Vehicle control method, electronic equipment, parking system, vehicle and storage medium

Abstract

The invention discloses a vehicle control method, electronic equipment, a parking system, a vehicle and a storage medium, and the vehicle control method comprises the steps: obtaining radar data of a vehicle-mounted millimeter wave radar and image data of a vehicle-mounted camera; a control decision is obtained according to the radar data and the image data, the control strategy is obtained according to a first control decision and a second control decision, the first control decision is obtained based on data fusion of the radar data and the image data, and the second control decision is obtained based on data fusion of the radar data and the image data. The second control decision is obtained based on a sub-decision determined based on the radar data and a sub-decision determined based on the image data; and controlling the vehicle according to the control decision. According to the method and the system, the perception level of the vehicle to the surrounding environment can be improved, the system can accurately identify and judge interferents in the environment, making of a more accurate and safer vehicle control decision is facilitated, and therefore the safety and the stability of the system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicles, and in particular, to a vehicle control method, an electronic device, a parking system, a vehicle, and a non-volatile readable storage medium. Background Art

[0002] In related technologies, some vehicle control decision-making methods, such as parking control, mainly rely on an ultrasonic radar system or a fusion system of an ultrasonic radar and a camera. However, the ultrasonic radar performs well in detecting at relatively short distances, but may be limited when the distance exceeds 5 meters, and does not have the ability to measure speed and angle. While the camera is greatly affected in extreme weather and lighting environments, which may lead to a decrease in the accuracy of recognition and judgment.

[0003] Therefore, controlling a vehicle based on an ultrasonic radar system or a fusion system of an ultrasonic radar and a camera has certain limitations and needs to be further improved. Summary of the Invention

[0004] The present invention aims to at least solve one of the technical problems existing in the prior art. To this end, an object of the present invention is to provide a vehicle control method, which can improve the vehicle's perception level of the surrounding environment, enable the system to accurately identify and judge interference objects in the environment, contribute to formulating more accurate and safer vehicle control decisions, and thus improve the safety and stability of the system.

[0005] A second object of the present invention is to provide an electronic device.

[0006] A third object of the present invention is to provide a parking system.

[0007] A fourth object of the present invention is to provide a vehicle.

[0008] A fifth object of the present invention is to provide a non-volatile readable storage medium.

[0009] To achieve the above object, the vehicle control method according to the first aspect embodiment of the present invention includes: acquiring radar data of an in-vehicle millimeter-wave radar and image data of an in-vehicle camera; obtaining a control decision according to the radar data and the image data, where the control decision is obtained according to a first control decision and a second control decision, the first control decision is obtained based on data fusion of the radar data and the image data, and the second control decision is obtained based on a sub-decision determined according to the radar data and a sub-decision determined according to the image data; controlling the vehicle according to the control decision.

[0010] According to the vehicle control method of an embodiment of the present invention, by acquiring radar data of an in-vehicle millimeter-wave radar and image data of an in-vehicle camera, the system realizes the fusion of multi-sensor data. This data fusion method constitutes the first control decision. Since the millimeter-wave radar and the camera have unique sensing characteristics respectively, by fusing their data, the system can understand the environment around the vehicle more comprehensively and accurately. At the same time, the second control decision can be sub-decisions determined based on the radar data and the image data respectively. The sub-decision of the radar data can involve specific target detection and tracking, while the sub-decision of the image data can include the extraction and classification of visual features of the target. This separate processing method gives full play to the advantages of the millimeter-wave radar and the camera under different environmental conditions. By integrating the first and second control decisions, the system forms a more accurate and safer control strategy, enhances the ability to identify interference objects, enables the vehicle to adapt to various complex environments, and thus improves the safety and stability of the system.

[0011] In some embodiments, the data fusion of the radar data and the image data includes: an association step of associating the radar data and the image data, a maintenance step of maintaining the target during association, and a filtering step of evaluating and predicting the state of the associated data through filtering. After the filtering step, it returns to the association step.

[0012] In some embodiments, before the association step and after the filtering step, it further includes: a step of filtering the data before association through a threshold value.

[0013] In some embodiments, the second control decision is obtained by comparing the weighted result of the first control sub-decision and the second control sub-decision with a first decision threshold; wherein, the first control sub-decision is determined according to the radar data, and the second control sub-strategy is determined according to the image data.

[0014] In some embodiments, the weighting values of the first control sub-decision and the second control sub-decision are determined according to the vehicle driving scenario.

[0015] In some embodiments, the first control sub-decision is determined by successively performing clustering, target bounding, and target classification on the radar data.

[0016] In some embodiments, the second control sub-decision is determined by performing target bounding and target classification on the image data.

[0017] In some embodiments, the control decision is determined by comparing the sum result of the first control decision and the second control decision with a second decision threshold.

[0018] In some embodiments, the vehicle is in a parking condition.

[0019] In some embodiments, in response to a parking instruction, the vehicle-mounted millimeter-wave radar switches to a parking parameter configuration mode. In this parking parameter configuration mode, the sweep bandwidth of the vehicle-mounted millimeter-wave radar is greater than that in the driving mode, the detection distance is less than that in the driving mode, the data frame period is less than that in the driving mode, and the detection angle of view is greater than that in the driving mode.

[0020] In some embodiments, in response to the vehicle-mounted camera detecting a target, the vehicle-mounted millimeter-wave radar starts data detection.

[0021] To achieve the above object, an electronic device according to an embodiment of the second aspect of the present invention includes: a processor; a memory communicatively connected to the processor; a computer program stored in the memory that can be executed by the processor, and when the processor executes the computer program, the vehicle control method described in the above embodiments is implemented.

[0022] According to the electronic device of the embodiment of the present invention, when the processor adopts the vehicle control method described in the above embodiments, the vehicle's perception level of the surrounding environment can be improved, enabling the system to accurately identify and judge interference objects in the environment, which helps to make more accurate and safer vehicle control decisions, thereby improving the safety and stability of the system.

[0023] To achieve the above object, a parking system according to an embodiment of the third aspect of the present invention includes the electronic device described in the above embodiments.

[0024] According to the parking system of the embodiment of the present invention, by adopting the electronic device described in the above embodiments, based on the comprehensive analysis of the radar data of the vehicle-mounted millimeter-wave radar and the image data of the vehicle-mounted camera, the system can more comprehensively and accurately perceive the target and its surrounding environment, enhance the ability to identify interference objects, effectively reduce the possibility of misjudgment, and help to make more accurate and safer vehicle control decisions, thereby improving the safety and stability of the system.

[0025] To achieve the above object, a vehicle according to an embodiment of the fourth aspect of the present invention includes the electronic device described in the above embodiments.

[0026] According to the vehicle of the embodiment of the present invention, by adopting the electronic device described in the above embodiments, the vehicle's perception level of the surrounding environment can be improved, enabling the system to accurately identify and judge interference objects in the environment, which helps to make more accurate and safer vehicle control decisions, thereby improving the safety and stability of the vehicle.

[0027] In some embodiments, the vehicle further includes: a millimeter-wave radar connected to the vehicle control device for collecting radar data; and a camera connected to the vehicle control device for collecting image data.

[0028] In some embodiments, the millimeter-wave radar is disposed on one side of the front bumper and / or the rear bumper of the vehicle close to the vehicle body.

[0029] To achieve the above object, a non-volatile readable storage medium according to a fifth aspect embodiment of the present invention stores a computer program, and when the computer program is executed, it implements the vehicle control method described in the above embodiments.

[0030] According to the non-volatile readable storage medium of the embodiments of the present invention, by adopting the vehicle control method described in the above embodiments, the perception level of the vehicle for the surrounding environment can be improved, so that the system can accurately identify and judge interference objects in the environment, which helps to make more accurate and safer vehicle control decisions, thereby improving the safety and stability of the system.

[0031] Additional aspects and advantages of the present invention will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above and / or additional aspects and advantages of the present invention will become apparent and be readily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0033] Figure 1 is a flowchart of a vehicle control method according to an embodiment of the present invention;

[0034] Figure 2 is a schematic diagram of the fusion process of an in-vehicle millimeter-wave radar and an in-vehicle camera according to an embodiment of the present invention;

[0035] Figure 3 is a schematic diagram of generating a control decision according to an embodiment of the present invention;

[0036] Figure 4 is a schematic diagram of the first control decision processing process according to an embodiment of the present invention;

[0037] Figure 5 is a schematic diagram of the second control decision processing process according to an embodiment of the present invention;

[0038] Figure 6 is a schematic diagram of the dual control decision processing process according to an embodiment of the present invention;

[0039] Figure 7Schematic diagram of a parking system incorporating a vehicle-mounted millimeter-wave radar according to an embodiment of the present invention;

[0040] Figure 8 Schematic diagram of the use of a front radar and a corner radar in a driving scenario according to an embodiment of the present invention;

[0041] Figure 9 Schematic diagram of the method of using a vehicle-mounted millimeter-wave radar in a parking scenario according to an embodiment of the present invention;

[0042] Figure 10 Flowchart of the implementation process of a parking system according to an embodiment of the present invention;

[0043] Figure 11 Block diagram of an electronic device according to an embodiment of the present invention;

[0044] Figure 12 Block diagram of a parking system according to an embodiment of the present invention;

[0045] Figure 13 Block diagram of a vehicle according to an embodiment of the present invention.

[0046] Reference signs:

[0047] Vehicle 100;

[0048] Parking system 1;

[0049] Electronic device 10; millimeter-wave radar 20; camera 30; vehicle control device 40;

[0050] Processor 11; memory 12. Detailed implementation manners

[0051] The embodiments of the present invention will be described in detail below. The embodiments described with reference to the drawings are exemplary. The embodiments of the present invention will be described in detail below.

[0052] Below, reference is made to Figures 1-10 Describe a vehicle control method according to an embodiment of the present invention.

[0053] Figure 1 It is a flowchart of a vehicle control method according to an embodiment of the present invention. As Figure 1 shown, the vehicle control method at least includes the following steps S1 - S3.

[0054] S1, Obtain radar data of the vehicle-mounted millimeter-wave radar and image data of the vehicle-mounted camera.

[0055] In some embodiments, the radar data may include information such as the distance, azimuth, speed, and RCS (Radar Cross Section, i.e., the reflection area of the target object to the radar wave) of the target object. Among them, for the distance of the target, the vehicle-mounted millimeter-wave radar can determine the distance between the target and the radar by measuring the round-trip time of the signal. For the azimuth of the target object, the vehicle-mounted millimeter-wave radar can provide the horizontal azimuth angle and vertical pitch angle of the target object relative to the radar. For the speed of the target object, the vehicle-mounted millimeter-wave radar can calculate the speed of the target object by measuring the positions of the target object at multiple moments. For the RCS of the target object, the RCS can reflect the reflection degree of the target object to the radar signal, that is, the reflection effect of the target object under the illumination of the radar wave, so as to provide the size and shape information of the target object. The larger the RCS value, the easier it is for the target object to be detected in the radar system. Different types of target objects usually have different RCSs. By measuring the RCS of the target object, the system can help identify and classify different types of target objects, such as vehicles, pedestrians, buildings, etc.

[0056] In some embodiments, the image data may include information such as target object feature information, road signs, and the surrounding environment. Among them, the target object feature information may be the visual features of the target object, including shape, color, texture, etc. The road signs may be traffic signs, lane markings, etc. The surrounding environment may include other vehicles, pedestrians, buildings, etc.

[0057] In some embodiments, as Figure 2 shown, the process of fusing the radar data of the vehicle-mounted millimeter-wave radar and the image data of the vehicle-mounted camera may include the vehicle-mounted camera capturing an image of the surrounding environment and extracting the feature information of the target object from the image. This may include features such as the shape, size, and color of the target object. The vehicle-mounted millimeter-wave radar can measure the distance, azimuth, speed, and RCS of the target object by transmitting millimeter waves and receiving the reflected signals. Together, these information provide the position and motion state of the target object in the vehicle coordinate system.

[0058] Furthermore, the data information obtained by the vehicle-mounted camera and the vehicle-mounted millimeter-wave radar are respectively transmitted to the processor (SOC, System on Chip) of the system for further processing. The SOC performs fusion processing on the data from the vehicle-mounted camera and the vehicle-mounted millimeter-wave radar. This may involve combining the feature information of the target object extracted by the camera with the distance, azimuth, speed, and RCS information of the target object detected by the millimeter-wave radar. Through the fusion processing, the system can more comprehensively and accurately understand the characteristics and motion state of the target object.

[0059] Furthermore, after the processing is completed, the system converts the fused target data into readable information, which usually includes the category of the target object (such as a ball) and specific position information (such as 15 meters away from the vehicle, azimuth 30°, speed 0 m / s).

[0060] In some embodiments, the data of the millimeter-wave radar is usually represented in the form of point clouds, including the x, y, z coordinates, speed, and RCS of the target object. Compared with image data, the point clouds of on-vehicle millimeter-wave radars may be relatively sparse. This is because the millimeter-wave has a short wavelength (usually in the millimeter range), which means its beam is relatively narrow. When the target object is small or far away, only a small amount of the target object's surface reflects back to the radar, resulting in a weak received reflection signal and relatively sparse point clouds. In addition, the weak resolution is also the reason for the sparse point clouds.

[0061] In some embodiments, for the characteristic of sparse point clouds of on-vehicle millimeter-wave radars, some special algorithms can be adopted for processing, such as super-resolution DOA algorithms like DML and MUSIC. By using virtual aperture technology to expand the antenna aperture to achieve channel separation, when applied to point cloud data, these technologies can involve mapping the angle information measured by the on-vehicle millimeter-wave radar to three-dimensional coordinates to improve the accuracy of target positioning and direction resolution.

[0062] S2. Obtain a control decision according to the radar data and the image data.

[0063] In some embodiments, as Figure 3 shown, the control decision can be obtained based on the first control decision and the second control decision, where the first control decision can be obtained through the data fusion of the radar data and the image data. Specifically, the system can convert the point cloud data of the on-vehicle millimeter-wave radar to the image plane to form a radar image. This radar image can include information such as the distance, azimuth, speed, and RCS of the target object, while the image data contains the visual features of the target object on the image. When performing fusion, the system adopts a data fusion strategy, which means that the system extracts features from the radar data and the image data and maps them to a common feature space to better understand the relationship between them.

[0064] In some embodiments, the extracted features are mapped to a common feature space. This is to ensure that the features from different data sources can be compared and fused under a unified framework. This mapping can be achieved through different mathematical transformations, such as linear transformation or non-linear mapping. The features mapped to the common feature space are fused together. This can include simple weighted addition or may also include more complex fusion algorithms, such as neural networks, machine learning models, etc. The fusion algorithm can learn the relationship between the data and generate a more comprehensive feature representation.

[0065] Therefore, based on the first control decision obtained from the data fusion of radar data and image data, the system can obtain more comprehensive and multi-dimensional information about the target object, improve the recognition accuracy of the target object, enhance the environmental perception ability, and provide a more abundant information basis for vehicle control decisions. This helps the system to more accurately judge the state of the target object, and thus formulate safer and more intelligent vehicle control strategies.

[0066] In some embodiments, the second control decision can be obtained from a sub-decision based on radar data and a sub-decision based on image data. Specifically, in decision-level fusion, the system can match the detection results from the on-vehicle camera and the on-vehicle millimeter-wave radar and calculate the similarity between them. This means that the system can compare the targets detected by the camera and the millimeter-wave radar respectively, find out their common features, and then use a similarity metric (such as a distance metric or other similarity measurement methods) to determine whether they represent the same target object. If the similarity of the detection results of the two sensors is high, the system can determine the existence of the target object and fuse this information to obtain a more comprehensive and reliable result.

[0067] Furthermore, the system integrates the sub-decision based on radar data and the sub-decision based on image data into the second control decision. This may involve assigning weights to the sub-decisions of the two or synthesizing their logical relationships to generate a more comprehensive and accurate control decision. The integrated decision can provide more abundant information, enabling the vehicle system to more intelligently respond to different scenarios.

[0068] Therefore, by integrating the detection results from different sensors (on-vehicle camera and on-vehicle millimeter-wave radar), the cognitive accuracy and reliability of the system for the target object can be improved. Thus, the performance of the entire system in target perception and decision-making can be enhanced.

[0069] In some embodiments, while the on-vehicle millimeter-wave radar transmits RAW data (raw data) to the SOC, it also transmits the target object information generated by its own algorithm. Here, the RAW data can refer to the unprocessed original signal received by the on-vehicle millimeter-wave radar, usually the radar waveform signal or point cloud data. These data contain the signal information reflected by the target object, but have not been filtered, feature-extracted, target-recognized, etc. The target information generated by its own algorithm can refer to that the on-vehicle millimeter-wave radar usually has some signal processing and target recognition algorithms built in. These algorithms can process the RAW data and identify information such as the position, speed, and size of the target. The processed target information can refer to the data automatically generated by the radar system based on the internal algorithm of the radar to describe the target characteristics.

[0070] Therefore, the in-vehicle millimeter-wave radar can transmit these two types of data to the SOC simultaneously. After transmitting the RAW data to the SOC, the SOC can use various algorithms to further process these data, such as filtering, feature extraction, target recognition, etc. On the other hand, the target information generated by the internal algorithm of the millimeter-wave radar can also be transmitted to the SOC for the SOC to refer to or verify. By comprehensively using these two types of data, the intelligent system can more accurately understand the surrounding environment, perform target recognition and tracking, so as to support applications such as autonomous driving and parking assistance.

[0071] In some embodiments, the control decision can be real-time to ensure timely response during vehicle driving and improve the safety of the system. Specifically, the in-vehicle millimeter-wave radar and the in-vehicle camera can usually collect data at a frequency of multiple times per second. By maintaining high-frequency data updates, the system can more timely sense the environment around the vehicle. Also, the data transmission from the sensor to the control system can be real-time. By adopting a high-bandwidth data transmission channel, such as Fast Ethernet or other dedicated communication protocols, it is ensured that the sensor data can be timely transmitted to the control unit for processing.

[0072] In addition, the generation of the control decision can be through real-time processing of the sensor data. This can involve efficient algorithms and parallel computing to ensure the analysis and understanding of complex data are completed in a short time, so that decisions can be quickly generated after receiving the latest sensor data. By combining these methods, the system can better achieve real-time perception of environmental changes and corresponding control decisions, thereby improving the safety of the vehicle system.

[0073] S3, control the vehicle according to the control decision.

[0074] Specifically, the system can perform corresponding vehicle control operations according to the control decision, which can include adjusting the vehicle speed, steering angle, or taking other measures to respond to the detected target object or environmental changes, so as to ensure that the vehicle makes an adaptive response to the surrounding environment. For example, if the system detects an obstacle ahead, according to the comprehensive data, the system can decide to decelerate or change lanes to avoid collision. If the system recognizes the scenario of an intersection, it can make corresponding control strategies, such as stopping, accelerating through, etc. In addition, the system can further consider specific scenarios, such as rainy days, snowy days or nights. In this case, the control decision may be adjusted to more cautiously respond to complex driving situations, such as decelerating under low visibility conditions to ensure the safe driving of the vehicle.

[0075] According to the vehicle control method of an embodiment of the present invention, by acquiring the radar data of the in-vehicle millimeter-wave radar and the image data of the in-vehicle camera, the system realizes the fusion of multi-sensor data. This data fusion method constitutes the first control decision. Since the millimeter-wave radar and the camera have unique sensing characteristics respectively, by fusing their data, the system can understand the environment around the vehicle more comprehensively and accurately. At the same time, the second control decision can be sub-decisions determined based on the radar data and the image data respectively. The sub-decision of the radar data can involve specific target detection and tracking, while the sub-decision of the image data can include the extraction and classification of the visual features of the target. This separate processing method gives full play to the advantages of the millimeter-wave radar and the camera under different environmental conditions. By integrating the first and second control decisions, the system forms a more accurate and safer control strategy, enhances the ability to identify interference objects, enables the vehicle to adapt to various complex environments, and thus improves the safety and stability of the system.

[0076] Figure 4 is a schematic diagram of the first control decision processing process according to an embodiment of the present invention, as Figure 4 shown, the data fusion of the radar data and the image data includes: an association step of associating the radar data and the image data, a maintenance step of maintaining the target during association, and a filtering step of evaluating and predicting the state of the associated data through filtering. After the filtering step, it returns to the association step.

[0077] Among them, in the association step, the system can associate the target detection result collected by the in-vehicle camera, that is, the image data, with the tracking target collected by the in-vehicle millimeter-wave radar, that is, the radar data, to determine whether they come from the same target object. In the maintenance step, the target can be maintained on the basis of the association to ensure the continuous tracking and update of the target. For example, when a new target is identified in the association step, this new target can be added to the target tracking system to ensure that the system can continuously track and manage this target. When a certain target is confirmed as an invalid target in the association step, this target can be deleted from the target tracking system. The situations of deleting the target can include that the target leaves the sensor detection range, is blocked by other objects, or is recognized as an invalid target by the system, etc.

[0078] In some embodiments, in the filtering step of state evaluation and prediction, a Kalman Filter can be used to process the state information of the target object. The Kalman Filter can be a recursive mathematical algorithm used to estimate the state of a dynamic system. Its main purpose is to evaluate the state of the current target object through known measurement values and predict the state of the target object at the next moment.

[0079] Specifically, the Kalman filter uses the measurement value at the current moment to estimate the state of the target object. This process is called "state evaluation", which uses the observation data provided by the sensor and combines the dynamic model of the system to estimate the state parameters of the target object at the current moment, such as position and velocity. By considering measurement errors and system noise, the Kalman filter can provide an optimal estimate of the state of the target object.

[0080] Furthermore, after estimating the state of the target object at the current moment, the Kalman filter uses the dynamic model of the system to predict the state of the target object at the next moment. This prediction process is called "state prediction". The Kalman filter can consider the dynamic characteristics of the system and predict the state parameters of the target object at future moments, such as position and velocity.

[0081] In this way, the Kalman filter can continuously update the estimate of the state of the target object, while considering the dynamic characteristics of the system and the uncertainty of sensor measurements, so as to provide an optimal estimate of the target state. In the target tracking system, this state estimation and prediction process helps to improve the accuracy and stability of target object tracking.

[0082] In some embodiments, during the process of target tracking, the data before association may contain noise or false detections from sensors. If these noisy data are not processed, they may interfere with subsequent target association and state estimation. Therefore, before the association step and after the filtering step, noise filtering of the data before the association step by a threshold value can be an effective preprocessing means, aiming to exclude those data that may be false detections or unreliable.

[0083] Specifically, an appropriate threshold value can be set, and this threshold value can be adjusted according to the actual requirements and performance of the system. The selection of the threshold value should consider the sensitivity of the target tracking system and the noise level in a specific environment. By applying the threshold value to the data before association, only the target objects with signal strength higher than the threshold value can be retained, and those signals lower than the threshold value are identified as noise or false detections to remove unreliable data.

[0084] Therefore, by performing the noise filtering step before the association step, not only can the possibility of false association be reduced, but also the utilization of computing resources can be optimized. This is because only the reliable data after noise filtering participate in the subsequent association calculation, thus reducing the computational amount.

[0085] In some embodiments, such as Figure 4As shown, the correlation step, the maintenance step, the filtering step for state evaluation and prediction, and the step of noise filtering for the data before correlation form a negative feedback loop. At each moment, the system will perform steps such as target correlation, target maintenance, state estimation and prediction, and noise filtering. After the above steps, the system will output an optimized decision value at the current moment. This decision value can include state information such as the position, direction, and speed of the target object, as well as the system's understanding of the current environment. This output decision value can be passed to the vehicle controller as its input. The controller can formulate real-time vehicle control strategies based on this information to adapt to the position and motion state of the current target object. By continuously obtaining data from sensors, performing target tracking and decision-making, the system can sense changes in the environment around the vehicle and take corresponding control strategies to adapt to different driving scenarios, thereby achieving safer and more stable vehicle control.

[0086] Figure 5 is a schematic diagram of the second control decision-making process according to an embodiment of the present invention. As Figure 5 shown, the second control decision can be obtained by comparing the weighted result of the first control sub-decision and the second control sub-decision with the first decision threshold. Among them, the first control sub-decision can be determined according to radar data, which means using the information of the on-vehicle millimeter-wave radar to generate a control strategy. The radar data provides information such as the distance, direction, and speed of the target object, which helps the system understand the relative position and motion state of the target object around the vehicle. The second control sub-strategy can be determined according to image data. This means that the image information provided by the on-vehicle camera can be used to generate another control strategy. The on-vehicle camera can identify the features, shapes, etc. of the target and provide a visual understanding of the target object.

[0087] Specifically, the first control sub-decision and the second control sub-decision are weighted to obtain a weighted result. This can be achieved by assigning weights to each sub-decision and performing a weighted sum on them. Then, this weighted result is compared with the set first decision threshold. If the weighted result exceeds or meets the first decision threshold, the system can take corresponding control actions such as braking, steering, etc. The first decision threshold can be a preset value used to decide whether to adopt the second control decision.

[0088] In some embodiments, the assignment of weights can be adjusted according to system performance and environmental requirements. For example, if the radar data is more reliable in certain situations, a higher weight can be assigned to the first control sub-decision. Such adjustments can be dynamically optimized according to the environment and driving conditions in which the vehicle is located.

[0089] In some embodiments, the weighted values ​​of the first control sub-decision and the second control sub-decision may be determined based on a vehicle driving scenario. The vehicle driving scenario may involve various situations, such as city streets, highways, complex intersections, etc. Each scenario may have different importance for different types of sensor data. For example, in an urban environment, image data from an onboard camera may be more important because more visual information is needed to cope with complex traffic conditions; while on a highway, radar data may be more critical because better distance and speed information is needed.

[0090] In addition, for real-time vehicle control systems, the adjustment of weight values ​​may also need to consider the real-time performance of the system. Even in the same scenario, different real-time conditions (such as traffic density, weather conditions) may lead to different weight distributions. For example, in rainy or sleety conditions, the on-board camera may be affected by raindrops, resulting in a decrease in image quality. At this time, radar data may be more reliable because millimeter-wave radar is not affected by raindrops and can provide relatively reliable distance and speed information. Therefore, in severe weather conditions, the weight of radar data can be increased and the reliance on image data can be reduced. In sunny conditions, the on-board camera can provide clear visual information, including vehicles, pedestrians, and road signs. At this time, the weight of image data can be appropriately increased to make better use of high-quality image data.

[0091] In some embodiments, the first control sub-decision may be determined by sequentially clustering, target framing, and target classification of radar data.

[0092] Among them, radar data can be clustered by DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm. DBSCAN can be a density-based clustering algorithm. It divides data points into clusters with similar density and can identify noise points (outliers) without presetting the number of clusters. Here, DBSCAN is used to cluster the radar data of millimeter-wave radar into target clusters to better identify targets.

[0093] Therefore, clustering can be the process of grouping the points in the radar data according to their characteristics. The vehicle-mounted millimeter-wave radar usually returns discrete point cloud data, each point containing information such as the position and speed of the target. Through the clustering algorithm, these point cloud data can be divided into different target clusters, each cluster representing a detected target. Through clustering, the detection results of the millimeter-wave radar can be better understood, different targets can be identified, and more accurate information can be provided for subsequent decision-making.

[0094] In some embodiments, target bounding can be to generate a bounding box for each target object based on clustering, which is used to represent the position and shape of the target object. By determining the bounding box of the target, the system can more conveniently perform subsequent feature extraction, classification, and decision-making.

[0095] In some embodiments, target classification can be to classify the targets to determine their types, such as vehicles, pedestrians, obstacles, etc. Among them, the vehicle-mounted millimeter-wave radar classifies the targets according to the bounding box and makes decisions according to the characteristics of the target objects, such as decelerating, braking, or changing lanes. The classification results are very important for formulating appropriate vehicle control strategies because different types of target objects may trigger different control strategies to ensure that the vehicle responds appropriately to the surrounding environment.

[0096] In some embodiments, the second control sub-decision can be determined by performing target bounding and target classification on the image data. Specifically, in the processing of the image data of the vehicle-mounted camera, explicit clustering is usually not required as in the case of the vehicle-mounted millimeter-wave radar. This is because the characteristics of the images of the vehicle-mounted camera itself already have spatial information and visual continuity, making the boundaries of the target objects clearly defined in the image. In contrast, the point cloud data of the vehicle-mounted millimeter-wave radar appears as sparse scattered points in the image, and clustering algorithms may be required to identify and group them.

[0097] In some embodiments, the vehicle-mounted camera can perform CNN convolution calculations according to the bounding box. Deep learning methods such as convolutional neural networks (CNNs) perform excellently in the field of image processing and can learn complex features from images to achieve accurate classification of different target categories. The target classification results can include different types of targets such as vehicles, pedestrians, traffic signs, etc. According to the characteristics of the target objects, the system can generate corresponding decisions, which can include operations such as decelerating, braking, and changing lanes for the vehicle.

[0098] Therefore, through the above steps, the system can obtain the spatial position information (bounding) of the target object and the semantic information (classification) of the target from the image data. Such target recognition and classification results will be used to generate the second control sub-decision. The second control sub-decision can provide detailed information about the target objects detected in the image data, supporting the vehicle control system to make more intelligent and adaptive decisions.

[0099] Figure 6 is a schematic diagram of a dual control decision processing process according to an embodiment of the present invention. As Figure 6 shown, the control decision can be determined by comparing the sum result of the first control decision and the second control decision with the second decision threshold.

[0100] Specifically, the radar data and the image data are fused at the data feature level to generate a first control decision. This stage may include steps such as correlating the radar data and the image data, target maintenance, state assessment, and prediction to obtain the perception of the overall environment and the estimation of the state of the target object. At the same time, the radar data and the image data are fused at the decision-making level to generate a second control decision. This stage may include respectively framing and classifying the targets for the image data and the radar data, which can provide a more detailed understanding of each detected target, including its type, location, speed, etc. The first control decision and the second control decision may include control strategies such as decelerating, braking, and lane-changing for the vehicle.

[0101] Furthermore, the results of the first control decision and the second control decision are summed to form a comprehensive control decision. This process may involve weighted summation of weights. The setting of the second decision threshold can be adjusted according to specific driving scenarios and system requirements. In different driving environments, different decision sensitivities may be required to balance the system's response speed and the risk of misoperation.

[0102] Furthermore, the comparison result will determine what action the system will ultimately take. If the summation result of the comprehensive decision exceeds the second decision threshold, the system can execute corresponding control operations, such as taking emergency braking, performing obstacle avoidance behavior, etc. Conversely, if the summation result does not reach the threshold, the system may maintain the current state or adopt a lighter control strategy to avoid unnecessary intervention.

[0103] For example, through the sensors in the vehicle sensor system, such as on-vehicle cameras, humidity sensors and other devices, the current environment is sensed, and it is judged that the current scene is rainy. At this time, the visibility is low, and the weight occupied by the on-vehicle camera is reduced, assumed to be 30%, while the weight of the on-vehicle millimeter-wave radar can be 70%. Then this weight can be brought into the strategy of fusing the two control decisions. Through fusion at the data feature level, it can be obtained that the weight ratio of the first control decision is 40%, and the first control decision is to decelerate. Through fusion at the decision-making level, it can be obtained that the weight ratio of the second control decision is 60%, and the second control decision is also to decelerate. By weighted summing the weights of the first control decision and the second control decision, a decision of 100% deceleration can be obtained. Comparing it with the set second decision threshold of 70%, the final decision can be made to decelerate.

[0104] For another example, if it is determined that the current scenario is a vehicle driving on a highway in clear weather, since the interference is relatively low at this time, the highway has high requirements for speed detection, while the speed measurement ability of the camera is weak. Therefore, the weight of the in-vehicle millimeter-wave radar is increased, assumed to be 70%, and the weight of the in-vehicle camera is 30%. Then, this weight can be brought into the strategy of fusing two control decisions. By fusing at the data feature layer, it can be obtained that the weight ratio of the first control decision is 30%, and the first control decision is to decelerate. By fusing at the decision layer, it can be obtained that the weight ratio of the second control decision is 70%, where the decision weight of the in-vehicle millimeter-wave radar is relatively high at 90%, and the in-vehicle camera is 10%. In the decision layer, when using the in-vehicle millimeter-wave radar alone, the decision weight is 90%; when using the in-vehicle camera alone, the decision weight is 60%. The weight determined by the decision layer is: (90% * 90% + 10% * 60%) * 70% = 60.9%. Therefore, by combining the two decision-making methods, the result is: 60.9% (the judgment made by the decision layer fusion) + 30% (the judgment made by the data feature layer fusion) = 90.9%. The same result is higher than the second decision threshold of 70%. By comparing the process of obtaining this value of 90.9% with the process of obtaining 100% previously, various factors of single-sensor limitations are comprehensively considered. Therefore, the result has better robustness and can better adapt to driving decisions under different weather and road conditions.

[0105] In some embodiments, the vehicle is in a parking condition. In the parking condition, the driver usually faces potential safety hazards that may be caused by vehicles driving quickly from the front or rear. Since the driver cannot comprehensively observe the dynamics of distant vehicles when reversing, they may overlook the vehicles approaching quickly, resulting in accidents. Therefore, by combining the fusion information of the in-vehicle millimeter-wave radar and the in-vehicle camera, more comprehensive and accurate parking assistance information can be provided to the driver to help the driver complete the parking action more safely.

[0106] Specifically, during the parking process, if the target object is within the detection range of the on-vehicle camera, the on-vehicle millimeter-wave radar can be activated based on the target information feedback by the on-vehicle camera at this time, and the relative coordinates of the interfering vehicle with respect to the vehicle being parked can be calculated. Specifically, the camera can compare multiple frames of data to analyze the moving distance, and then calculate the approximate speed based on the time interval between frames. If the interfering vehicle is a potential target, switch to the on-vehicle millimeter-wave radar for precise speed measurement and label the target vehicle. If the interfering vehicle poses a potential threat to the parked vehicle, once the system detects this situation, it can promptly notify the driver of the information. This can be achieved through means such as the on-vehicle display screen, sound prompts, or vibrations. After notifying the driver, the driver can promptly understand the running conditions of the vehicle in the distance, thereby taking corresponding preventive measures to avoid collisions or accidents with the approaching vehicle quickly, improving the safety of the parking process.

[0107] As Figure 7 shown, the on-vehicle millimeter-wave radar can provide accurate distance, angle, speed and other information of the target object in the parking scenario, and is not easily affected by the external environment, which is crucial for precise parking in a limited space. Therefore, the radar data can help the system detect surrounding obstacles, boundaries and other vehicles, as well as evaluate their positions and motion states relative to the vehicle. The on-vehicle camera provides image data, which is crucial for visual recognition and understanding in the parking scenario. The image data can be used to detect and identify important elements in the environment such as parking spaces, signs, pedestrians, etc.

[0108] By integrating the data of these two sensors, the system can establish a comprehensive understanding of the parking scenario. During the parking process, the system can use this information for operations such as path planning, obstacle avoidance, and vehicle position adjustment. Among them, the distance information provided by the on-vehicle millimeter-wave radar helps to determine the positions of surrounding obstacles, while the image information provided by the on-vehicle camera can help the system understand the structure of the surrounding environment, such as the position and size of the parking space. These two sensors provide complementary information, and their combined use can enhance the perception of the parking scenario to ensure that the vehicle can complete the parking operation safely and efficiently.

[0109] In some embodiments, in response to a parking instruction issued by a driver's manual operation or an autonomous driving system, the in-vehicle millimeter-wave radar can switch its function through the reverse gear, switching from the angle radar and front radar functions originally used during normal driving to a parking parameter configuration mode for short-distance requirements, so as to improve the detection performance of nearby target objects. The purpose of this switching design is that in the driving mode, due to the requirements of the ADAS (Advanced Driving Assistance System) scenario function, the in-vehicle millimeter-wave radar needs to ensure a longer detection distance, a larger detection speed range, a higher target resolution, etc. However, for the parking scenario, these performances are not that important or the parking scenario has its own parameter requirements for the in-vehicle millimeter-wave radar. Therefore, the in-vehicle millimeter-wave radar can be switched to the parking parameter configuration mode to ensure the functional requirements of the parking scenario.

[0110] In some embodiments, in the parking parameter configuration mode, the in-vehicle millimeter-wave radar adjusts its parameters to meet the requirements of the parking scenario. These parameter adjustments can include sweep bandwidth, detection distance, data frame period, detection angle of view, etc.

[0111] Among them, the sweep bandwidth of the in-vehicle millimeter-wave radar in the parking mode is greater than that in the driving mode. The larger sweep bandwidth helps to improve the resolution of the radar system for nearby target objects, enabling the system to more accurately detect surrounding obstacles in a limited parking space. The detection distance in the parking mode is less than that in the driving mode. This is because the parking scenario usually involves approaching stationary or slowly moving obstacles, and the detection distance may be shortened to improve the perception accuracy at close range.

[0112] In addition, the data frame period in the parking mode is less than that in the driving mode. In an embodiment, the data frame period generally refers to the time required for the radar system to process and output a complete data frame. The in-vehicle millimeter-wave radar uses FMCW (Frequency Modulated Continuous Wave) signals to measure information such as the distance and speed of a target. The FMCW radar sends a continuously frequency-modulated signal within a cycle, then receives the reflected signal, and obtains target information by measuring the frequency change.

[0113] In some embodiments, for a traditional parking system based on in-vehicle ultrasonic radar, due to the inherent characteristics of the material, there will be an inherent blind spot of about 20 cm. However, for a parking system based on in-vehicle millimeter-wave radar, since electromagnetic waves are used, electromagnetic waves do not involve the after-vibration problem faced by mechanical waves. Therefore, there is no inherent blind spot, and only the period of the FMCW signal and the algorithm processing need to be adjusted to the shortest based on the scenario to reduce the detection blind spot.

[0114] Specifically, by analyzing the parking scenario in detail, including the site layout, vehicle parking positions, characteristics of obstacles, etc. Understanding the scenario characteristics helps optimize the configuration of in-vehicle millimeter-wave radar. By adjusting the period of the FMCW signal, it can be optimized according to the scenario characteristics. A shorter period helps improve the system's resolution for near targets, increases the system's update frequency for target objects, helps obtain the dynamic information of the target more timely, and improves the perception ability of the target state during parking.

[0115] In some embodiments, the detection view angle in the parking mode is greater than that in the driving mode. As Figure 8 shown, in the driving scenario, a longer detection distance is required, so multi-antenna beamforming is needed to increase the detection distance. However, the increase in detection distance may lead to the narrowing of the FOV (Field of Vision), that is, the detection view angle range becomes narrower. In the parking scenario, a larger FOV is required to reduce the near-blind area. Therefore, antenna beamforming is not required to ensure a wider view angle and reduce the blind area.

[0116] Therefore, as Figure 9 shown, for the parking scenario, by configuring the parking parameters of the in-vehicle millimeter-wave radar, the detection blind area can be significantly reduced, enabling the in-vehicle millimeter-wave radar to more comprehensively perceive the environment around the vehicle. At the same time, it can also make it have the functions of traditional ultrasonic radars in some aspects, such as near-obstacle detection and precise distance measurement, thus greatly improving the safety of parking.

[0117] In some embodiments, in response to the in-vehicle camera detecting a target, the in-vehicle millimeter-wave radar starts data detection. This means that when the in-vehicle camera identifies a potential target object in the surrounding environment, which may be other vehicles, pedestrians, or obstacles, the system can trigger the data detection process of the in-vehicle millimeter-wave radar. This may include the radar system starting to scan the target, obtaining data such as distance, angle, and speed, and generating corresponding data frames. These data can be used for further target recognition, tracking, and decision-making to support the vehicle control system to make corresponding response actions, such as decelerating, braking, or changing lanes. Therefore, by starting the data detection of the millimeter-wave radar when the camera detects a target, the system can more reliably perceive and understand the surrounding environment and improve the ability to identify and respond to potential dangers or collision risks.

[0118] Figure 10 is a flowchart of the implementation process of a parking system according to an embodiment of the present invention. As Figure 10 shown, the implementation process of the parking system at least includes the following steps S10 - S14.

[0119] S10, start parking.

[0120] S11. The in-vehicle camera detects information on target objects in the front and rear.

[0121] S12. Activate the in-vehicle millimeter-wave radar to detect the target object.

[0122] S13. Determine whether it will interfere with the parking process. If so, go to step S14; if not, return to step S11.

[0123] S14. Remind the driver to take corresponding control decisions.

[0124] In summary, by comprehensively utilizing the advantages of the camera and the millimeter-wave radar and fusing the data of both, the accuracy and reliability of target detection are improved to ensure that the vehicle safely and efficiently completes the parking operation, avoid collisions or accidents with vehicles approaching rapidly, and enhance the safety of the parking process.

[0125] Based on the vehicle control method of the above embodiment, the following refers to Figure 11 Describe the electronic device of the embodiment of the present invention.

[0126] Figure 11 is a block diagram of an electronic device according to an embodiment of the present invention. As Figure 11 shown, the electronic device 10 includes: a processor 11 and a memory 12.

[0127] Among them, the processor 11 can be the core component responsible for executing computer programs. The processor 11 can be a central processing unit (CPU) or other processing units suitable for executing instructions.

[0128] In some embodiments, the memory 12 is communicatively connected to the processor 11 and is used to store computer programs and other related data. The memory 12 can include random access memory (RAM), read-only memory (ROM), flash memory, etc. A computer program executable by the processor 11 is stored in the memory 12, and when the processor 11 executes the computer program, the vehicle control method described in the above embodiment is implemented.

[0129] For the electronic device 10 according to the embodiment of the present invention, when the processor 11 adopts the vehicle control method described in the above embodiment, the vehicle's perception level of the surrounding environment can be improved, enabling the system to accurately identify and judge interference objects in the environment, which helps to make more accurate and safer vehicle control decisions, thereby improving the safety and stability of the system.

[0130] The following refers to Figure 12 Describe the parking system of the embodiment of the present invention.

[0131] Figure 12 is a block diagram of a parking system according to an embodiment of the present invention. AsFigure 12 As shown in Figure 12 , the parking system 1 includes: an electronic device 10.

[0132] In some embodiments, the parking system 1 can be a system for assisting or automating the vehicle parking process. It can include various sensors, actuators, and control units to provide intelligent control and support for the vehicle during the parking process.

[0133] According to the parking system 1 of the embodiments of the present invention, by adopting the electronic device 10 described in the above embodiments, based on the comprehensive analysis of the radar data of the vehicle-mounted millimeter-wave radar and the image data of the vehicle-mounted camera, the system can more comprehensively and accurately perceive the target and its surrounding environment, enhance the ability to identify interference objects, effectively reduce the possibility of misjudgment, and contribute to formulating more accurate and safer vehicle control decisions, thereby improving the safety and stability of the system.

[0134] In practical applications, the parking system 1 can include an automatic parking function, which intelligently controls the movement of the vehicle in the parking scenario through the electronic device 10 to achieve automatic parking. In addition, the parking system 1 can also communicate with other vehicles to achieve cooperative parking and improve the parking efficiency.

[0135] Next, refer to Figure 13 to describe the vehicle of the embodiments of the present invention.

[0136] Figure 13 is a block diagram of a vehicle according to an embodiment of the present invention, as Figure 13 shown, the vehicle 100 includes the electronic device 10, the millimeter-wave radar 20, and the camera 30 described in the above embodiments.

[0137] Among them, the electronic device 10 includes a processor 11, a memory 12 communicatively connected to the processor 11, and a computer program stored therein that can be executed by the processor 11 to implement the vehicle control method described in the above embodiments. The millimeter-wave radar 20 is connected to the vehicle control device 40 and is used to collect radar data. The millimeter-wave radar 20 can provide high-precision distance, speed, and direction information about the surrounding environment, which is crucial for target detection and environment understanding during the parking process. The camera 30 is connected to the vehicle control device 40 and is used to collect image data. The camera 30 can provide visual information, including features such as the shape and color of the target, which helps the system to more comprehensively understand the environment around the vehicle.

[0138] According to the vehicle 100 of the embodiments of the present invention, by adopting the electronic device 10 described in the above embodiments, the vehicle 100 can improve the perception level of the surrounding environment, enable the system to accurately identify and judge interference objects in the environment, contribute to formulating more accurate and safer vehicle control decisions, thereby improving the safety and stability of the vehicle 100.

[0139] In some embodiments, the millimeter-wave radar 20 is disposed on one side of the front bumper and / or the rear bumper of the vehicle 100 close to the vehicle body, and has the ability to penetrate the bumper with electromagnetic wave radiation. However, the ultrasonic radar does not have the ability to penetrate the bumper. Therefore, compared with the traditional ultrasonic radar, the millimeter-wave radar 20 has better detection performance and does not need to worry about the influence of detection performance caused by rain, snow, or dirt coverage. This improves the stability and reliability of the detection system under different weather conditions. In addition, the millimeter-wave radar 20 is disposed on one side of the front bumper and / or the rear bumper of the vehicle 100 close to the vehicle body, which can make the radar more concealed, reduce the impact on the appearance of the vehicle 100, and still be able to effectively sense the environment in front of and behind the vehicle 100.

[0140] In some embodiments, the vehicle 100 may be configured with multiple millimeter-wave radars 20, which are distributed at different positions to provide omnidirectional environmental perception. For example, the millimeter-wave radar 20 on the front bumper can be used for detecting front targets, and the radar on the rear bumper can be used for detecting rear targets. This multi-radar configuration helps to achieve a more comprehensive and accurate perception of the parking scenario.

[0141] In addition, the millimeter-wave radar 20 can also be combined with other sensors, such as the camera 30 or the ultrasonic sensor, to provide more information for environmental perception and target detection. This configuration of integrated sensors helps to improve the robustness and applicability of the system and adapt to different parking scenarios and complex traffic environments.

[0142] In some embodiments of the present invention, a non-volatile readable storage medium is also proposed, on which a computer program is stored. When the computer program is executed, the vehicle control method described in the above embodiments is implemented.

[0143] In some embodiments, the non-volatile readable storage medium may be a medium for storing a computer program, which is non-volatile, that is, the data remains stored after power-off or shutdown. This can be a storage medium similar to a flash memory, a hard disk drive, a CD-ROM, etc.

[0144] According to the non-volatile readable storage medium of the embodiments of the present invention, by adopting the vehicle control method described in the above embodiments, the perception level of the vehicle 100 of the surrounding environment can be improved, the system can accurately identify and judge the interference objects in the environment, which helps to make more accurate and safer vehicle control decisions, thereby improving the safety and stability of the system.

[0145] In the description of this specification, the descriptions with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic descriptions of the above terms do not necessarily refer to the same embodiment or example.

[0146] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A vehicle control method, characterized in that, Including: Obtaining radar data of an in-vehicle millimeter-wave radar and image data of an in-vehicle camera; Obtaining a control decision based on the radar data and the image data, wherein the control decision is obtained based on a first control decision and a second control decision, the first control decision is obtained based on data fusion of the radar data and the image data, and the second control decision is obtained based on a sub-decision determined based on the radar data and a sub-decision determined based on the image data; Controlling the vehicle according to the control decision.

2. The vehicle control method according to claim 1, characterized in that: The data fusion of the radar data and the image data includes: an association step of associating the radar data and the image data, a maintenance step of maintaining the target during association, and a filtering step of performing state evaluation and prediction on the associated data through filtering, and then returning to the association step after the filtering step.

3. The vehicle control method according to claim 2, wherein Before the association step and after the filtering step, there is also included: a step of performing noise filtering on the data before association through a threshold value.

4. The vehicle control method according to claim 1, characterized in that: The second control decision is obtained by comparing the weighted result of a first control sub-decision and a second control sub-decision with a first decision threshold; wherein, the first control sub-decision is determined based on the radar data, and the second control sub-strategy is determined based on the image data.

5. The vehicle control method according to claim 4, wherein The weighting values of the first control sub-decision and the second control sub-decision are determined according to the vehicle driving scenario.

6. The vehicle control method according to claim 4, wherein, The first control sub-decision is determined by successively performing clustering, target bounding, and target classification on the radar data.

7. The vehicle control method according to claim 4, characterized in that, The second control sub-decision is determined by performing target bounding and target classification on the image data.

8. The vehicle control method according to claim 1, characterized in that, The control decision is determined by comparing the sum result of the first control decision and the second control decision with a second decision threshold.

9. The vehicle control method according to any one of claims 1-8, characterized in that, The vehicle is in a parking working condition.

10. The vehicle control method according to claim 9, wherein In response to a parking instruction, the in-vehicle millimeter-wave radar switches to a parking parameter configuration mode, wherein in the parking parameter configuration mode, the sweep bandwidth of the in-vehicle millimeter-wave radar is greater than that in the driving mode, the detection distance is less than that in the driving mode, the data frame period is less than that in the driving mode, and the detection angle is greater than that in the driving mode.

11. The vehicle control method according to claim 10, characterized in that, In response to the in-vehicle camera detecting a target, the in-vehicle millimeter-wave radar starts data detection.

12. An electronic device, characterized in that, Including: A processor; A memory communicatively connected to the processor; A computer program executable by the processor is stored in the memory, and when the processor executes the computer program, the vehicle control method according to any one of claims 1-11 is implemented.

13. A parking system, characterized in that, Including the electronic device according to claim 12.

14. A vehicle, characterized in that, Including the electronic device according to claim 13.

15. The vehicle according to claim 14, characterized in that, The vehicle further includes: A millimeter-wave radar, which is connected to the vehicle control device and is used for collecting radar data; A camera, which is connected to the vehicle control device and is used for collecting image data.

16. The vehicle according to claim 15, characterized in that, The millimeter-wave radar is arranged on one side of the front bumper and / or the rear bumper of the vehicle close to the vehicle body.

17. A non-volatile readable storage medium, on which a computer program is stored, characterized in that, When the computer program is executed, it implements the vehicle control method according to any one of claims 1-11.

Images