CN111332298A

CN111332298A - Method, device and equipment for determining travelable area and storage medium

Info

Publication number: CN111332298A
Application number: CN202010103250.3A
Authority: CN
Inventors: 陈至元; 付骁鑫; 李旭健; 朱振广; 马霖
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2020-02-19
Filing date: 2020-02-19
Publication date: 2020-06-26
Anticipated expiration: 2040-02-19
Also published as: CN111332298B

Abstract

The application discloses a method, a device, equipment and a storage medium for determining a travelable area, and relates to the field of automatic driving. The specific implementation scheme is as follows: the method is applied to electronic equipment, a sensing system is mounted on a target vehicle, the electronic equipment is communicated with the sensing system, and the method comprises the following steps: acquiring current frame environmental data of a target vehicle acquired by a sensing system; if the adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extracting current frame environment characteristic data on the adjacent lane; predicting the current frame danger of the adjacent lane according to the current frame environment characteristic data to obtain a current frame initial danger value; calculating a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame; and calculating the current frame drivable area of the adjacent lane according to the final danger value of the current frame. The target vehicle cannot be driven in a hurry direction, and the stability of the target vehicle and the riding experience of a user are improved.

Description

Method, device and equipment for determining travelable area and storage medium

Technical Field

The application relates to the technical field of data processing, in particular to an automatic driving technology.

Background

With the maturity of artificial intelligence technology, the automatic driving technology has also been developed rapidly. In the automatic driving technique, the reasonable determination of the drivable area is an important issue regarding the driving safety.

When the travelable area is determined, if the target vehicle meets the front obstacle and needs to avoid the obstacle, the space of the adjacent lane needs to be borrowed temporarily.

In the prior art, when the situation is faced, the drivable area of the adjacent lane is determined according to the expansion strategy or the contraction strategy by independently judging each frame of the target vehicle in the driving process through manually writing a rule, so that the drivable area of the adjacent lane is suddenly changed to a large extent when the expansion strategy and the contraction strategy are switched in the driving process of the target vehicle, the target vehicle is easy to slam, the stability of the target vehicle is poor, and the riding experience of a user is poor.

Disclosure of Invention

The embodiment of the application provides a method, a device, equipment and a storage medium for determining a drivable area, and solves the technical problems that in the prior art, when switching between an expansion strategy and a contraction strategy is carried out in the driving process of a target vehicle, the drivable area of an adjacent lane can be suddenly changed to a large extent, the target vehicle is likely to slam to the direction, the stability of the target vehicle is poor, and the riding experience of a user is poor.

A first aspect of an embodiment of the present application provides a travelable area determining method, where the method is applied to an electronic device, a sensing system is mounted on a target vehicle, and the electronic device communicates with the sensing system, and the method includes:

acquiring current frame environmental data of the target vehicle acquired by the perception system; if the adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extracting current frame environment characteristic data on the adjacent lane; predicting the current frame danger of the adjacent lane according to the current frame environment characteristic data to obtain a current frame initial danger value; calculating a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame; and calculating the current frame drivable area of the adjacent lane according to the current frame final danger value.

In the embodiment of the application, after the current frame danger of the adjacent lane is predicted according to the current frame environment characteristic data to obtain the current frame initial danger value, the current frame initial danger value and the corresponding previous frame danger reference value are fused to calculate the current frame final danger value, and the current frame drivable area of the adjacent lane is calculated according to the current frame final danger value. Therefore, the final danger value of each frame is fused with the danger reference value of the previous frame, so that the final danger value between two adjacent frames does not jump but is changed gradually, and further, the determined each frame drivable area of the adjacent lanes is also changed gradually rather than being changed suddenly by a large amplitude. And then the target vehicle can not be driven to make a violent movement, and the stability of the target vehicle and the riding experience of a user are improved.

Further, the method as described above, the predicting a current frame risk of the adjacent lane according to the current frame environmental characteristic data to obtain a current frame initial risk value includes:

inputting the current frame environmental characteristic data into a danger prediction model trained to be convergent so as to predict the danger of the current frame by adopting the danger prediction model trained to be convergent; and outputting the current frame initial risk value through the risk prediction model trained to be convergent.

In the embodiment of the application, the danger of the adjacent lane is predicted by adopting a danger prediction model trained to be convergent according to the environmental characteristic data, and then the current frame initial danger value of the adjacent lane is obtained. As long as the data of the training samples for training the risk prediction model are comprehensive enough, the prediction effect of the risk prediction model is ensured. If the target vehicle encounters a scene which is not seen in the driving process and causes poor performance, the accuracy of the prediction of the initial danger value of the current frame can be improved as long as the environmental characteristic data corresponding to the new scene is added into the training set to train the danger prediction model again, so that the accuracy of the determination of the drivable area is improved, and the maintenance cost can be controlled.

Further, in the method described above, the risk prediction model is a generalized linear logistic regression model.

In the embodiment of the application, because the generalized linear logistic regression model does not forcibly change the natural measurement of data, the predicted initial risk value of the current frame can be kept stable under the condition that the training sample is not changed much, and the jump is not easy to occur.

Further, the method as described above, before inputting the current frame environmental feature data into the risk prediction model trained to converge, further includes:

acquiring training samples corresponding to a danger prediction model, wherein the training samples are each frame of historical environment characteristic data for carrying out danger labeling on the adjacent lanes; training an initial risk prediction model by adopting the training sample; and if the trained danger prediction model meets the preset model convergence condition, determining the danger prediction model meeting the preset model convergence condition as the danger prediction model from training to convergence.

In the embodiment of the application, each frame of environmental characteristic data on adjacent roads is extracted as a training sample through the actual environment of the target vehicle in the running process, so that the training sample data are comprehensive enough, and the accuracy of prediction of the initial danger value of the current frame is improved when a danger prediction model from training to convergence is adopted for prediction.

Further, the method as described above, wherein the previous frame hazard reference value comprises: a final risk value of a previous frame and a decision value of the previous frame; the calculating the final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame comprises the following steps:

determining the decision value of the previous frame according to the final danger value of the previous frame; determining weights corresponding to the initial risk value of the current frame, the final risk value of the previous frame and the decision value of the previous frame; and performing weighted summation calculation on the current frame initial risk value, the previous frame final risk value and the previous frame decision value to obtain the current frame final risk value.

In the embodiment of the present application, the risk reference value of the previous frame includes: and when the final risk value of the current frame is calculated, the product of the final risk value of the previous frame and the weight enables the decision switching to be more stable, and the product of the final risk value of the previous frame and the weight pulls the final risk value of the current frame to the same direction as the final risk value of the previous frame, so that the final risk value of each frame is not easy to jump but is a gradually changing process, the target vehicle cannot rush to the direction, the anti-noise performance of the target vehicle is further improved, and the stability of the target vehicle and the riding experience of a user are further improved.

Further, the method as described above, the determining the previous frame decision value according to the previous frame final risk value includes:

acquiring a final danger value of the previous frame; if the final risk value of the previous frame is greater than a preset decision threshold, determining that the decision value of the previous frame is a first numerical value; and if the final risk value of the previous frame is less than or equal to the preset decision threshold, determining the decision value of the previous frame as a second numerical value.

In the embodiment of the application, the decision value of the previous frame is calculated by adopting the final risk value of the previous frame, and the decision value of the previous frame is only provided with two numerical values, so that the product term of the decision value of the previous frame and the weight can better play a role of pulling in the same direction as the final risk value of the previous frame.

Further, in the method as described above, the weight corresponding to the decision value of the previous frame is smaller than the weight corresponding to the initial risk value of the current frame.

In the embodiment of the application, the weight corresponding to the decision value of the previous frame is set to be smaller than the weight corresponding to the initial danger value of the current frame, so that decision switching can be performed on the same obstacle in the same adjacent lane, and the problem that all judgment except the final danger value of the first frame is invalid is avoided.

Further, the method as described above, the calculating a current frame travelable region of the adjacent lane according to the current frame final risk value includes:

acquiring a preset continuous function, wherein the preset continuous function represents the mapping relation between the final danger value of the current frame and the drivable area of the current frame; and inputting the final danger value of the current frame into the preset continuous function to calculate the travelable region of the current frame.

In the embodiment of the application, the mapping relation between the final risk value of the current frame and the travelable area of the current frame is set as the preset continuous function, so that the travelable area of the current frame is also a gradually changing process in the process that the final risk value of each frame is gradually changed, and the stability of the target vehicle and the riding experience of a user are further improved.

Further, in the method as described above, the current frame environmental characteristic data includes any one or more of the following:

the direction deviation of the obstacle in the current frame from the current driving lane, the width of the obstacle in the current frame, the longitudinal speed of the obstacle in the current frame, the longitudinal relative speed of the obstacle in the current frame and the target vehicle, the longitudinal distance between the obstacle in the current frame and the target vehicle, and the transverse distance between the obstacle in the current frame and the target vehicle.

In the embodiment of the application, the current frame environmental characteristic data comprises any one or more of the above, so that the current frame environmental data can accurately express the relationship between the target vehicle and the obstacle of the adjacent lane, and the current frame danger of the adjacent lane can be more accurately predicted according to the current frame environmental characteristic data.

A second aspect of the embodiments of the present application provides a travelable area determining apparatus, where the apparatus is located in an electronic device, a sensing system is mounted on a target vehicle, and the electronic device communicates with the sensing system, and the apparatus includes:

the environment data acquisition module is used for acquiring the current frame environment data of the target vehicle acquired by the perception system;

the characteristic data extraction module is used for extracting current frame environmental characteristic data on an adjacent lane if the adjacent lane exists on the current driving lane of the target vehicle according to the current frame environmental data;

the initial danger value prediction module is used for predicting the current frame danger of the adjacent lane according to the current frame environment characteristic data so as to obtain a current frame initial danger value;

the final danger value calculation module is used for calculating a final danger value of the current frame according to the initial danger value of the current frame and the corresponding danger reference value of the previous frame;

and the area calculation module is used for calculating the current frame drivable area of the adjacent lane according to the current frame final danger value.

Further, in the above apparatus, the initial risk value prediction module is specifically configured to:

Further, in the above-described apparatus, the risk prediction model is a generalized linear logistic regression model.

Further, the apparatus as described above, further comprising: a model training module to:

Further, the apparatus as described above, the previous frame hazard reference value comprising: a final risk value of a previous frame and a decision value of the previous frame;

the final risk value calculation module is specifically configured to:

Further, in the above apparatus, when determining the decision value of the previous frame according to the final risk value of the previous frame, the final risk value calculation module is specifically configured to:

Further, in the apparatus as described above, the weight corresponding to the decision value of the previous frame is smaller than the weight corresponding to the initial risk value of the current frame.

Further, in the apparatus as described above, the region calculating module is specifically configured to:

Further, in the apparatus as described above, the current frame environmental characteristic data includes any one or more of the following:

A third aspect of the embodiments of the present application provides an electronic device, including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of the first aspects.

A fourth aspect of embodiments of the present application provides a non-transitory computer-readable storage medium having stored thereon computer instructions for causing a computer to perform the method of any one of the first aspects.

A fifth aspect of embodiments of the present application provides a computer program comprising program code for performing the method according to the first aspect when the computer program is run by a computer.

Drawings

The drawings are included to provide a better understanding of the present solution and are not intended to limit the present application. Wherein:

fig. 1 is a first scene diagram of a travelable region determination method that can implement an embodiment of the present application;

fig. 2 is a second scene diagram of a travelable region determination method that can implement an embodiment of the present application;

fig. 3 is a schematic flowchart of a travelable region determination method according to a first embodiment of the present application;

fig. 4 is a schematic flowchart of a travelable region determination method according to a second embodiment of the present application;

fig. 5 is a schematic flowchart of step 201 in a travelable region determining method according to a second embodiment of the present application;

fig. 6 is a schematic flowchart of step 206 in a travelable region determination method according to a second embodiment of the present application;

fig. 7 is a signaling flowchart of a travelable region determination method according to a third embodiment of the present application;

fig. 8 is a schematic structural diagram of a travelable region determination apparatus according to a fourth embodiment of the present application;

fig. 9 is a schematic structural view of a travelable region determination apparatus according to a fifth embodiment of the present application;

fig. 10 is a block diagram of an electronic device for implementing the travelable region determination method according to the embodiment of the present application.

Detailed Description

The following description of the exemplary embodiments of the present application, taken in conjunction with the accompanying drawings, includes various details of the embodiments of the application for the understanding of the same, which are to be considered exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

For a clear understanding of the technical solutions of the present application, a detailed description of the prior art solutions is first provided. When the automatic driving vehicle runs on a road, the sensing system collects environmental data around the target vehicle according to the collection period, the front decision-making system determines a travelable area according to each frame of environmental data, and when the travelable area is determined, if no barrier is in front of the vehicle, the travelable area in the current travelling lane of the target vehicle can be determined. However, in the driving process of the target vehicle, the situation that an obstacle exists on the current driving lane where the target vehicle is located is often encountered. If the obstacle is a vehicle parked illegally or a construction cone drum and the like. When determining the travelable region in this case, therefore, it is also necessary to determine the travelable region of the adjacent lane. In the prior art, in order to determine the drivable area in the adjacent lane, the drivable area of the adjacent lane is generally determined according to a manually written rule by setting the manually written rule into a front decision system of the target vehicle. When the travelable area of the adjacent lane is determined, complex conditions such as whether the adjacent lane has side-by-side vehicles or not and whether the rear side vehicle of the target vehicle in the adjacent lane collides with the target vehicle or not are comprehensively considered, so that the manually written rule may have dozens or even dozens of rules, the rule is complex, and the maintenance cost is high. However, this also fails to exhaustively enumerate all possible road conditions in advance, resulting in poor performance when unexpected situations are encountered.

Secondly, in the prior art, when the drivable area of the adjacent lane is determined by adopting a manually written rule, whether the drivable area of the adjacent lane is determined by independently judging whether each frame of the target vehicle is according to an expansion strategy or a contraction strategy in the driving process is independently judged. The determination strategy of the travelable region is made discontinuous. The drivable area of the adjacent lane can be suddenly changed to a large extent when the expansion strategy and the contraction strategy are switched in the driving process of the target vehicle, the target vehicle is easy to hurry, the stability of the target vehicle is poor, and the riding experience of a user is poor.

Moreover, since the determination strategy of the travelable region in the prior art is discontinuous, if the determination strategy of the previous frame is the first strategy, and when the next frame is judged by the rule, and the environmental data of the next frame collected by the target vehicle sensing system is deviated and just meets the second strategy, switching between the two strategies is performed, the travelable region of the adjacent lane is suddenly changed by a large amplitude, and the target vehicle slamming direction is easily caused. The determination strategy of the travelable region in the prior art is also poor in noise immunity.

Therefore, aiming at the technical problem of the method for determining the travelable region in the prior art, the inventor finds that, in the research, in order to avoid the problem that the travelable region of the adjacent lane is suddenly changed by a large amplitude when the discontinuous strategies are switched in the traveling process of the target vehicle, the target vehicle hurry direction and the anti-noise performance are poor. The travelable region determination strategy determined for the current frame may be fused with the travelable region determination strategy determined for the previous frame. Making the determination strategy of each frame a gradual process. Alternatively, the decision strategy as per frame may employ the final hazard representation per frame on the adjacent lanes. In such a gradual change travelable region determination strategy, the travelable region is also a gradual change process, rather than a sudden change of a large magnitude. And then the target vehicle can not make the target vehicle fierce to hit the direction, the anti-noise performance of the target vehicle is improved, and the stationarity of the target vehicle and the riding experience of a user are also improved.

In order to avoid the problems that all possible road conditions cannot be listed in detail in advance due to the manually written rules, the performance is often poor when the conditions are not predicted in advance, and the rules are complicated and the maintenance cost is high. The inventor further researches and discovers that environmental characteristic data on adjacent roads can be extracted through the actual environment of the target vehicle in the running process, and the danger of the adjacent lanes is predicted according to the environmental characteristic data by adopting a danger prediction model trained to be convergent, so that a determination strategy of the travelable area of the adjacent lanes is obtained. Alternatively, the decision strategy per frame may employ the final hazard representation per frame on the adjacent lanes. In the mode, as long as the data of the training samples for training the danger prediction model are comprehensive enough, the prediction effect of the danger prediction model is ensured. If the target vehicle encounters a scene which is not seen in the driving process and causes poor performance, the accuracy of determining the drivable area can be improved and the maintenance cost can be controlled as long as the environmental characteristic data corresponding to the new scene is added into the training set to train the danger prediction model again.

The inventor proposes a technical scheme of the application based on the creative discovery. An application scenario of the travelable area determination method provided by the embodiment of the present application is described below. As shown in fig. 1, an application scenario corresponding to the travelable region determining method provided in the embodiment of the present application includes: an electronic device 1 and a target vehicle 2. A sensing system may be mounted on the target vehicle 2. The perception system is used for periodically collecting current frame environment data of the target vehicle. A decision making system may also be mounted on the target vehicle 2. The electronic device 1 communicates with the perception system and the decision making system of the target vehicle 2. The electronic equipment 1 acquires current frame environmental data of a target vehicle acquired by a sensing system, and if adjacent lanes exist on a current driving lane of the target vehicle according to the current frame environmental data, extracting current frame environmental characteristic data on the adjacent lanes; predicting the current frame danger of the adjacent lane according to the current frame environment characteristic data to obtain a current frame initial danger value; calculating a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame; and calculating the current frame drivable area of the adjacent lane according to the final danger value of the current frame. As shown in fig. 1, the target vehicle 2 is currently running in the middle lane, and the electronic device 1 acquires the current frame environmental data of the target vehicle collected by the sensing system, and determines that the adjacent lanes existing on the current running lane of the target vehicle 2 include a left lane and a right lane. Extracting the current frame environmental characteristic data on the left lane and the right lane respectively. And respectively predicting the current frame dangerousness of the left lane and the right lane according to the current frame environment characteristic data so as to respectively obtain current frame initial danger values of the left lane and the right lane. And calculating a current frame final danger value of the left lane according to the current frame initial danger value of the left lane and the corresponding previous frame danger reference value, and calculating a current frame drivable area of the left lane according to the current frame final danger value of the left lane. And calculating a current frame final danger value of the right lane according to the current frame initial danger value of the right lane and the corresponding previous frame danger reference value, and calculating a current frame travelable area of the right lane according to the current frame final danger value of the right lane. Because the sensing system behind the left lane does not acquire the obstacle information, the sensing system behind the right lane acquires the obstacle information, and the obstacle is the vehicle 3. Therefore, the transverse distance between the determined current travelable region of the left lane and the nearest boundary of the current traveling lane of the target vehicle is 2 meters, and the transverse distance between the determined current travelable region of the right lane and the nearest boundary of the current traveling lane of the target vehicle is 1 meter. After the current travelable area of the adjacent lane is determined, the determined current travelable area of the adjacent lane is sent to a decision-making system, and the decision-making system plans the travelable path of the target vehicle according to the current travelable area of the adjacent lane and the current travelable area on the current travelling lane of the target vehicle, so that the target vehicle travels according to the travelable path. As in fig. 1, since there is no obstacle ahead on the current travel lane of the target vehicle, although the left lane has a larger current-frame travelable area, the travelable path 51 planned by the target vehicle does not pass through the left lane, but passes on the current travel lane of the target vehicle 2. After the sensing system collects the next frame of environmental data, the target vehicle 2 still runs on the intermediate road, so the electronic device determines that the left lane and the right lane still exist on the current running lane of the target vehicle. Since the sensing system does not acquire the obstacle information yet behind the left lane, the lateral distance between the current travelable area of the determined left lane and the nearest boundary of the current traveling lane of the target vehicle is 2 meters, and since the speed of the obstacle behind the right lane is slow, the sensing system does not acquire the obstacle information, that is, the information of the vehicle 3, yet. However, in the solution of the present application, when the electronic device calculates the final risk value of the next frame according to the initial risk value of the next frame and the corresponding risk reference value of the current frame, the final risk value of the next frame does not jump but has a small difference with the final risk value of the current frame, and the final risk value of each frame is a gradual change process. The lateral distance between the next frame travelable region of the determined right lane and the nearest boundary of the current traveling lane of the target vehicle is 1.2 meters. The determination of the travelable region between two frames is a gradual process.

Fig. 2 is a second scenario diagram for implementing the travelable region determining method according to the embodiment of the present application, and as shown in fig. 2, in the application scenario shown in fig. 2, different from the application scenario shown in fig. 1, the electronic device 1 in fig. 2 determines that there is only a left lane in an adjacent lane on the current traveling lane of the target vehicle 2 according to the current frame environment data, and then extracts the current frame environment feature data on the left lane. And predicting the current frame danger of the left lane according to the current frame environment characteristic data to obtain the initial danger value of the current frame of the left lane. And calculating a current frame final danger value of the left lane according to the current frame initial danger value of the left lane and the corresponding previous frame danger reference value, and calculating a current frame drivable area of the left lane according to the current frame final danger value of the left lane. The determined nearest boundary between the current travelable region of the left lane and the current travelling lane of the target vehicle 2 is 2 meters. After the current travelable area of the left lane is determined, the determined current travelable area of the left lane is sent to the decision-making system, and the decision-making system plans the travelable path of the target vehicle according to the current travelable area of the left lane and the current travelable area on the current travelling lane of the target vehicle 2, so that the target vehicle travels according to the travelable path. As shown in fig. 2, since the target vehicle 2 has an obstacle in the front 1m of the current driving lane, the obstacle is a construction cone 4. Therefore, the planned travelable path of the target vehicle passes through the current frame travelable region of the left lane, and the target vehicle travels on the current frame travelable path 52 that turns into the left lane. When the sensing system collects the next frame of environment data, the target vehicle is already shifted to the left lane to drive, so that the adjacent lane existing on the next frame of driving lane of the target vehicle is determined to be the right lane according to the next frame of environment data. The process for determining the next frame of drivable area is similar to the process for determining the current frame of drivable area, and is not repeated here, for example, it is determined that the nearest boundary between the next frame of drivable area of the right lane and the current driving lane of the target vehicle is 1 meter.

According to the method and the device, after the current frame danger of the adjacent lane is predicted according to the current frame environment characteristic data to obtain the current frame initial danger value, the current frame initial danger value and the corresponding previous frame danger reference value are fused to calculate the current frame final danger value, and the current frame drivable area of the adjacent lane is calculated according to the current frame final danger value. Therefore, the final danger value of each frame is fused with the danger reference value of the previous frame, so that the final danger value between two adjacent frames does not jump but is changed gradually, and further, the determined each frame drivable area of the adjacent lanes is also changed gradually rather than being changed suddenly by a large amplitude. And then the target vehicle can not make the target vehicle fierce to hit the direction, the anti-noise performance of the target vehicle is improved, and the stationarity of the target vehicle and the riding experience of a user are also improved.

Embodiments of the present application will be described below in detail with reference to the accompanying drawings.

Example one

Fig. 3 is a flowchart illustrating a travelable region determining method according to a first embodiment of the present application, and as shown in fig. 3, an executing subject of the embodiment of the present application is a travelable region determining apparatus, which may be integrated in an electronic device. The travelable region determination method provided by the present embodiment includes the following steps.

Step 101, acquiring current frame environmental data of a target vehicle, which is acquired by a sensing system.

In this embodiment, the sensing system may periodically collect the current frame environmental data of the target vehicle according to the collection frequency. High-precision map information can be collected when the current frame environmental data of the target vehicle is collected. The current frame environment data may include: whether an obstacle exists around the target vehicle; information on the size, advancing speed, position and orientation of the obstacle; information such as the position, boundary line, and orientation of the current driving lane of the target vehicle; the information such as the sign, position, boundary line and orientation of the adjacent lane of the current driving lane, and the information such as the position, orientation and speed of the target vehicle.

In this embodiment, the electronic device obtains the current frame environmental data of the target vehicle acquired by the sensing system by communicating with the sensing system. The electronic device communicates with the sensing System of the target vehicle, and the communication mode may be Global System for Mobile communication (GSM), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Time Division-Synchronous Code Division multiple Access (TD-SCDMA), Long Term Evolution (Long Term Evolution, LTE), or future 5G. It can be understood that the communication mode of the electronic device and the perception system may also be a wireless communication mode, and the wireless communication mode may be zigbee communication, bluetooth BLE communication, or wifi communication of an action hotspot.

And step 102, if the adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extracting the current frame environment characteristic data on the adjacent lane.

In this embodiment, after acquiring the current frame environment data, the electronic device determines that an adjacent lane exists on the current driving lane of the target vehicle according to information such as the identification, the position, the boundary line, and the orientation of the adjacent lane of the current driving lane in the current frame environment data.

It is to be understood that, as an alternative embodiment, if the fields of the information indicating the identification, position, boundary line, and orientation of the adjacent lane of the current driving lane are all 0, it is determined that there is no adjacent lane on the current driving lane of the target vehicle. Otherwise, determining that an adjacent lane exists on the current driving lane of the target vehicle. And can determine the adjacent lane as a left lane and/or a right lane according to the mark of the adjacent lane of the current driving lane.

In this embodiment, if it is determined that an adjacent lane exists on the current driving lane of the target vehicle, the current frame environmental feature data on the adjacent lane is extracted according to the current frame environmental data.

Wherein, the current frame environment feature data may include: the direction deviation of the obstacle in the current frame from the current driving lane, the width of the obstacle in the current frame, the longitudinal speed of the obstacle in the current frame, the longitudinal relative speed of the obstacle in the current frame and the target vehicle, the longitudinal distance between the obstacle in the current frame and the target vehicle, and the transverse distance between the obstacle in the current frame and the target vehicle.

It should be noted that the obstacle of the current frame environmental characteristic data is an obstacle on an adjacent lane.

It is understood that the current frame environment feature data may also include other feature data, which is not limited in this embodiment.

And 103, predicting the current frame danger of the adjacent lane according to the current frame environment characteristic data to obtain a current frame initial danger value.

In this embodiment, the current frame environmental characteristic data may be input into the risk prediction model, and the risk prediction model predicts the current frame risk of the adjacent lane according to the current frame environmental characteristic data. And outputting a prediction result by the danger prediction model, wherein the prediction result is the current frame initial danger value of the adjacent lane.

It is noted that the risk prediction model may be a machine learning model or a deep learning model, etc. If the model is a machine learning model, the model may be a statistical prediction model, such as a generalized linear logistic regression model or other statistical prediction models. If the model is a deep learning model, the model may be a neural network model, such as a convolutional neural network model, a cyclic neural network model, or the like.

And 104, calculating the final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame.

In this embodiment, when the final risk value of each frame is calculated, the final risk value of each frame is stored. And after the initial danger value of the current frame is obtained, obtaining the final danger value of the previous frame corresponding to the adjacent road, calculating the danger reference value of the previous frame according to the final danger value of the previous frame, and calculating the final danger value of the current frame according to the initial danger value of the current frame and the corresponding danger reference value of the previous frame.

And 105, calculating the current frame drivable area of the adjacent lane according to the current frame final danger value.

In this embodiment, a mapping relationship between the final risk value of the current frame and the travelable region of the current frame may be stored in advance, and a value of the travelable region of the current frame having a mapping relationship with the final risk value of the current frame may be determined according to the mapping relationship and the calculated final risk value of the current frame.

The mapping relationship between the final risk value of the current frame and the travelable region of the current frame may be stored in the form of a mapping table, or stored in the form of a preset function, which is not limited in this embodiment.

In the method for determining a drivable area provided by the embodiment, current frame environmental data of a target vehicle, which is acquired by a sensing system, is acquired; if the adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extracting current frame environment characteristic data on the adjacent lane; predicting the current frame danger of the adjacent lane according to the current frame environment characteristic data to obtain a current frame initial danger value; calculating a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame; and calculating the current frame drivable area of the adjacent lane according to the final danger value of the current frame. After the current frame danger of the adjacent lane is predicted according to the current frame environmental characteristic data to obtain the current frame initial danger value, the current frame initial danger value and the corresponding previous frame danger reference value are fused to calculate the current frame final danger value, and the current frame travelable area of the adjacent lane is calculated according to the current frame final danger value. Therefore, each frame of final danger value is fused with the previous frame of danger reference value, so that the final danger value between two adjacent frames does not jump but is changed gradually, and further, each frame of drivable area of the determined adjacent lanes is also changed gradually rather than being changed suddenly to a large extent, so that the target vehicle does not rush to the direction, and the stability of the target vehicle and the riding experience of a user are improved.

Example two

Fig. 4 is a schematic flowchart of a method for determining a travelable region according to a second embodiment of the present application, and as shown in fig. 4, the method for determining a travelable region according to the present embodiment is further detailed in steps 103 to 105 on the basis of the method for determining a travelable region according to the first embodiment of the present application. And further comprising the step of training the risk prediction model. The travelable region determination method provided by the present embodiment includes the following steps.

Step 201, determining a risk prediction model trained to converge.

As an alternative embodiment, as shown in fig. 5, step 201 includes the following steps:

in step 2011, training samples corresponding to the risk prediction model are obtained, and the training samples are each frame of historical environmental characteristic data for risk labeling on the adjacent lane.

Optionally, in this embodiment, the risk prediction model is a generalized linear logistic regression model. Because the generalized linear logistic regression model does not forcibly change the natural measurement of the data, the predicted initial risk value of the current frame can be kept stable under the condition that the training sample has little variation, and the jump is not easy to occur.

Wherein, each frame of historical environmental characteristic data is each frame of environmental characteristic data which has occurred. Each frame of historical environmental characteristic data comprises any one or more of the following:

the direction deviation of the obstacle in the history frame from the current driving lane, the width of the obstacle in the history frame, the longitudinal speed of the obstacle in the history frame, the longitudinal relative speed of the obstacle in the history frame to the target vehicle, the longitudinal distance of the obstacle in the history frame to the target vehicle, and the lateral distance of the obstacle in the history frame to the target vehicle.

In this embodiment, each frame of historical environmental characteristic data is obtained by the electronic device acquiring each frame of environmental data of the target vehicle acquired by the sensing system during the driving process of the target vehicle, and extracting each frame of environmental data. After each frame of historical environment characteristic data is obtained, the dangers corresponding to each frame of historical environment characteristic data can be labeled. If a frame of historical environmental feature data is dangerous historical environmental feature data, the danger of the frame of historical environmental feature data can be marked as 1. If the historical environmental feature data of a certain frame is safe historical environmental feature data, the risk of the historical environmental feature data of the certain frame can be marked as 0.

Step 2012, training the initial risk prediction model using the training samples.

Step 2013, if it is determined that the trained risk prediction model meets the preset model convergence condition, determining the risk prediction model meeting the preset model convergence condition as the risk prediction model trained to converge.

Optionally, in this embodiment, the initial risk prediction model is an initial generalized linear logistic regression model. And inputting each training sample into the initial generalized linear logistic regression model, and predicting the current frame risk of the adjacent lane by the initial generalized linear logistic regression model according to each training sample. And changing parameters in the initial generalized linear logistic regression model to train the initial generalized linear logistic regression model. And judging whether the trained generalized linear logistic regression model meets the preset model convergence condition or not according to the preset model convergence condition in the training process. And if the preset model convergence condition is not met, continuing training the generalized linear logistic regression model. And if the preset model convergence condition is met, determining the danger prediction model meeting the preset model convergence condition as a danger prediction model trained to be converged.

The preset model convergence condition may be that the iteration number reaches a preset iteration number, or that the accuracy of the prediction result reaches a preset accuracy, or other preset model convergence conditions, which is not limited in this embodiment.

It should be noted that, after the risk prediction model trained to converge is determined, step 201 does not need to be executed each time the travelable region determining method according to the embodiment of the present application is executed. When the target vehicle encounters an unseen scene in the driving process to cause poor performance, in order to improve the accuracy of determining a drivable area, the environmental characteristic data corresponding to the new scene can be added into a training set to be used as a training sample to train the danger prediction model again so as to determine the danger prediction model trained to be convergent.

In the embodiment, each frame of environmental characteristic data on the adjacent roads is extracted as the training sample through the actual environment of the target vehicle in the driving process, so that the training sample data are comprehensive enough, and the accuracy of the prediction of the initial risk value of the current frame is improved when the prediction is carried out by adopting a risk prediction model from training to convergence.

Step 202, acquiring current frame environmental data of the target vehicle, which is acquired by a perception system.

In this embodiment, the implementation manner of step 202 is similar to that of step 101 in the first embodiment of the present application, and is not described herein again.

And step 203, if the adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extracting the current frame environment characteristic data on the adjacent lane.

Optionally, in this embodiment, the current frame environment feature data includes any one or more of the following:

It is understood that the current frame environmental feature data includes a feature data type consistent with each frame of historical environmental feature data.

Wherein the adjacent lane may be a left lane and/or a right lane of a current driving lane of the target vehicle.

For example, as shown in fig. 1, since the current driving lane of the target vehicle lane is the center lane, information such as the identification, position, boundary line, and orientation of the left lane, and information such as the identification, position, boundary line, and orientation of the right lane are included in the current frame environment data. The adjacent lanes of the target traveling vehicle are the left lane and the right lane. As shown in fig. 2, since the current driving lane of the target vehicle lane is the rightmost lane, the current frame environment data includes information such as the identification, position, and orientation of the left lane. The adjacent lane of the target traveling vehicle is the left lane.

And 204, inputting the environmental characteristic data of the current frame into the danger prediction model trained to be converged, and predicting the danger of the current frame by adopting the danger prediction model trained to be converged.

And step 205, outputting the initial risk value of the current frame through a risk prediction model trained to be convergent.

It should be noted that steps 204-205 are an alternative implementation of step 103 in the first embodiment of the present application.

Optionally, in this embodiment, the risk prediction model trained to converge is a generalized linear logistic regression model trained to converge.

Specifically, in this embodiment, the current frame environmental feature data is input into a generalized linear logistic regression model trained to converge, the generalized linear logistic regression model trained to converge predicts the current frame risk of the adjacent lane according to the current frame environmental feature data, and a prediction result is output, where the prediction result is the current frame initial risk value.

The initial danger value of the current frame is a numerical value between 0 and 1, and the larger the numerical value is, the more dangerous the current adjacent lane is. Conversely, a smaller value indicates that the current adjacent lane is safer.

In this embodiment, the risk prediction model trained to converge is used to predict the risk of the adjacent lane according to the environmental feature data, so as to obtain the current frame initial risk value of the adjacent lane. As long as the data of the training samples for training the risk prediction model are comprehensive enough, the prediction effect of the risk prediction model is ensured. If the target vehicle encounters a scene which is not seen in the driving process and causes poor performance, the accuracy of the prediction of the initial danger value of the current frame can be improved as long as the environmental characteristic data corresponding to the new scene is added into the training set to train the danger prediction model again, so that the accuracy of the determination of the drivable area is improved, and the maintenance cost can be controlled.

And step 206, calculating a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame.

Wherein the previous frame hazard reference value comprises: the final risk value of the previous frame and the decision value of the previous frame. As an alternative embodiment, as shown in fig. 6, step 206 includes the following steps:

step 2061, determining the decision value of the previous frame according to the final danger value of the previous frame.

As an alternative implementation, in this embodiment, step 2061 includes the following steps:

in step 2061a, the final risk value of the previous frame is obtained.

In step 2061b, if the final risk value of the previous frame is greater than the predetermined decision threshold, the decision value of the previous frame is determined to be the first value.

In step 2061c, if the final risk value of the previous frame is less than or equal to the predetermined decision threshold, the decision value of the previous frame is determined as the second value.

Specifically, in this embodiment, the final risk value of the previous frame is a value between 0 and 1. And after the final danger value of the previous frame is obtained, comparing the final danger value of the previous frame with a preset decision threshold, if the final danger value of the previous frame is greater than the preset decision threshold, indicating that the decision of the previous frame is a decision for contracting the drivable area of the previous frame, and determining that the decision value of the previous frame is a first numerical value which is 1. And if the final risk value of the previous frame is less than or equal to the preset decision threshold, the decision of the previous frame is a decision for expanding the travelable area of the previous frame, and the decision value of the previous frame is determined to be a second numerical value which is 0.

Step 2062, determining weights corresponding to the initial risk value of the current frame, the final risk value of the previous frame and the decision value of the previous frame.

The weights corresponding to the current frame initial risk value, the previous frame final risk value and the previous frame decision value can be respectively represented as k1, k2 and k 3.

In order to ensure that decision switching occurs on the same obstacle in the same adjacent lane, the decision except the final risk value of the first frame is not completely failed, and the weight k3 corresponding to the decision value of the previous frame is smaller than the weight k1 corresponding to the initial risk value of the current frame.

It is understood that the numerical values of k1, k2 and k3 can be determined according to the accuracy of determining the travelable region per frame and the performance of the target vehicle during traveling.

Step 2063, performing weighted summation calculation on the initial risk value of the current frame, the final risk value of the previous frame and the decision value of the previous frame to obtain the final risk value of the current frame.

Specifically, in this embodiment, the current frame initial risk value, the previous frame final risk value, and the previous frame decision value are weighted and summed to obtain the current frame final risk value, which can be expressed as formula (1):

L_{finally, the product is processed}＝k1*L_Initial+k2*N_{Finally, the product is processed}+k3*N_{Decision making}(1)

Wherein L is_{Finally, the product is processed}Is the final risk value, L, of the current frame_InitialFor the initial risk value of the current frame, N_{Finally, the product is processed}For the final risk value of the previous frame, N_{Decision making}The value is decided for the previous frame.

In particular, k2 × N_{Finally, the product is processed}The effect of this term is also to make the decision making switch more robust, but the value of k2 is appropriate. The larger the value of k2, the less easy the decision to switch, resulting in sluggish changes between drivable regions per frame. k 3N_{Decision making}The function of the term will pull the final risk value of the current frame in the same direction as the final risk value of the previous frame. This part has a threshold-like effect so that the final risk value of the current frame does not jump if the confidence of the initial risk value of the current frame is not high enough. And when the initial risk value of the current frame is slightly deviated to the direction different from the final risk value of the previous frame, the final risk value of the current frame is not jumped due to the action of the item, so that the anti-noise performance is strong, and the final risk value of each frame is prevented from jumping.

In this embodiment, the dangerous reference value of the previous frame includes: and when the final risk value of the current frame is calculated, the product of the final risk value of the previous frame and the weight enables the decision switching to be more stable, and the product of the final risk value of the previous frame and the weight pulls the final risk value of the current frame to the same direction as the final risk value of the previous frame, so that the final risk value of each frame is not easy to jump but is a gradually changing process, the target vehicle cannot rush to the direction, the anti-noise performance of the target vehicle is further improved, and the stability of the target vehicle and the riding experience of a user are further improved.

And step 207, acquiring a preset continuous function, wherein the preset continuous function represents the mapping relation between the final risk value of the current frame and the travelable area of the current frame.

Specifically, in this embodiment, a continuous function representing a mapping relationship between the final risk value of the current frame and the travelable region of the current frame is preset, and the continuous function may be a linear function, an exponential function, or the like. The independent variable may be a final risk value of the current frame, and the dependent variable may be a size of a travelable region of the current frame, or a current lateral distance between the travelable region of the current frame and the target vehicle, a lateral distance between the travelable region of the current frame and a nearest boundary of a current travel lane of the target vehicle, and the like.

And step 208, inputting the final danger value of the current frame into a preset continuous function to calculate a travelable area of the current frame.

It should be noted that steps 207-208 are an alternative implementation of step 105 in the first embodiment of the present application.

Specifically, in this embodiment, the final risk value of the current frame is input as an independent variable into a preset continuous function, and the preset continuous function calculates a value of a corresponding dependent variable, where the value of the dependent variable is the calculated travelable area of the current frame.

It will be appreciated that the predetermined continuous function is a monotonically decreasing function. Namely, the larger the final risk value of the current frame is, the smaller the travelable area of the current frame is. Otherwise, the smaller the final risk value of the current frame is, the larger the drivable area of the current frame is.

Step 209, the current frame travelable area is sent to the decision system of the target vehicle, so that the decision system plans the travelable path according to the current frame travelable area and the current frame travelable area on the current travel lane of the target vehicle.

Further, in this embodiment, the determined current frame travelable area of the adjacent lane is sent to the decision system, and the decision system plans the travelable path of the target vehicle according to the current frame travelable area of the adjacent lane and the current frame travelable area on the current traveling lane of the target vehicle, so that the target vehicle travels according to the travelable path.

As in fig. 1, since there is no obstacle ahead on the current driving lane of the target vehicle, although the left lane has a larger current-frame travelable area, the travelable path planned by the target vehicle does not pass through the left lane but passes on the current driving lane of the target vehicle.

As another example, in fig. 2, the current travelable area of the left lane is 2 meters, and an obstacle exists 1m ahead of the current road on which the target vehicle is currently traveling, the travelable path planned by the decision system passes through the current travelable area of the left lane to bypass the obstacle existing 1m ahead of the current road on which the target vehicle is currently traveling.

According to the method for determining the drivable area, the mapping relation between the final risk value of the current frame and the drivable area of the current frame is set as the preset continuous function, so that the drivable area of the current frame is gradually changed in the process that the final risk value of each frame is gradually changed, and the stability of the target vehicle and the riding experience of a user are further improved.

EXAMPLE III

Fig. 7 is a signaling flowchart of a travelable region determining method according to a third embodiment of the present application, and as shown in fig. 7, the travelable region determining method according to the present embodiment includes the following steps:

step 301, a sensing system collects current frame environmental data of a target vehicle.

Step 302, the perception system sends the current frame environmental data of the target vehicle to the electronic device.

As an optional implementation manner, in this embodiment, the electronic device may be an in-vehicle device, and is also mounted on the target vehicle, and the sensing system transmits the current frame environment data of the target vehicle to the electronic device.

Step 303, if the electronic device determines that an adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extracting current frame environment feature data on the adjacent lane.

And step 304, the electronic equipment predicts the current frame danger of the adjacent lane according to the current frame environment characteristic data to obtain a current frame initial danger value.

In step 305, the electronic device calculates a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame.

And step 306, the electronic equipment stores the final risk value of the current frame.

Further, in this embodiment, the electronic device stores the final risk value of the current frame, so that the final risk value of the current frame is obtained from the storage area when the final risk value of the next frame is calculated, and the decision value of the current frame is calculated according to the final risk value of the current frame. The calculation of the current frame travelable area is a continuous cycle process in the traveling process of the target vehicle.

Optionally, in this embodiment, the current frame final risk values of the left lane and the right lane may be stored separately, so as to obtain the current frame final risk values on the corresponding adjacent lanes more quickly.

And 307, the electronic equipment calculates the current frame drivable area of the adjacent lane according to the current frame final danger value.

And 308, the electronic equipment sends the current frame driving-capable area of the adjacent lane to the decision-making system.

In this embodiment, the implementation manners of steps 303 to 305 and steps 307 to 308 are similar to the implementation manners of the related steps in the second embodiment of the present application, and are not described in detail here.

Step 309, the decision system plans the drivable path of the target vehicle according to the current frame drivable area of the adjacent lane and the current frame drivable area on the current driving lane of the target vehicle, so that the target vehicle drives according to the drivable path.

It should be noted that the specific manner in which the decision system plans the drivable path of the target vehicle according to the current frame drivable area of the adjacent lane and the current frame drivable area on the current driving lane of the target vehicle is not limited in this embodiment.

As an optional implementation manner, when the decision system determines the current frame drivable area on the current driving lane, the decision system may acquire current frame environmental data acquired by the sensing system, extract current frame environmental characteristic data on the current driving lane of the target vehicle, and determine the current frame drivable area on the current driving lane according to the current frame environmental characteristic data.

Example four

Fig. 8 is a schematic structural diagram of a travelable region determining apparatus according to a fourth embodiment of the present application, and as shown in fig. 8, the travelable region determining apparatus according to this embodiment is located on an electronic device, a sensing system is mounted on a target vehicle, and the electronic device is in communication with the sensing system, and the travelable region determining apparatus 800 according to this embodiment includes: an environmental data acquisition module 801, a feature data extraction module 802, an initial risk value prediction module 803, a final risk value calculation module 804 and an area calculation module 805.

The environment data acquiring module 801 is configured to acquire current frame environment data of the target vehicle acquired by the sensing system. The feature data extraction module 802 is configured to, if it is determined that an adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extract current frame environment feature data on the adjacent lane. The initial risk value prediction module 803 is configured to predict the risk of the current frame of the adjacent lane according to the current frame environment characteristic data, so as to obtain an initial risk value of the current frame. And a final risk value calculation module 804, configured to calculate a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame. The region calculating module 805 is configured to calculate a current frame drivable region of the adjacent lane according to the current frame final risk value.

The travelable area determining apparatus provided in this embodiment may implement the technical solution of the method embodiment shown in fig. 3, and the implementation principle and technical effect of the travelable area determining apparatus are similar to those of the method embodiment shown in fig. 3, and are not described in detail here.

Fig. 9 is a schematic structural diagram of a travelable region determining device according to a fifth embodiment of the present application, and as shown in fig. 9, the travelable region determining device according to the present embodiment further includes, in addition to the travelable region determining device according to the fourth embodiment: a model training module 901.

In the travelable region determining apparatus provided in this embodiment, in the feature data extraction module 802, the current frame environment feature data includes any one or more of the following:

Further, the initial risk value prediction module 803 is specifically configured to:

inputting the current frame environmental characteristic data into a risk prediction model trained to be convergent, and predicting the risk of the current frame by adopting the risk prediction model trained to be convergent; and outputting the initial risk value of the current frame through a risk prediction model trained to be converged.

Optionally, the risk prediction model is a generalized linear logistic regression model.

Further, the model training module 901 is configured to:

acquiring training samples corresponding to the danger prediction model, wherein the training samples are each frame of historical environment characteristic data for carrying out danger labeling on adjacent lanes; training the initial risk prediction model by adopting a training sample; and if the trained danger prediction model meets the preset model convergence condition, determining the danger prediction model meeting the preset model convergence condition as the danger prediction model from training to convergence.

Further, the previous frame hazard reference value includes: the final risk value of the previous frame and the decision value of the previous frame.

Accordingly, the final risk value calculation module 804 is specifically configured to:

determining a decision value of a previous frame according to the final danger value of the previous frame; determining weights corresponding to the initial risk value of the current frame, the final risk value of the previous frame and the decision value of the previous frame; and carrying out weighted summation calculation on the initial risk value of the current frame, the final risk value of the previous frame and the decision value of the previous frame to obtain the final risk value of the current frame.

Further, when determining the decision value of the previous frame according to the final risk value of the previous frame, the final risk value calculation module 804 is specifically configured to:

acquiring a final danger value of a previous frame; if the final risk value of the previous frame is larger than a preset decision threshold, determining the decision value of the previous frame as a first numerical value; and if the final risk value of the previous frame is less than or equal to the preset decision threshold, determining the decision value of the previous frame as a second numerical value.

Further, the weight corresponding to the decision value of the previous frame is smaller than the weight corresponding to the initial risk value of the current frame.

Further, the region calculating module 805 is specifically configured to:

acquiring a preset continuous function, wherein the preset continuous function represents the mapping relation between the final risk value of the current frame and the travelable area of the current frame; and inputting the final danger value of the current frame into a preset continuous function to calculate a travelable region of the current frame.

The travelable area determining apparatus provided in this embodiment may implement the technical solutions of the method embodiments shown in fig. 3 to 7, and the implementation principles and technical effects thereof are similar to those of the method embodiments shown in fig. 3 to 7, and are not described in detail here.

According to an embodiment of the present application, an electronic device and a readable storage medium are also provided.

As shown in fig. 10, the present invention is a block diagram of an electronic device according to a travelable region determining method according to an embodiment of the present application. Electronic devices are intended for various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.

As shown in fig. 10, the electronic apparatus includes: one or more processors 1001, memory 1002, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions for execution within the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to the interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple electronic devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). Fig. 10 illustrates an example of one processor 1001.

The memory 1002 is a non-transitory computer readable storage medium provided herein. The memory stores instructions executable by the at least one processor to cause the at least one processor to perform the travelable region determination methods provided herein. A non-transitory computer-readable storage medium of the present application stores computer instructions for causing a computer to execute the travelable region determination method provided by the present application.

The memory 1002, which is a non-transitory computer-readable storage medium, may be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as program instructions/modules corresponding to the travelable region determination method in the embodiment of the present application (for example, the environmental data acquisition module 801, the feature data extraction module 802, the initial risk value prediction module 803, the final risk value calculation module 804, and the region calculation module 805 shown in fig. 8). The processor 1001 executes various functional applications of the server and data processing, i.e., implements the travelable region determination method in the above-described method embodiment, by running non-transitory software programs, instructions, and modules stored in the memory 1002.

The memory 1002 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the electronic device of fig. 10, and the like. Further, the memory 1002 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 1002 may optionally include memory located remotely from the processor 1001, which may be connected to the electronic device of FIG. 10 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The electronic device of fig. 10 may further include: an input device 1003 and an output device 1004. The processor 1001, the memory 1002, the input device 1003, and the output device 1004 may be connected by a bus or other means, and the bus connection is exemplified in fig. 10.

The input device 1003 may receive input voice, numeric, or character information and generate key signal inputs related to user settings and function controls of the electronic apparatus of fig. 10, such as an input device like a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointer stick, one or more mouse buttons, a track ball, a joystick, etc. The output devices 1004 may include a voice playing device, a display device, auxiliary lighting devices (e.g., LEDs), and tactile feedback devices (e.g., vibrating motors), among others. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device can be a touch screen.

Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

These computer programs (also known as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.

The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

According to the technical scheme of the embodiment of the application, after the current frame danger of the adjacent lane is predicted according to the current frame environment characteristic data to obtain the current frame initial danger value, the current frame initial danger value and the corresponding previous frame danger reference value are fused to calculate the current frame final danger value, and the current frame drivable area of the adjacent lane is calculated according to the current frame final danger value. Therefore, the final danger value of each frame is fused with the danger reference value of the previous frame, so that the final danger value between two adjacent frames does not jump but is changed gradually, and further, the determined each frame drivable area of the adjacent lanes is also changed gradually rather than being changed suddenly by a large amplitude. And then the target vehicle can not be driven to make a violent movement, and the stability of the target vehicle and the riding experience of a user are improved.

It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present application may be executed in parallel, sequentially, or in different orders, and the present invention is not limited thereto as long as the desired results of the technical solutions disclosed in the present application can be achieved.

The above-described embodiments should not be construed as limiting the scope of the present application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A travelable region determination method applied to an electronic device with a sensing system mounted on a target vehicle, the electronic device being in communication with the sensing system, the method comprising:

acquiring current frame environmental data of the target vehicle acquired by the perception system;

if the adjacent lane exists on the current driving lane of the target vehicle according to the current frame environment data, extracting current frame environment characteristic data on the adjacent lane;

predicting the current frame danger of the adjacent lane according to the current frame environment characteristic data to obtain a current frame initial danger value;

calculating a final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame;

and calculating the current frame drivable area of the adjacent lane according to the current frame final danger value.

2. The method of claim 1, wherein the predicting the current frame risk of the adjacent lane according to the current frame environment feature data to obtain a current frame initial risk value comprises:

inputting the current frame environmental characteristic data into a danger prediction model trained to be convergent so as to predict the danger of the current frame by adopting the danger prediction model trained to be convergent;

and outputting the current frame initial risk value through the risk prediction model trained to be convergent.

3. The method of claim 2, wherein the risk prediction model is a generalized linear logistic regression model.

4. The method according to claim 2 or 3, wherein before inputting the current frame environmental feature data into the risk prediction model trained to converge, further comprising:

acquiring training samples corresponding to a danger prediction model, wherein the training samples are each frame of historical environment characteristic data for carrying out danger labeling on the adjacent lanes;

training an initial risk prediction model by adopting the training sample;

and if the trained danger prediction model meets the preset model convergence condition, determining the danger prediction model meeting the preset model convergence condition as the danger prediction model from training to convergence.

5. The method of claim 1, wherein the previous frame hazard reference value comprises: a final risk value of a previous frame and a decision value of the previous frame; the calculating the final risk value of the current frame according to the initial risk value of the current frame and the corresponding risk reference value of the previous frame comprises the following steps:

determining the decision value of the previous frame according to the final danger value of the previous frame;

determining weights corresponding to the initial risk value of the current frame, the final risk value of the previous frame and the decision value of the previous frame;

and performing weighted summation calculation on the current frame initial risk value, the previous frame final risk value and the previous frame decision value to obtain the current frame final risk value.

6. The method of claim 5, wherein determining the previous frame decision value based on the previous frame final risk value comprises:

acquiring a final danger value of the previous frame;

if the final risk value of the previous frame is greater than a preset decision threshold, determining that the decision value of the previous frame is a first numerical value;

and if the final risk value of the previous frame is less than or equal to the preset decision threshold, determining the decision value of the previous frame as a second numerical value.

7. The method according to claim 5 or 6, wherein the weight corresponding to the decision value of the previous frame is smaller than the weight corresponding to the initial risk value of the current frame.

8. The method according to any one of claims 1-3, wherein said calculating a current frame travelable region of said adjacent lane according to said current frame final risk value comprises:

acquiring a preset continuous function, wherein the preset continuous function represents the mapping relation between the final danger value of the current frame and the drivable area of the current frame;

and inputting the final danger value of the current frame into the preset continuous function to calculate the travelable region of the current frame.

9. The method according to any one of claims 1 to 3, wherein the current frame environment feature data comprises any one or more of the following:

10. A travelable region determining apparatus, characterized in that the apparatus is located in an electronic device, a sensing system is mounted on a target vehicle, and the electronic device communicates with the sensing system, the apparatus comprising:

11. The apparatus of claim 10, wherein the initial risk value prediction module is specifically configured to:

12. The apparatus of claim 11, wherein the risk prediction model is a generalized linear logistic regression model.

13. The apparatus of claim 11 or 12, further comprising: a model training module to:

14. The apparatus of claim 10, wherein the previous frame hazard reference value comprises: a final risk value of a previous frame and a decision value of the previous frame;

the final risk value calculation module is specifically configured to:

15. The apparatus according to claim 14, wherein the final risk value calculating module, when determining the previous frame decision value according to the previous frame final risk value, is specifically configured to:

16. The apparatus according to claim 14 or 15, wherein the weight corresponding to the decision value of the previous frame is smaller than the weight corresponding to the initial risk value of the current frame.

17. The apparatus according to any one of claims 10 to 12, wherein the region calculation module is specifically configured to:

18. The apparatus according to any one of claims 10-12, wherein the current frame environment feature data comprises any one or more of:

19. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.

20. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-9.