converging value, so that when the spatial distance of the tracking target is relatively long, the stability of target tracking can be ensured, and the real-time performance of target movement can be considered.

The curve shown in fig. 3 also reflects the change characteristics of the correction factors at different azimuth angles at a certain distance, the larger the azimuth angle is, the larger the expansibility of the tracked target and the uncertainty of the current motion state measurement value are, in this case, the filter ratio correction factor can be adjusted to be smaller appropriately, and it can also be seen in combination with fig. 3 that the larger the azimuth angle is, the smaller the stable value of the final convergence of the filter correction factor is. In other words, in the case of radial distance determination, the larger the azimuth angle, the more the tracking result is relatively emphasized on the current motion state measurement value, so that continuous tracking of the target can be realized.

And S130, generating a dynamic filter ratio of the current tracking target according to the initial filter ratio and the filter ratio correction factor.

And the dynamic filtering ratio represents the fusion ratio of the current motion state predicted value and the current motion state measured value of the current tracking target.

In one embodiment, the dynamic filter ratio may be obtained according to the following formula:

wherein k represents a dynamic filter ratio, factor represents a filter ratio correction factor,for the initial filter ratio of this filtering process, alpha_iRepresenting the filtered mean error, beta_iRepresenting the velocity filtered mean error. Obviously, other functional relations capable of presenting the characteristic of the formula 2 can also be applied to the embodiment of the present invention. Referring to fig. 4, a schematic diagram of a curve of a dynamic filter ratio provided by an embodiment of the present invention along with a change of a spatial distance of a tracked target is shown, as shown in fig. 4, a curve 1 is a corresponding curve when a conventional filter ratio is set to a fixed value, and a curve 2 is a corresponding curve of the dynamic filter ratio provided by the present invention.

As can be seen from the two curves shown in fig. 4, in the conventional manner of setting the filter ratio to a fixed value, the filter ratio is fixed at different target spatial distances. The dynamic filtering ratio of the invention is positively correlated with the target space distance.

Obviously, the dynamic filtering ratio setting of the invention is more in line with the actual tracking requirements. In this case, it is obviously necessary to set a larger filter ratio for suppressing the influence of measurement noise and realizing the stability of target tracking.

And S140, comparing the current motion state predicted value and the current motion state measured value of the current tracking target by using dynamic filtering, and performing filtering processing to obtain target state data of the current tracking target.

In one embodiment of the invention, the α - β filtering process is performed according to the following equation:

X＝k·X_pre+X_m(formula 3)

In formula 3, X represents the result after the filtering process, X_preA current motion state prediction value, X, representing a current tracked target_mRepresents the current motion state measurement value of the current tracking target, and k represents the dynamic filtering ratio.

In one embodiment of the invention, the combinationThe motion characteristic of the static target, the target motion state can be generally defined as a 4-dimensional vector, and the motion state predicted value X of the tracking target of the current frame can be obtained by combining the motion state transition model of the tracking target_pre＝[x_p,v_xp,y_p,v_yp]^TWherein x is_pRepresenting the component of the spatial distance of the tracked object in the x-axis, v_xpRepresenting the velocity component, y, of the velocity vector of the tracked object in the x-axis_pRepresenting the component of the spatial distance of the tracked object in the y-axis, v_ypRepresenting the velocity component of the tracked object in the y-axis.

Usually, the motion state measured value of the radar to the tracking target is a value under a polar coordinate system, and the assumed radar measured data is represented as X under the polar coordinate system_m＝[r,v,θ]^TWherein r represents the radial distance of the tracking target from the radar, v represents the radial velocity component of the real-time velocity of the tracking target, and θ is the azimuth angle of the current tracking target relative to the radar.

In practical application, the measured value of the motion state of the tracking target needs to be subjected to coordinate transformation and converted into a coordinate system shown as X in fig. 2_m＝[x_m,v_xm,y_m,v_ym]^T。

According to the target tracking data filtering processing method provided by the embodiment, different initial filtering ratios are set according to different motion states of the self-vehicle and the current tracking target. And, the filter ratio correction factor is determined based on the relative spatial position information between the current tracked target and the own vehicle, and the initial filter ratio is corrected by using the filter ratio correction factor, so that the finally obtained dynamic filter ratio is positively correlated with the filter ratio correction factor, and the filter ratio correction factor gradually increases and approaches a certain convergence value as the spatial distance increases, that is, the dynamic filter ratio increases with the increase of the spatial distance and approaches a convergence value. As the distance of the currently tracked target increases, the measurement noise also increases, and a larger filter ratio is set to suppress the influence of the measurement noise. Moreover, the filter ratio correction factor with the convergence characteristic can avoid the phenomenon that the actual position of the tracking target is really changed, and the tracking real-time performance is deteriorated due to a larger filter ratio, so that the real-time performance and continuity of target tracking are finally realized, and the track of the tracking target obtained by tracking processing is smooth and stable.

In an exemplary embodiment of the present invention, the preset corresponding relationship between the motion state of the vehicle and the track tracking stable stage information of the current tracking target and the initial filter ratio may be the relationship shown in table 1, and obviously, in practical application, the method is not limited to the manner shown in the example.

TABLE 1

In table 1, ω is a lateral oscillation angular velocity of the host vehicle, and is used to determine whether the host vehicle is in a turning state; and tick is the frame number of continuous tracking of the target in the process of tracking the current tracking target and is used for representing the stability of tracking the tracking target, and the larger the value is, the more stable the tracking of the target is represented. Combining these two parameters allows setting the initial value of the filter ratio, i.e. the initial filter ratio, according to different situations. The specific values of the thresholds of the two parameters in table 1 can be obtained by statistics according to the data measured by the actual sensors, such as the vehicle motion state discrimination threshold ω₁、ω₂The specific value of the transverse swinging angular velocity for judging whether the vehicle is in turning or not has a certain difference among different vehicle types.

1) The bicycle is in a turning state

When omega > omega₁Where the motion data of the vehicle is statistically known, ω₁Usually 0.05rad/s, in this case, the motion relationship between the self-vehicle and the tracked target is complex, it is difficult to satisfy a single motion state transition model, and the accuracy of the predicted value of the current motion state obtained by predicting according to the motion state transition model is difficult to guarantee, so that the initial filtering ratio is set to a relatively small value, for example, k₀＝α₁/β₁1-3. The filter ratio can realize the rapid tracking convergence of the tracked target, and avoid the target tracking loss or the track pseudo convergence of the targetThe phenomenon of (2).

2) The bicycle is in a state of turning over or changing lanes

When ω is₂＜ω＜ω₁During the running, the self-vehicle is determined to be in the excessive turning state or the lane changing state, and the motion data of the self-vehicle is counted to obtain omega₂Typically 0.02 rad/s.

In this case, although the own vehicle has a certain yaw angular velocity, the relative motion relationship between the own vehicle and the tracking target is not complicated, and therefore, the track stability of the tracking target, that is, the tick parameter in table 1 can be appropriately considered when setting the initial filter ratio.

If tick < tick₂In general tick₂10-15, the tracked object is considered to be in the first initial stage, i.e. the track stability of the tracked object is not high, in which case the initial filter ratio needs to be set with a certain emphasis on the measured values, e.g. k₀＝α₂₁/β₂₁＝3～5。

If tick is greater than or equal to tick₂Then, the tracked target is considered to be in the first stable stage, i.e. the track stability of the tracked target is high, in this case, the initial filtering ratio should be set to focus on the predicted value, for example, the initial filtering ratio is usually k₀＝α₃₁/β₃₁＝5～7。

3) The bicycle is in a straight-going state

When omega < omega₂When the initial filtering ratio is set, track stability of the tracked target is considered, but the track stability does not need to be emphasized to a measured value to a certain extent, and thus the track obtained by target tracking has high stability.

If tick < tick₃₁In general tick₃₁And considering that the tracking target is in the second initial stage and the stability is not high, in this case, setting the initial filtering ratio may emphasize the measurement value to a certain extent, but is not suitable for too much measuringValues, too many of which tend to cause significant track jitter at the start of the flight, e.g. initial filter ratio k₀Can be set to be 4-5.

If tick₃₁≤tick＜tick₃₂In general tick₃₂Considering the tracking target in a relatively stable stage 10-25, in this case, the initial filtering ratio should be set to be appropriate amount to focus on the predicted value, for example, k₀Can be set to 5-7.

If tick is greater than or equal to tick₃₂Then the tracked target is considered to be in the second stable phase, in which case setting the initial filter ratio should focus on the predicted value, e.g., k₀The setting can be 7 ~ 9.

Based on the initial filter ratio setting criterion, the motion state of the vehicle and the track tracking stability stage information of the tracked target are combined, and a proper initial filter ratio kappa is selected according to the corresponding relation shown in table 1₀＝α_i:β_i。

According to the initial filter ratio setting criterion, the initial filter ratio set by the method can be dynamically adjusted according to the motion state of the vehicle and the track tracking stable stage information of the tracked target. Obviously, the correspondence between the motion state of the host vehicle and the track tracking stabilization phase information of the tracked target and the initial filter ratio may also be represented by a function capable of presenting the relationship characteristics shown in table 1, for example, a piecewise function. Moreover, the preset corresponding relationship is not limited to the relationship shown in table 1, and other corresponding relationships may be adopted as long as the finally obtained filtering ratio can obtain the technical effect of the embodiment of the present invention.

Corresponding to the embodiment of the target tracking data filtering processing method, the invention also provides an embodiment of a target tracking data filtering processing device.

Referring to fig. 5, a schematic structural diagram of a target tracking data filtering processing apparatus according to an embodiment of the present invention is shown, and as shown in fig. 5, the apparatus may include: an initial filter ratio determination module 110, a filter ratio correction factor determination module 120, a dynamic filter ratio generation module 130, and a filter processing module 140.

And the initial filtering ratio determining module 110 is configured to determine an initial filtering ratio of an α - β filtering algorithm according to the motion state of the vehicle and the track tracking stabilization phase information of the current tracking target.

In one embodiment of the present invention, as shown in fig. 6, the initial filter ratio determining module 110 includes: a first initial filter ratio determination submodule 111, a second initial filter ratio determination submodule 112 and a third initial filter ratio determination submodule 113.

The first initial filter ratio determining submodule 111 is configured to determine the initial filter ratio of the α - β filter algorithm as the first filter ratio when the motion state of the host vehicle is the turning state.

Wherein the first filtering ratio emphasizes the target state data over the current motion state measurement.

And the second initial filter ratio determining submodule 112 is configured to determine a motion stage where the current tracking target is located according to the track tracking stable stage information when the motion state of the self vehicle is the turning transition state or the lane change state, and determine an initial filter ratio of the alpha-beta filter algorithm according to the motion stage where the current tracking target is located.

When the motion state of the self-vehicle is a turning transition state or a lane changing state, the motion phase of the current tracking target is divided into a first initial phase and a first stable phase.

When the current tracking target is in a first initial stage, determining the initial filtering ratio as a second filtering ratio, wherein the second filtering ratio enables the target state data to be heavier than the current motion state measured value;

and when the current tracking target is in the first stable stage, determining the initial filtering ratio as a third filtering ratio, wherein the third filtering ratio is larger than the second filtering ratio.

In one embodiment of the present invention, when the target continuous tracking frame number of the current tracking target is smaller than the first frame number threshold, it is determined that the current tracking target is currently in the first initial stage; when the target continuous tracking frame number of the current tracking target is greater than or equal to the first frame number threshold value, determining that the current tracking target is in a first stable stage.

And a third initial filter ratio determining submodule 113, configured to determine, when the motion state of the host vehicle is a straight-ahead state, a motion stage where the current tracking target is located according to the track tracking stabilization stage information, and determine, according to the motion stage where the current tracking target is located, an initial filter ratio of the α - β filter algorithm.

When the motion state of the self-vehicle is a straight-going state, the motion phase of the current tracking target is divided into a second starting phase, a relative stable phase and a second stable phase.

When the motion stage of the current tracked target is in a second initial stage, determining the initial filtering ratio as a fourth filtering ratio, wherein the fourth filtering ratio enables the target state data to be heavier than the current motion state measured value;

when the motion stage of the current tracking target is in a relatively stable stage, determining the initial filtering ratio as a fifth filtering ratio, wherein the fifth filtering ratio is greater than the fourth filtering ratio;

and when the motion phase of the current tracking target is in a second stable phase, determining the initial filtering ratio as a sixth filtering ratio, wherein the sixth filtering ratio is larger than the fifth filtering ratio.

In an embodiment of the present invention, when the target continuous tracking frame number of the current tracking target is smaller than the second frame number threshold, it is determined that the current tracking target is currently in the second initial stage; when the target continuous tracking frame number of the current tracking target is greater than or equal to the second frame number threshold and smaller than the third frame number threshold, determining that the current tracking target is in a relatively stable stage; and when the target continuous tracking frame number of the current tracking target is greater than or equal to the third frame number threshold, determining that the current tracking target is in a second stable stage.

And a filter ratio correction factor determining module 120, configured to determine a filter ratio correction factor according to the relative spatial position information of the vehicle and the current tracking target.

The relative spatial position information comprises an spatial distance and an azimuth angle, the filter ratio correction factor is gradually increased and converged along with the increase of the spatial distance under the same azimuth angle, the larger the azimuth angle is, the smaller the filter ratio correction factor is, the azimuth angle is the angle of the current tracking target detected by the vehicle-mounted radar, and the spatial distance is the radial distance between the current tracking target and the vehicle-mounted radar.

In an embodiment of the present invention, the filter ratio correction factor determining module 120 is specifically configured to:

and calculating a filter ratio correction factor according to the following formula:

wherein, range is the space distance, a, b are the adjustment coefficient respectively, theta is the azimuth.

And a dynamic filter ratio generating module 130, configured to generate a dynamic filter ratio of the current tracking target according to the initial filter ratio and the filter ratio correction factor.

The dynamic filter ratio is positively correlated with the filter ratio modifier.

In an embodiment of the present invention, the dynamic filter ratio generating module 130 is specifically configured to:

and calculating the dynamic filtering ratio according to the following formula:

wherein k represents a dynamic filter ratio, factor represents a filter ratio correction factor,is the initial filter ratio of the current filtering process.

And the filtering processing module 140 is configured to perform filtering processing on the current motion state predicted value and the current motion state measured value of the current tracking target by using dynamic filtering to obtain target state data of the current tracking target.

The filtering processing module 140 is specifically configured to: and calculating the product of the dynamic filtering ratio and the predicted value of the current motion state, and calculating the sum of the product and the measured value of the current motion state to obtain target state data.

The target tracking data filtering processing device provided by the embodiment sets different initial filtering ratios according to different motion states of the self-vehicle and the current tracking target. And, the filter ratio correction factor is determined based on the relative spatial position information between the current tracked target and the own vehicle, and the initial filter ratio is corrected by using the filter ratio correction factor, so that the finally obtained dynamic filter ratio is positively correlated with the filter ratio correction factor, and the filter ratio correction factor gradually increases and approaches a certain convergence value as the spatial distance increases, that is, the dynamic filter ratio increases with the increase of the spatial distance and approaches a convergence value. As the distance of the currently tracked target increases, the measurement noise also increases, and a larger filter ratio is set to suppress the influence of the measurement noise. Moreover, the filter ratio correction factor with the convergence characteristic can avoid the phenomenon that the actual position of the tracking target is really changed, and the tracking real-time performance is deteriorated due to a larger filter ratio, so that the real-time performance and continuity of target tracking are finally realized, and the track of the tracking target obtained by tracking processing is smooth and stable.

The invention provides an apparatus comprising a processor and a memory having stored therein a program executable on the processor. The processor implements the target tracking data filtering processing method provided in the above embodiment when running the program stored in the memory.

The present invention also provides a storage medium executable by a computing device, the storage medium storing a program, the program implementing the above-described target tracking data filtering processing method when executed by the computing device.

The present invention also provides a computer program which, when executed by an apparatus, implements the above-described target tracking data filtering processing method.

While, for purposes of simplicity of explanation, the foregoing method embodiments have been described as a series of acts or combination of acts, it will be appreciated by those skilled in the art that the present invention is not limited by the illustrated ordering of acts, as some steps may occur in other orders or concurrently with other steps in accordance with the invention. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required by the invention.

It should be noted that technical features described in the embodiments in the present specification may be replaced or combined with each other, each embodiment is mainly described as a difference from the other embodiments, and the same and similar parts between the embodiments may be referred to each other. For the device-like embodiment, since it is basically similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

The steps in the method of each embodiment of the invention can be sequentially adjusted, combined and deleted according to actual needs.

The device and the modules and sub-modules in the terminal in the embodiments of the present invention can be combined, divided and deleted according to actual needs.

In the embodiments provided in the present invention, it should be understood that the disclosed terminal, apparatus and method may be implemented in other ways. For example, the above-described terminal embodiments are merely illustrative, and for example, the division of a module or a sub-module is only one logical division, and there may be other divisions when the terminal is actually implemented, for example, a plurality of sub-modules or modules may be combined or integrated into another module, or some features may be omitted or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or modules, and may be in an electrical, mechanical or other form.

The modules or sub-modules described as separate parts may or may not be physically separate, and parts that are modules or sub-modules may or may not be physical modules or sub-modules, may be located in one place, or may be distributed over a plurality of network modules or sub-modules. Some or all of the modules or sub-modules can be selected according to actual needs to achieve the purpose of the solution of the present embodiment.

In addition, each functional module or sub-module in each embodiment of the present invention may be integrated into one processing module, or each module or sub-module may exist alone physically, or two or more modules or sub-modules may be integrated into one module. The integrated modules or sub-modules may be implemented in the form of hardware, or may be implemented in the form of software functional modules or sub-modules.

Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.