CN111307170A - Unmanned vehicle driving planning method, device, equipment and medium - Google Patents

Abstract

The invention discloses a driving planning method, device, equipment and medium of an unmanned vehicle. The method includes: receiving an intelligent driving instruction sent by a target vehicle, and obtaining positioning information of the target vehicle; making a binocular camera of the target vehicle collect synchronous image data within a preset monitoring range; determining the dynamic status of the target vehicle according to the positioning information of the target vehicle Receiving range, and obtain the shared image data within the dynamic receiving range from the cloud database; obtain the synchronous monitoring data generated by decompressing and inversely transforming the shared image data; according to the synchronous image data and the synchronous monitoring data, drive the target vehicle planning. The invention realizes the rapid transmission and sharing of data in the area, which is conducive to timely early warning and judgment, and timely and effectively handles emergencies around the vehicle, thereby improving the safety of the unmanned vehicle.

Description

Unmanned vehicle driving planning method, device, equipment and medium

technical field

The invention relates to the field of artificial intelligence, in particular to a driving planning method, device, equipment and medium for an unmanned vehicle.

Background technique

For the Advanced Driver Assistance Systems (ADAS) of driverless vehicles, sensors are a key component of driverless vehicles, which can provide the central controller with vehicle image data in front of the vehicle, rear and side of the vehicle. The sources of vehicle data can include: radar, ultrasonic waves, lasers, cameras, etc. The data obtained by the camera can observe changes in the dynamic driving process in real time, and because of its low cost, large amount of information collection, and relatively high resolution. It is very popular for its high standard.

At present, the image data obtained by the unmanned vehicle driving on the road with the help of the advanced driving assistance system only comes from a single vehicle body, and the handling of emergencies is relatively simple, such as between vehicles or between vehicles and moving obstacles due to different The effective visual information of both parties can be obtained in time to cause traffic accidents; at the same time, the amount of image data and information obtained within the scope of the single-view lens between vehicles is relatively small, the judgment of the road is single, and it is impossible to communicate in time to make an early warning Therefore, it needs to be optimized continuously.

SUMMARY OF THE INVENTION

Embodiments of the present invention provide a driving planning method, device, device and medium for an unmanned vehicle, which can efficiently transmit synchronous image data and share synchronous monitoring data with other vehicles during the unmanned driving of a target vehicle, so that according to the synchronous image data The target vehicle can be controlled by synchronizing health data to achieve unmanned driving, and at the same time, emergencies can be handled in time according to it, which improves the safety of unmanned vehicles.

A driving planning method for an unmanned vehicle, comprising:

Receive the intelligent driving instruction sent by the target vehicle, and obtain the positioning information of the target vehicle;

Acquiring synchronous image data within a preset monitoring range collected by the binocular camera of the target vehicle;

The dynamic receiving range of the target vehicle is determined according to the positioning information of the target vehicle, and the shared image data within the dynamic receiving range is obtained from the cloud database; the shared image data refers to the synchronous transmission of all vehicles in the current period to image data in the cloud database;

obtaining synchronous monitoring data generated after decompressing and inverse transforming the shared image data;

Carrying out vehicle travel planning for the target vehicle according to the synchronized image data and the synchronized monitoring data.

An unmanned vehicle driving planning device, comprising:

a positioning module, configured to receive the intelligent driving instruction sent by the target vehicle, and obtain the positioning information of the target vehicle;

an acquisition module, configured to acquire synchronous image data within a preset monitoring range acquired by the binocular camera of the target vehicle;

a receiving module, configured to determine the dynamic receiving range of the target vehicle according to the positioning information of the target vehicle, and obtain the shared image data within the dynamic receiving range from the cloud database; the shared image data refers to the current time period All vehicles are synchronously transmitted to the image data in the cloud database;

a decompression module, configured to obtain synchronous monitoring data generated after decompressing and inverse transforming the shared image data;

A planning module, configured to perform vehicle travel planning on the target vehicle according to the synchronous image data and the synchronous monitoring data.

A computer device comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, the processor implementing the above-mentioned unmanned vehicle when executing the computer-readable instructions Driving planning method.

A computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, implements the above-mentioned driving planning method for an unmanned vehicle.

The driving planning method, device, device and medium for an unmanned vehicle provided by the present invention, by acquiring the synchronous image data collected by the binocular camera of the target vehicle and receiving the synchronous monitoring data within the dynamic receiving range of the target vehicle, combined with the target vehicle itself The synchronous image data of the target vehicle and the synchronous monitoring data within the dynamic acceptance range of the target vehicle are used for vehicle driving planning, and the amount of obtained image data and information is more abundant; the present invention can efficiently transmit the synchronous image data during the unmanned driving of the target vehicle. And share the synchronous monitoring data with other vehicles, so as to control the target vehicle to realize unmanned driving according to the synchronous image data and the synchronous health data. At the same time, the present invention can also timely make early warning and judgment according to the above data, and timely and effectively deal with the sudden changes around the vehicle. incidents, improving the safety of driverless vehicles.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the following briefly introduces the drawings that are used in the description of the embodiments of the present invention. Obviously, the drawings in the following description are only some embodiments of the present invention. , for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative effort.

1 is a schematic diagram of an application environment of a driving planning method for an unmanned vehicle according to an embodiment of the present invention;

2 is a flowchart of a driving planning method for an unmanned vehicle in an embodiment of the present invention;

3 is a flowchart of step S20 of a driving planning method for an unmanned vehicle in an embodiment of the present invention;

4 is a flowchart of step S203 of the driving planning method for an unmanned vehicle in an embodiment of the present invention;

5 is a flowchart of a compressed forward transformation of a driving planning method for an unmanned vehicle according to an embodiment of the present invention;

6 is a flowchart of step S40 of the driving planning method for an unmanned vehicle according to an embodiment of the present invention;

7 is a flowchart of decompression and inverse transformation of a driving planning method for an unmanned vehicle according to an embodiment of the present invention;

8 is a schematic block diagram of a driving planning device for an unmanned vehicle according to an embodiment of the present invention;

9 is a schematic block diagram of a collection module of an unmanned vehicle driving planning device according to an embodiment of the present invention;

FIG. 10 is a schematic diagram of a computer device in an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The driving planning method for an unmanned vehicle provided by the present invention can be applied in the application environment as shown in FIG. 1 , wherein the unmanned vehicle communicates with a server through a network. The server can be implemented as an independent server or a server cluster composed of multiple servers.

In one embodiment, as shown in FIG. 2, a driving planning method for an unmanned vehicle is provided, and the method is applied to the server in FIG. 1 as an example for description, including the following steps:

S10. Receive an intelligent driving instruction sent by a target vehicle, and acquire positioning information of the target vehicle.

The intelligent driving instruction refers to the operating parameters and destination location information entered by the owner of the unmanned vehicle (that is, the target vehicle) on the control panel set by the unmanned vehicle, and triggering the pre-operation by clicking or sliding. The designated automatic driving mode button or the manual driving mode button sends the command to the server.

Preferably, when receiving the intelligent driving instruction sent by the target vehicle, the positioning information of the target vehicle is obtained through the positioning system set by the unmanned vehicle. At this time, the positioning information and the destination position information can be stored in a preset electronic Mark on the map, and receive multiple initial travel routes planned by the electronic map, and then automatically select a better route among the multiple travel routes (for example, the parameters such as the fastest time, the lowest charge or the shortest distance are better) , make the unmanned vehicle run according to the preferred route with the operating parameters, and then adjust the preferred route according to the image data acquired in real time in the S40.

S20. Acquire synchronous image data within a preset monitoring range collected by the binocular camera of the target vehicle.

Wherein, the binocular cameras include, but are not limited to, binocular cameras arranged in the front and rear, left and right sides, upper left, lower right, upper right and lower left of the unmanned vehicle, and each of the binocular cameras is provided with Two single cameras, each of the binocular cameras collects one channel of image data within its monitoring range. That is, the synchronous image data refers to the front and rear, left and right sides of the unmanned vehicle that are collected in parallel and synchronously by the binocular cameras arranged at the front and rear, left and right sides, upper left and lower right, upper right and lower left of the unmanned vehicle. Image data for both sides, upper left and lower right, and upper right and lower left. The synchronized image data includes, but is not limited to, environmental image data around the vehicle, as well as people flow data, surrounding vehicle information, and the like extracted from the environmental image data around the vehicle.

Specifically, when the unmanned vehicle (that is, the target vehicle) is not started, the binocular camera is turned off or in an energy-saving state. When the unmanned vehicle is started, the binocular camera is awakened, that is, the binocular camera is automatically turned off from the energy-saving state or off. The state is switched to the working state, and the binocular camera is made to capture synchronous image data according to the preset image acquisition cycle (that is, to collect once every preset time period).

S30. Determine the dynamic receiving range of the target vehicle according to the positioning information of the target vehicle, and acquire shared image data within the dynamic receiving range from a cloud database; the shared image data refers to the synchronization of all vehicles in the current period Image data transferred to the cloud database.

Wherein, the dynamic receiving range is to follow the positioning information of the target vehicle to determine the receiving range of the synchronous monitoring data, and the dynamic receiving range is set according to requirements, such as: taking the target vehicle as the center, obtaining a circle within 2 kilometers The area range serves as the dynamic reception range.

The cloud database is used to store the synchronous image data collected by all vehicles with the same sharing authority through the binocular cameras, and provides a query function by region and time period. It is understandable that, in an embodiment, each car company of the driverless vehicle can create its own cloud database, and each car company needs to provide sharing permission for all the driverless vehicles under its name.

Further, after compressing the synchronous image data of the target vehicle, it can be transmitted to a cloud database that is communicatively connected to the server, and the compressed synchronous image data of the target vehicle (which is marked as shared image data after compression) After transmission to the cloud database, if the target vehicle is in the dynamic receiving range of other vehicles during the current period, the synchronized image data of the target vehicle transmitted to the cloud server (it has been compressed and marked as a shared image before transmission to the cloud server) data) can be used as the synchronous monitoring data of other vehicles (after the shared image data is retrieved from the cloud server, it is decompressed and inversely transformed to generate synchronous monitoring data); if the target vehicle moves from the first location to After the second location, the first location is within the dynamic acceptance range of the second location. At this time, after the synchronized image data captured at the first location is transmitted to the cloud database, the synchronized image of the target vehicle at the first location can be The data is used as the synchronous monitoring data of the second location, but the synchronous image data of the target vehicle in the first location can be obtained directly from the cloud database, or directly from the database of the target vehicle (the target vehicle is in the first location). The synchronized image data of the location is pre-stored in the database, and it is faster to extract directly from the database).

Preferably, when starting the unmanned vehicle, the time information of the target vehicle (the current time point and the current time period at which the current time point is located) is obtained, and a query instruction is generated according to the time information and the positioning information, and the The query command is sent to the cloud database, and the shared image data transmitted back by the cloud database is received. The time information is the current time point and the current time period corresponding to the current time point.

S40. Acquire synchronization monitoring data generated after decompressing and inverse transforming the shared image data.

In this embodiment, after the shared image data corresponding to the time information and the positioning information is automatically obtained from the cloud database according to the time information and the positioning information, the obtained shared image data is decompressed. transmitted to the target vehicle. It is understandable that the shared image data stored in the cloud database is the compressed synchronized image data. When the shared image data is transmitted to the vehicle, the shared image data needs to be decompressed to obtain the data available for the vehicle (ie, synchronized image data). monitoring data).

S50. Perform vehicle travel planning on the target vehicle according to the synchronous image data and the synchronous monitoring data.

Wherein, the vehicle driving planning includes route control and vehicle control; the route control includes changing lanes, route planning/adjustment, etc.; the vehicle control includes vehicle acceleration, vehicle acceleration, vehicle turning, vehicle U-turn, starting vehicle, stopping vehicles etc.

It is understandable that the shared image data of vehicles within a dynamic receiving range (a short-range range) has temporal and spatial correlations. Taking each vehicle itself as the central node, different vehicles within the dynamic receiving range are selected for continuous time. The image data can make more reasonable vehicle driving planning.

Exemplarily, after obtaining the image data of the road ahead through the binocular camera disposed directly in front of the target vehicle, the distance between the target vehicle and the vehicle in front is determined according to the image data of the front road, and a short-range deceleration warning is issued; further, The distance between the target vehicle and the preceding vehicle is obtained in real time, and when it is detected that the distance between the target vehicle and the preceding vehicle is less than a preset deceleration distance threshold, the target vehicle is automatically controlled to decelerate.

Exemplarily, in the case of low visibility in extreme weather (such as heavy fog, sand and dust, etc.), the traffic light signal is captured from the binocular camera set directly in front of the target vehicle and a voice prompt is given. Further, when the traffic light is obtained as a green light according to the traffic light signal, the maximum passing probability that the vehicle can pass the traffic light is obtained according to the obtained road image data, the current running parameters of the vehicle, the distance between the vehicle and the traffic light, and other information, When it is detected that the maximum passing probability is less than a preset green light deceleration threshold, the target vehicle is automatically controlled to decelerate.

To sum up, the driving planning method for an unmanned vehicle provided by the present invention simultaneously acquires the synchronous image data of the target vehicle itself and the shared image data around the target vehicle itself, and combines the synchronous image data of the target vehicle itself and the target vehicle itself. The surrounding shared image data is used for vehicle driving planning, and the amount of obtained image data and information is more abundant, which is conducive to handling emergencies, and can communicate with the server in time to make early warning judgments.

In one embodiment, as shown in FIG. 3 , the step S20 specifically includes the following steps:

S201. Order multiple sets of binocular cameras to synchronously collect the synchronous image data in multiple ways; the binocular cameras include binocular cameras disposed at the front and rear, left and right sides, upper left, lower right, and upper right and lower left of the target vehicle , and each of the binocular cameras includes two single cameras;

That is to say, the binocular cameras installed on the front and rear, left and right sides, upper left, lower right, and upper right and lower left of the target vehicle are all a set of binocular cameras, and multiple sets of binocular cameras are synchronized to collect multiple simultaneous images in parallel. data, which can ensure the real-time validity of image data.

S202. Save the synchronized image data collected by each of the binocular cameras into a two-dimensional point matrix; the two-dimensional point matrix refers to a two-dimensional matrix described by pixels.

That is, each single camera of the binocular camera set on the target vehicle can be used as a collection node to collect synchronous image data of each collection node respectively, and save the synchronous image data using a two-dimensional point matrix.

S203 , after performing compression forward transformation on the two-dimensional point matrix, obtain a two-dimensional component matrix; the two-dimensional component matrix refers to a two-dimensional matrix described by non-zero dimension components.

Wherein, the compression forward transformation is used to convert the synchronous image data collected by the binocular camera into shared image data, and includes row transformation, column transformation, high and low frequency component transformation and component replacement.

In this embodiment, after the synchronous image data is stored using a two-dimensional dot matrix in step S202, row transformation and column transformation are performed on the two-dimensional dot matrix, and four corresponding synchronous image data can be obtained. Dimension components, remove the highest component, and keep the remaining three dimension components as valid dimension components.

S204. Mark the two-dimensional component matrix as shared image data, and transmit the shared image data to the cloud database.

To sum up, the driving planning method for an unmanned vehicle provided by the present invention obtains the synchronous image data on all the collection nodes, and saves the synchronous image data corresponding to each collection node using a two-dimensional point matrix, so as to obtain the synchronous image data of each collection node. After the two-dimensional point matrix is compressed and transformed, shared image data is obtained, and the shared image data is transmitted to the cloud database, which can save transmission energy and improve transmission efficiency.

In an embodiment, as shown in FIG. 4 , the step S203 specifically includes the following steps;

S2031. Perform row transformation using the two-dimensional point matrix as the original signal, and obtain a row transformed signal including a first detail signal and a first approximation signal.

Exemplarily, the compressed forward transform is shown in FIG. 5 .

First, the row pixels in the original signal are split to obtain a first odd-numbered signal and a first even-numbered signal. That is, entering the splitting stage, the original signal f(t) is split into the first odd-numbered signal xo _m (t) and the first even-numbered signal xe _m (t) according to the sampling interval τ, as shown in the following formula (1):

Next, predict the first odd-numbered signal according to a preset first predictor to obtain a first detail signal. That is, enter the prediction stage, keep the first even signal unchanged, predict the first odd signal by the first prediction operator P _m [ ], and define the difference between the predicted value and the actual value as the first detail signal d _m (t). It is shown in the following formula (2):

d _m (t)=xo _m (t)-P _m [xe _m (t)] (2)

Finally, the first even-number information is updated according to the preset first update operator and the first detail signal to obtain a first approximation signal. That is, entering the update stage, firstly introduce the first update operator U _m [ ], and use the above-mentioned first detail signal to update the original first even signal, thereby obtaining a first approximation signal cm ( _t ), as follows Formula (3) shows:

c _m (t)=xe _m (t)+U _m [d _m (t)] (3)

Preferably, when performing line transformation on the original signal, the first detail signal is a high-frequency component; the first approximation signal is a low-frequency component, and the first predictor P _m [·] is xe _m (t), that is, the first even-numbered signal itself, and the first update operator U _m [·] is d _m (t), that is, the first detail signal itself.

S2032. Perform column transformation on the row-transformed signal to obtain a column-transformed signal including a second detail signal and a second approximation signal.

Specifically, split processing is performed on the column pixels in the row-transformed signal to obtain the second odd-numbered signal yon ( _t ) and the second even _- numbered signal yeon (t); using the preset second prediction operator P _n [·] Predict the second odd-numbered signal yon (t) to obtain the second detail signal d _n ( _t ), and use the preset second update operator U _n [·] and the second detail signal d _n ( t) Update the second even-numbered signal y _n (t) to obtain a second approximation signal c _n (t). Preferably, when performing column transformation on the row-transformed signal, the second detail signal is a high-frequency component, the second approximation signal is a low-frequency component, and the second predictor P _n [·] is y _n (t), that is, the second even signal itself, and the update operator U _n [·] is d _n (t), that is, the second detail signal itself.

其中y₁(t₀)为第一列第一项的偶数信号，y₂(t₀)为第二列第一项的偶数信号，y₁(t₁)为第一列第一项的奇数信号，y₂(t₁)为第二列第一项的奇数信号；此时，数组H进行列变换可以得到

其中c₁(t₁)和d₁(t₁)为分别根据y₁(t₀)和y₁(t₁)进行列变换得到的第一列第一项的逼近信号和细节信号，c₂(t₂)和d₂(t₂)为分别为根据y₂(t₀)和y₂(t₁)进行列变换得到的第二列第一项的逼近信号和细节信号。Exemplarily, if some of the column pixels in a two-dimensional point matrix form an array

where y ₁ (t ₀ ) is the even signal of the first item in the first column, y ₂ (t ₀ ) is the even signal of the first item in the second column, and y ₁ (t ₁ ) is the odd signal of the first item in the first column signal, y ₂ (t ₁ ) is the odd signal of the first item in the second column; at this time, the column transformation of the array H can be obtained

where c ₁ (t ₁ ) and d ₁ (t ₁ ) are the approximation signal and detail signal of the first item of the first column obtained by column transformation according to y ₁ (t ₀ ) and y ₁ (t ₁ ) respectively, c ₂ (t ₂ ) and d ₂ (t ₂ ) are the approximation signal and the detail signal of the first item of the second column obtained by column transformation according to y ₂ (t ₀ ) and y ₂ (t ₁ ), respectively.

S2033. Perform component conversion and component replacement on the second detail signal and the second approximation signal in the column-transformed signal to obtain a two-dimensional component matrix.

That is, the pixel points described by the detail signal and the approximation signal are described by high and low frequency components, and the highest one dimension component (high and low frequency components) is replaced with zero components.

It is understandable that after performing row transformation and column transformation processing on each pixel in the synchronized image data, the high-frequency components (values) and low-frequency components (values) of the rows and columns of the synchronized image data are obtained. area, the smaller the high-frequency coefficient value (the inverse of the high-frequency component) is, you can choose to discard it; on the contrary, in the area with lower approximation, the high-frequency coefficient value can be selected to be retained. At this time, by adapting different prediction algorithms Different degrees of compression can be obtained by selecting different high-frequency coefficients, update operators, and discarding different high-frequency coefficients.

其中x₁(t₀)第一行的第一像素点，x₁(t₁)为第一行的第二个像素点，x₂(t₀)为第二行的第一个像素点，x₂(t₁)为第二行的第二个像素点，此时，数组A的变换过程的步骤如下：Exemplarily, if some of the pixels in a two-dimensional point matrix are arrays

where x ₁ (t ₀ ) is the first pixel in the first row, x ₁ (t ₁ ) is the second pixel in the first row, and x ₂ (t ₀ ) is the first pixel in the second row, x ₂ (t ₁ ) is the second pixel of the second row. At this time, the steps of the transformation process of array A are as follows:

其中，c₁(t₁)和d₁(t₁)分别为根据x₁(t₀)和x₁(t₁)进行行变换得到的第一项的第一逼近信号和第一项的第一细节信号，c₂(t₂)和d₂(t₂)为分别为根据x₂(t₀)和x₂(t₁)进行行变换得到的第二项的第一逼近信号和第二项的第二细节信号；且第一项的第一逼近信号c₁(t₁)和第一项的第一细节信号d₁(t₁)将作为步骤2(列变换)中的偶数信号；第二项的第一逼近信号c₂(t₂)和第二项的第一细节信号d₂(t₂)将作为步骤2(列变换)中的奇数信号。Step 1: Array A can be obtained by row transformation

Among them, c ₁ (t ₁ ) and d ₁ (t ₁ ) are the first approximation signal of the first item and the first approximation signal of the first item obtained by row transformation according to x ₁ (t ₀ ) and x ₁ (t ₁ ), respectively A detail signal, c ₂ (t ₂ ) and d ₂ (t ₂ ) are the first approximation signal and the second approximation signal of the second term obtained by row transformation according to x ₂ (t ₀ ) and x ₂ (t ₁ ), respectively the second detail signal of the term; and the first approximation signal c ₁ (t ₁ ) of the first term and the first detail signal d ₁ (t ₁ ) of the first term will be used as even signals in step 2 (column transformation); The first approximation signal c ₂ (t ₂ ) of the second term and the first detail signal d ₂ (t ₂ ) of the second term will be used as odd signals in step 2 (column transformation).

其中，c₁c(t₁)和d₁c(t₁)分别为根据f₁(t₀)和f₁(t₁)进行列变换之后得到的第一项的第二逼近信号和第一项的第二细节信号，c₂d(t₂)和d₂d(t₂)分别为根据f₂(t₀)和f₂(t₁)进行列变换之后得到的第二项的第二逼近信号和第二项的第二细节信号；Step 2: Column transformation of the array A' can get the array

Among them, c _{1 c} (t ₁ ) and d _{1 c} ( _{t 1 )} are the second approximation signal and first The _second detail signal of the term, c ₂ d(t ₂ ) and d ₂ d( _{t 2 )} are the _second the approximation signal and the second detail signal of the second term;

其中，L-L为低-低频分量，H-L为高-低频分量，L-H为低-高频分量，H-H为高-高频分量；Step 3: Convert the components of the array A" to get the array

Among them, LL is the low-low frequency component, HL is the high-low frequency component, LH is the low-high frequency component, and HH is the high-frequency component;

Step 4: Replace the components of array B (that is, replace the high-frequency components with 0) to obtain an array

In one embodiment, as shown in FIG. 6 , the step S40 specifically includes the following steps:

S401. Acquire the two-dimensional component matrix corresponding to the shared image data.

The shared image data is a two-dimensional component matrix (or image data) described by high and low frequency components (dimension components) obtained after compression and forward transformation.

It is understandable that when other vehicles and other vehicles transmit the shared image data described by high and low frequency components to the target vehicle, the shared image data needs to be restored to synchronous monitoring data (that is, two-dimensional points described by pixel points) through decompression and inverse transformation. matrix).

S402. Acquire a component signal including the second detail signal and the second approximation signal according to the two-dimensional component matrix.

That is, the component signal is similar to the column-transformed signal obtained by the row transformation and column transformation of the original signal.

S403. Perform inverse line transformation on the component signal to obtain a signal after inverse line transformation that includes the second even signal and the second odd signal.

Exemplarily, the decompressed forward and inverse transform is shown in FIG. 8 .

First, the second even signal is obtained according to the second update operator, the second detail signal and the second approximation signal. That is, the second update operator U _n [·] is introduced first, and the original second even signal _y′en (t) is obtained by using the known second detail signal and the second approximation signal in the component signal. It is shown in the following formula (4):

y′e _n (t)=c _n (t)-U _n [d _n (t)] (4)

Then, a second odd-numbered signal of the row of pixels is obtained according to the second predictor, the second even-numbered signal, and the second approximation signal. That is, the second predictor P _n [·] is introduced to obtain the original second odd signal _y′on (t ). It is shown in the following formula (5):

y′on (t)=d _n (t)+P _n [ _{y′e n (} t)] (5)

S404. Perform inverse column transformation on the inverse row-transformed signal, obtain a two-dimensional dot matrix including the first even signal and the first odd signal, and mark the two-dimensional dot matrix as the synchronization monitoring data.

Specifically, the first detail signal and the first approximation signal before column inverse transformation are obtained according to the second even signal and the second odd signal in the signal after inverse row transformation, and then according to the The first update operator U _m [ ], the first detail signal and the first approximation signal obtain the original first even signal x'e _m (t), and according to the first predictor P _m [·], the first even-numbered signal and the first approximation signal to obtain the original first odd-numbered signal x'o _m (t), and then restore the first odd-numbered signal x'om (t) obtained from the component signal through inverse row transformation and inverse column transformation. An even-numbered signal x'e _m (t) and a first odd-numbered signal x'o _m (t) are combined to obtain an original signal y'(t), that is, the two-dimensional dot matrix.

To sum up, in the driving planning method for unmanned vehicles provided by the present invention, the synchronous image is compressed and transformed into shared image data and transmitted to the cloud database, and the shared image data within the dynamic receiving range is obtained from the cloud database for processing. The decompression and inverse transformation is converted into synchronous monitoring data, so that unmanned vehicles can quickly transmit and share image data within their respective fields of view in crowded intersections and areas with high traffic flow, realize timely and effective processing of the body environment, improve the safety of driverless vehicles.

In one embodiment, as shown in FIG. 9 , an unmanned vehicle driving planning apparatus is provided, and the unmanned vehicle driving planning apparatus corresponds one-to-one with the driving planning method of the unmanned vehicle in the above-mentioned embodiment. The unmanned vehicle driving planning device includes a positioning module 110 , a collection module 120 , a receiving module 130 and a planning module 140 . The detailed description of each functional module is as follows:

The positioning module 110 is configured to receive the intelligent driving instruction sent by the target vehicle, and obtain the positioning information of the target vehicle.

The acquisition module 120 is configured to acquire synchronized image data within a preset monitoring range acquired by the binocular camera of the target vehicle.

The receiving module 130 is configured to determine the dynamic receiving range of the target vehicle according to the positioning information of the target vehicle, and obtain the shared image data within the dynamic receiving range from the cloud database; the shared image data refers to the current time period All vehicles in the vehicle are synchronously transmitted to the image data in the cloud database.

The decompression module 140 is configured to acquire synchronization monitoring data generated after decompressing and inverse transforming the shared image data.

The planning module 150 is configured to perform vehicle travel planning on the target vehicle according to the synchronized image data and the synchronized monitoring data.

In one embodiment, as shown in FIG. 9 , in the driving planning device for unmanned vehicles, the acquisition module 120 includes the following sub-modules, and each functional sub-module is described in detail as follows:

The synchronous acquisition sub-module 121 is used for multiple sets of binocular cameras to synchronously collect the synchronous image data in multiple ways; the binocular cameras include the front and rear, the left and right sides, the upper left, the lower right and the upper right of the target vehicle. The lower left binocular cameras, and each of the binocular cameras includes two single cameras.

The saving sub-module 122 is configured to save the synchronized image data collected by each of the binocular cameras into a two-dimensional point matrix; the two-dimensional point matrix refers to a two-dimensional matrix described by pixels.

The compression sub-module 123 is configured to obtain a two-dimensional component matrix after performing a compressed forward transformation on the two-dimensional point matrix; the two-dimensional component matrix refers to a two-dimensional matrix described by non-zero dimension components.

The transmission sub-module 124 is configured to mark the two-dimensional component matrix as shared image data, and transmit the shared image data to the cloud database.

In one embodiment, in the driving planning device for an unmanned vehicle, the compression sub-module includes the following units, and each functional unit is described in detail as follows:

A row transformation unit, configured to perform row transformation on the two-dimensional point matrix as the original signal, and obtain a row transformed signal including a first detail signal and a first approximation signal.

A column transformation unit, configured to perform column transformation on the row transformed signal to obtain a column transformed signal including a second detail signal and a second approximation signal.

A component processing unit, configured to perform component conversion and component replacement on the second detail signal and the second approximation signal in the column-transformed signal to obtain a two-dimensional component matrix.

In one embodiment, in the driving planning device for an unmanned vehicle, the row conversion unit includes the following subunits, and each functional subunit is described in detail as follows:

The first splitting subunit is used for splitting the row pixels in the original signal to obtain the first odd-numbered signal and the first even-numbered signal.

A first prediction subunit, configured to predict the first odd-numbered signal according to a preset first prediction operator, and obtain a first detail signal.

A first update subunit, configured to update the first even-number information according to a preset first update operator and the first detail signal, and obtain a first approximation signal.

In one embodiment, in the driving planning device for unmanned vehicles, the column transformation unit includes the following subunits, and each functional subunit is described in detail as follows:

The second splitting subunit is used for splitting the column pixels in the row-transformed signal to obtain a second odd-numbered signal and a second even-numbered signal.

The second prediction subunit is configured to predict the second odd-numbered signal according to a preset second prediction operator, and obtain a second detail signal.

The second update subunit is configured to update the second even-number information according to a preset second update operator and the second detail signal to obtain a second approximation signal.

In one embodiment, in the driving planning device for unmanned vehicles, the receiving module 130 includes the following sub-modules, and each functional sub-module is described in detail as follows:

A data acquisition sub-module for acquiring the two-dimensional component matrix corresponding to the shared image data.

A signal extraction submodule, configured to acquire a component signal including the second detail signal and the second approximation signal according to the two-dimensional component matrix.

The first inverse transform sub-module is configured to perform line inverse transform on the component signal, and obtain a line inversely transformed signal including the second even signal and the second odd signal.

The second inverse transform sub-module obtains a two-dimensional point matrix including the first even signal and the first odd signal according to the inverse column transform of the row inversely transformed signal, and converts the two-dimensional point matrix into Mark as the synchronization monitoring data.

For the specific limitation of the driving planning device for the unmanned vehicle, reference may be made to the limitation on the driving planning method for the unmanned vehicle above, which will not be repeated here. Each module in the above-mentioned driving planning device for an unmanned vehicle can be implemented in whole or in part by software, hardware and combinations thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

In one embodiment, a computer device is provided, and the computer device may be a server, and its internal structure diagram may be as shown in FIG. 10 . The computer device includes a processor, memory, a network interface and a database connected by a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium, an internal memory. The non-volatile storage medium stores an operating system, computer readable instructions and a database. The internal memory provides an environment for the execution of the operating system and computer-readable instructions in the non-volatile storage medium. The computer readable instructions, when executed by the processor, implement a method for planning a trip of an unmanned vehicle.

In one embodiment, a computer device is provided, comprising a memory, a processor, and computer-readable instructions stored on the memory and executable on the processor, and the processor implements the following steps when executing the computer-readable instructions:

Receive the intelligent driving instruction sent by the target vehicle, and obtain the positioning information of the target vehicle;

Acquiring synchronous image data within a preset monitoring range collected by the binocular camera of the target vehicle;

The dynamic receiving range of the target vehicle is determined according to the positioning information of the target vehicle, and the shared image data within the dynamic receiving range is obtained from the cloud database; the shared image data refers to the synchronous transmission of all vehicles in the current period to image data in the cloud database;

obtaining synchronous monitoring data generated after decompressing and inverse transforming the shared image data;

Carrying out vehicle travel planning for the target vehicle according to the synchronized image data and the synchronized monitoring data.

In one embodiment, a computer-readable storage medium is provided on which computer-readable instructions are stored, and when the computer-readable instructions are executed by a processor, the following steps are implemented:

Receive the intelligent driving instruction sent by the target vehicle, and obtain the positioning information of the target vehicle;

Acquiring synchronous image data within a preset monitoring range collected by the binocular camera of the target vehicle;

The dynamic receiving range of the target vehicle is determined according to the positioning information of the target vehicle, and the shared image data within the dynamic receiving range is obtained from the cloud database; the shared image data refers to the synchronous transmission of all vehicles in the current period to image data in the cloud database;

obtaining synchronous monitoring data generated after decompressing and inverse transforming the shared image data;

Carrying out vehicle travel planning for the target vehicle according to the synchronized image data and the synchronized monitoring data.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a non-volatile computer. In the readable storage medium, the computer-readable instructions, when executed, may include the processes of the foregoing method embodiments. Wherein, any reference to memory, storage, database or other medium used in the various embodiments provided by the present invention may include non-volatile and/or volatile memory. Nonvolatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units or modules is used for illustration. Module completion, that is, dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be used for the foregoing implementations. The technical solutions described in the examples are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the within the protection scope of the present invention.

Claims (10)

1. A method for planning the driving of an unmanned vehicle, comprising:

receiving an intelligent driving instruction sent by a target vehicle, and acquiring positioning information of the target vehicle;

acquiring synchronous image data in a preset monitoring range acquired by a binocular camera of the target vehicle;

determining a dynamic receiving range of the target vehicle according to the positioning information of the target vehicle, and acquiring shared image data in the dynamic receiving range from a cloud database; the shared image data refers to image data which are synchronously transmitted to the cloud database by all vehicles in the current time period;

acquiring synchronous monitoring data generated after the shared image data is subjected to decompression and inverse transformation;

and planning vehicle running of the target vehicle according to the synchronous image data and the synchronous monitoring data.

2. The unmanned vehicle driving planning method of claim 1, wherein the acquiring of the synchronous image data within the preset monitoring range collected by the binocular camera of the target vehicle comprises:

enabling a plurality of groups of binocular cameras to acquire the synchronous image data in a multipath synchronous manner; the binocular cameras comprise binocular cameras arranged in the front and back, the left and right sides, the upper left and lower right sides and the upper right and lower left sides of the target vehicle, and each binocular camera comprises two single cameras;

storing the synchronous image data acquired by each binocular camera into a two-dimensional point matrix; the two-dimensional point matrix is a two-dimensional matrix described by pixel points;

after the two-dimensional point matrix is subjected to compression forward transformation, a two-dimensional component matrix is obtained; the two-dimensional component matrix is a two-dimensional matrix described by using non-zero dimension components;

and marking the two-dimensional component matrix as shared image data, and transmitting the shared image data to the cloud database.

3. The unmanned vehicle driving planning method of claim 2, wherein obtaining a two-dimensional component matrix after the compressive forward transformation of the two-dimensional point matrix comprises:

performing row transformation on the two-dimensional point matrix serving as the original signal to obtain a row-transformed signal containing a first detail signal and a first approximation signal;

performing column transformation on the row-transformed signal to obtain a column-transformed signal containing a second detail signal and a second approximation signal;

and performing component conversion and component replacement on the second detail signal and the second approximation signal in the signals after the column transformation to obtain a two-dimensional component matrix.

4. The unmanned vehicle driving planning method of claim 3, wherein the performing line transformation on the two-dimensional point matrix as the original signal to obtain a line-transformed signal including a first detail signal and a first approximation signal comprises:

splitting the row pixels in the original signal to obtain a first odd signal and a first even signal;

predicting the first odd-numbered signal according to a preset first predictor to obtain a first detail signal;

and updating the first even number information according to a preset first updating operator and the first detail signal to obtain a first approximation signal.

5. The unmanned vehicle driving planning method of claim 3 wherein said performing a column transformation on said row transformed signals to obtain column transformed signals comprising second detail signals and second approximation signals comprises:

splitting the column pixels in the row-transformed signals to obtain a second odd signal and a second even signal;

predicting the second odd-numbered signal according to a preset second predictor to obtain a second detail signal;

and updating the second even number information according to a preset second updating operator and the second detail signal to obtain a second approximation signal.

6. The unmanned vehicle driving planning method of claim 3, wherein said obtaining the synchronized monitoring data generated after said decompressing inverse transformation of the shared image data comprises:

obtaining the two-dimensional component matrix corresponding to the shared image data;

acquiring a component signal comprising the second detail signal and the second approximation signal according to the two-dimensional component matrix;

performing inverse row transform on the component signals to obtain inverse row transformed signals including the second even-numbered signals and the second odd-numbered signals;

and performing column inverse transformation on the signals subjected to the row inverse transformation to obtain a two-dimensional point matrix containing the first even-numbered signal and the first odd-numbered signal, and marking the two-dimensional point matrix as the synchronous monitoring data.

7. An unmanned vehicle driving planning apparatus, comprising:

the positioning module is used for receiving an intelligent driving instruction sent by a target vehicle and acquiring positioning information of the target vehicle;

the acquisition module is used for acquiring synchronous image data in a preset monitoring range acquired by a binocular camera of the target vehicle;

the receiving module is used for determining the dynamic receiving range of the target vehicle according to the positioning information of the target vehicle and acquiring shared image data in the dynamic receiving range from a cloud database; the shared image data refers to image data which are synchronously transmitted to the cloud database by all vehicles in the current time period;

the decompression module is used for acquiring synchronous monitoring data generated after the shared image data is subjected to decompression inverse transformation;

and the planning module is used for planning the vehicle running of the target vehicle according to the synchronous image data and the synchronous monitoring data.

8. The unmanned vehicle driving planning apparatus of claim 7 wherein said acquisition module comprises:

the synchronous acquisition submodule is used for enabling a plurality of groups of binocular cameras to acquire the synchronous image data synchronously in a multipath manner; the binocular cameras comprise binocular cameras arranged in the front and back, the left and right sides, the upper left and lower right sides and the upper right and lower left sides of the target vehicle, and each binocular camera comprises two single cameras;

the storage submodule is used for storing the synchronous image data acquired by each binocular camera into a two-dimensional point matrix; the two-dimensional point matrix is a two-dimensional matrix described by pixel points;

the compression submodule is used for obtaining a two-dimensional component matrix after performing compression forward transformation on the two-dimensional point matrix; the two-dimensional component matrix is a two-dimensional matrix described by using non-zero dimension components;

and the transmission sub-module is used for marking the two-dimensional component matrix as shared image data and transmitting the shared image data to the cloud database.

9. A computer device comprising a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, wherein the processor when executing the computer readable instructions implements the driverless vehicle travel planning method of any of claims 1-6.

10. A computer readable storage medium storing computer readable instructions, wherein the computer readable instructions, when executed by a processor, implement a method for planning the driving of an unmanned vehicle according to any of claims 1 to 6.

Images