Vehicle control method, device, storage medium and processor

Abstract

The present invention discloses a vehicle control method, device, storage medium, and processor. The method comprises: obtaining environmental information of at least one road object collected during a current time period; inputting this environmental information into a trajectory prediction model to obtain a predicted trajectory of the road object in a future time period; and controlling the vehicle's driving state based on the predicted trajectory of the road object. This invention addresses the technical problem of low vehicle trajectory prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the field of intelligent vehicle technology, and in particular to a vehicle control method, device, storage medium and processor. Background Art

[0002] With the rapid development of artificial intelligence technology, autonomous driving has become a major development direction for future transportation. In autonomous driving scenarios, predicting the driving trajectory of other roads can help vehicles make correct decisions and improve driving safety.

[0003] Currently, the main method for predicting the driving state of road objects is to use their state and a pre-established behavioral rule library. However, due to the limited road scenarios covered by this behavioral rule library, it is difficult to accurately predict the driving trajectory of road objects in complex traffic conditions. In such cases, the vehicle cannot determine its own driving state based on the driving trajectories of other road objects, resulting in low vehicle trajectory prediction accuracy.

[0004] Currently, no effective solution has been proposed to the above-mentioned problem of low accuracy in vehicle trajectory prediction. Summary of the Invention

[0005] Embodiments of the present invention provide a vehicle control method, device, storage medium, and processor to at least solve the technical problem of low accuracy in vehicle trajectory prediction.

[0006] According to one aspect of an embodiment of the present invention, a vehicle control method is provided. The method may include: obtaining environmental information of at least one road object collected during a current time period, wherein the road object is a road object within the vehicle's information collection range; inputting the environmental information into a trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object during a future time period, wherein the trajectory prediction model is pre-trained on an original trajectory prediction model based on historical environmental information and historical driving trajectory information of the road object during a historical time period; and controlling the vehicle's driving state based on the predicted driving trajectory of the road object.

[0007] Optionally, the environmental information is input into a trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in a future time period, including: adding feature values ​​to the environmental information to obtain a feature vector corresponding to the environmental information; and inputting the feature vector into the trajectory prediction model for processing to obtain a predicted driving trajectory.

[0008] Optionally, the historical time period includes a first historical time period and a second historical time period, wherein the second historical time period is a time period after the first historical time period and adjacent to the first historical time period. The method further includes: inputting environmental information of the road object in the first historical time period into the original trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in the second historical time period; and adjusting parameters of the original trajectory prediction model based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period to obtain a trajectory prediction model.

[0009] Optionally, adjusting the parameters of the original trajectory prediction model based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period includes: determining the actual driving trajectory of the road object in the second historical time period based on the historical driving trajectory information of the road object in the historical time period; determining a matching rate between the predicted driving trajectory and the actual driving trajectory based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period; and adjusting the parameters of the original trajectory model based on the matching rate.

[0010] Optionally, the method further includes: in response to a matching rate between the predicted driving trajectory and the actual driving trajectory being less than a first threshold, adjusting the parameters of the original trajectory prediction model and increasing the number of training rounds of the original trajectory prediction model until the matching rate between the predicted driving trajectory and the actual driving trajectory is not less than the first threshold, and then determining the current original trajectory prediction model as the trajectory prediction model.

[0011] Optionally, controlling the driving state of the vehicle based on the predicted driving trajectory of the road object includes: determining the driving strategy of the vehicle based on the predicted driving trajectory of the road object, wherein the driving strategy is used to characterize the driving trajectory of the vehicle in a future time period; controlling the driving state of the vehicle based on the driving strategy, wherein the driving state includes one of the following: left turn, right turn, lane change to the left, lane change to the right, cruising, emergency stop, U-turn and unknown state.

[0012] According to another aspect of an embodiment of the present invention, a vehicle control device is also provided, including: an acquisition module for acquiring environmental information of at least one road object collected within a current time period, wherein the road object is a road object within the information collection range of the vehicle; a prediction module for inputting the environmental information into a trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in a future time period, wherein the trajectory prediction model is obtained by pre-training an original trajectory prediction model based on historical environmental information and historical driving trajectory information of the road object in a historical time period; and a control module for controlling the driving state of the vehicle based on the predicted driving trajectory of the road object.

[0013] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device containing the computer-readable storage medium is controlled to execute the vehicle control method according to an embodiment of the present invention.

[0014] According to another aspect of an embodiment of the present invention, a processor is provided, wherein the processor is configured to run a program, wherein when the program is run, the vehicle control method according to an embodiment of the present invention is executed.

[0015] According to another aspect of an embodiment of the present invention, a vehicle is provided. The vehicle is configured to execute the vehicle control method according to an embodiment of the present invention.

[0016] In an embodiment of the present invention, environmental information of at least one road object collected within the current time period is obtained, wherein the road object is a road object within the information collection range of the vehicle; the obtained environmental information is input into a trajectory prediction model for prediction to obtain the driving trajectory of the road object in the future time period; and the driving state of the vehicle is controlled based on the predicted driving trajectory of the road object. In other words, the vehicle control method provided by the embodiment of the present invention can predict the driving trajectory of the road object based on the environmental information obtained of at least one road object within the information collection range of the current vehicle, and then control the driving state of the current vehicle based on the predicted driving trajectory of the road object. Since the predicted driving trajectory of the road object is predicted based on the environmental information obtained in real time, the predicted driving trajectory of the road object is more accurate, and then controlling the driving state of the vehicle based on the predicted driving trajectory of the road object can improve the safety of vehicle driving and solve the technical problem of low accuracy in vehicle driving trajectory prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 is a flow chart of a vehicle control method according to an embodiment of the present invention;

[0019] Figure 2 is a schematic diagram of a trajectory prediction model according to an embodiment of the present invention;

[0020] Figure 3 is a schematic diagram of a vehicle control device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] Example 1

[0024] According to an embodiment of the present invention, an embodiment of a vehicle control method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0025] Figure 1 is a flow chart of a vehicle control method according to an embodiment of the present invention. Figure 1 As shown, the method may include the following steps:

[0026] Step S101 : Acquire environmental information of at least one road object collected in a current time period, wherein the road object is a road object within an information collection range of a vehicle.

[0027] In the technical solution provided in step S101 of the present invention, environmental information of at least one road object collected during a current time period is obtained. The control device may use a radar and / or image acquisition device on the vehicle to collect environmental information of at least one road object within the vehicle's information collection range during the current time period. The at least one road object may include pedestrians, motor vehicles, and non-motor vehicles within the vehicle's information collection range, without specific limitation herein.

[0028] Optionally, the radar and / or image acquisition device on the vehicle may collect environmental information of road objects within the vehicle's information collection range at a preset time interval. The preset time interval may be pre-set, for example, 0.5s or 0.8s, which is not limited herein.

[0029] Optionally, the control device obtains environmental information of road objects collected during a current time period, where the current time period may be a time period prior to the current moment, for example, 1 second prior to the current moment or 2 seconds prior to the current moment, without specific limitation herein. Based on this, the control device may obtain environmental information of road objects collected at preset time intervals by a radar and / or image acquisition device on the vehicle during the current time period.

[0030] Optionally, when there is only one road object within the vehicle's information collection range, the acquired environmental information is the environmental information of the single road object; when there are multiple road objects within the vehicle's information collection range, the acquired environmental information is the environmental information of multiple road objects. The acquired environmental information of the road object includes at least one of the following: speed information of the road object, scene information, traffic light information ahead, speed and distance information of surrounding road objects, and map information of the road object's location.

[0031] Step S102: Input the environmental information into a trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in a future time period.

[0032] In the technical solution provided in the above step S102 of the present invention, the environmental information of the road object is input into the trajectory prediction model for prediction, wherein the trajectory prediction model is a model obtained by pre-training the original trajectory prediction model based on the historical environmental information and historical driving trajectory information of the road object in a historical time period. The trajectory prediction model can predict the driving trajectory of the road object based on the environmental information of the road object and output the predicted driving trajectory of the road object.

[0033] Optionally, before inputting the environmental information of the road object into the trajectory prediction model for prediction, feature values ​​can be added to the environmental information of the road object obtained within the current time period to obtain a feature vector corresponding to the environmental information of the road object, and then the obtained feature vector is input into the trajectory prediction model for processing to obtain the predicted driving trajectory of the road object.

[0034] Step S103: Control the driving state of the vehicle based on the predicted driving trajectory of the road object.

[0035] In the technical solution provided in step S103 of the present invention, the vehicle's driving state is controlled based on the predicted driving trajectory of a road object. As described above, the road object may be one or multiple. When there is only one road object, the control device may determine the vehicle's driving strategy based on the predicted driving trajectory of the single road object. This driving strategy may be used to characterize the vehicle's driving trajectory over a future time period. When there are multiple road objects, the control device may also determine the vehicle's driving strategy based on the predicted driving trajectories of the multiple road objects.

[0036] After determining the vehicle's driving strategy, the control device can control the vehicle's driving state based on the vehicle's driving strategy. The vehicle's driving state can include any one of a left turn, a right turn, a left lane change, a right lane change, cruising, an emergency stop, a U-turn, and an unknown state.

[0037] In the above steps S101 to S103 of the present application, environmental information of at least one road object collected within the current time period is obtained, wherein the road object is a road object within the information collection range of the vehicle; the obtained environmental information is input into the trajectory prediction model for prediction to obtain the driving trajectory of the road object in the future time period; and the driving state of the vehicle is controlled based on the predicted driving trajectory of the road object. In other words, the vehicle control method provided by the embodiment of the present invention can predict the driving trajectory of the road object based on the environmental information of at least one road object within the information collection range of the current vehicle, and then control the driving state of the current vehicle based on the predicted driving trajectory of the road object. Since the predicted driving trajectory of the road object is predicted based on the environmental information obtained in real time, the predicted driving trajectory of the road object is more accurate, and then the driving state of the vehicle is controlled based on the predicted driving trajectory of the road object, which can improve the safety of vehicle driving and solve the technical problem of low accuracy in vehicle driving trajectory prediction.

[0038] The above method of this embodiment is further introduced below.

[0039] As an optional implementation method, step S102 inputs the environmental information into the trajectory prediction model for prediction to obtain the predicted driving trajectory of the road object in the future time period, including: adding feature values ​​to the environmental information to obtain a feature vector corresponding to the environmental information; inputting the feature vector into the trajectory prediction model for processing to obtain the predicted driving trajectory.

[0040] In this embodiment, the environmental information acquired by the control device for each road object includes at least one of the following: the road object's speed, the scene in which it is located, information about the traffic ahead, speed and distance information for surrounding road objects, and map information of the road object's location. In this case, the control device may first compile the environmental information acquired for each road object at preset time intervals during the current time period into an environmental information set. The control device may then input each environmental information set into a pre-trained target processing algorithm for processing. This target processing algorithm may add feature values ​​to the information included in each environmental information set and output a feature vector corresponding to each environmental information set.

[0041] For example, taking a certain moment in the current time period as an example, assuming that the environmental information of a road object obtained by the control device at this moment includes the speed information of the road object, the scene information, and the traffic light information ahead. Among them, the speed information of the road object can be expressed as A t Indicates that the scene information of the road object can be expressed as B t Indicates that the traffic light information ahead can be used with C t Indicates that the environmental information set of the road object can be expressed as X t Based on this, the environmental information set of the road object can be expressed as X t ={A t 、B t 、C t The control device can collect the environmental information X t ={A t 、B t 、C t The target processing algorithm then adds features and normalizes the environmental information set, outputting a feature vector corresponding to the environmental information set. Using the same method, the control device can obtain feature vectors corresponding to the environmental information set for the road object at different times within the current time period.

[0042] After obtaining the feature vectors corresponding to the environmental information set of the road object at different times in the current time period, the control device can input the multiple feature vectors obtained into the trajectory prediction model. The trajectory prediction model can process the feature vectors corresponding to the environmental information set of the road object at different times input by the control device, and then output the predicted driving trajectory of the road object in the future time period.

[0043] As an optional embodiment, before step S102, the process further includes training the original trajectory prediction model based on the historical environmental information and historical driving trajectory information of the road object in the historical time period to obtain the trajectory prediction model. Training the original trajectory prediction model based on the historical environmental information and historical driving trajectory information of the road object to obtain the trajectory prediction model includes: obtaining the historical environmental information and historical trajectory information of the road object in the historical time period, the historical time period including a first historical time period and a second historical time period, wherein the second historical time period is a time period after the first historical time period and adjacent to the first historical time period; inputting the environmental information of the road object in the first historical time period into the original trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in the second historical time period; and adjusting the parameters of the original trajectory prediction model based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period to obtain the trajectory prediction model.

[0044] In this embodiment, the historical time period is a time period before the current time. The historical time period can be a time period before the current time period, or it can include the current time period, without specific limitation. The control device can obtain historical environmental information and historical driving trajectory information for road objects within the vehicle's information collection range during the historical time period using the vehicle's radar and / or image acquisition device. After obtaining the historical environmental information and historical driving trajectory information for the road objects during the historical time period, the control device can divide the historical time period. For example, the historical time period can be divided into two adjacent time periods. For ease of explanation, the two adjacent historical time periods can be referred to as a first historical time period and a second historical time period. The first historical time period and the second historical time period can be of equal or unequal length. The control device can then obtain historical environmental information for the road objects during the first historical time period. The environmental information for the road objects during the first historical time period includes at least one of the following: speed information for the road objects, scene information, traffic light information ahead, speed and distance information for surrounding road objects, and map information for the road objects' location. The control device may also obtain historical driving trajectory information of the road object in the second historical time period, wherein the historical driving trajectory information includes an actual driving trajectory of the road object in the second historical time period.

[0045] Optionally, the control device may input the historical environmental information sets of the road objects at different times acquired within the first historical time period into the target processing algorithm based on the method described above, and obtain multiple feature vectors corresponding to the multiple historical environmental information sets of the road objects within the first historical time period. The control device may input the multiple feature vectors into the original trajectory prediction model. The original trajectory prediction model may include a bidirectional GRU neural network model and a feedforward neural network model, wherein, after the control device inputs the feature vectors corresponding to the multiple environmental information sets of the road objects within the first historical time period into the original trajectory prediction model, the bidirectional GRU neural network model is first used for processing, and the information processed by the bidirectional GRU neural network model is then processed by the feedforward neural network, and the feedforward neural network model may output the predicted driving trajectory of the road object within the second historical time period. It should be noted that when the control device inputs multiple feature vectors of the road object in the first historical time period into the original trajectory prediction model, it can also input the actual driving trajectory of the road object in the second historical time period into the original trajectory prediction model. Based on this, after the original trajectory prediction model predicts the driving trajectory of the road object in the second historical time period, the predicted driving trajectory in the second historical time period can be compared with the actual driving trajectory, and then the parameters of the original trajectory prediction model can be adjusted to obtain a trajectory prediction model.

[0046] Optionally, based on the predicted driving estimate and the actual driving trajectory of the road object in the second historical time period, adjusting the parameters of the original trajectory prediction model to obtain the trajectory prediction model includes: determining the actual driving trajectory of the road object in the second historical time period based on the historical driving trajectory of the road object in the historical time period; determining a matching rate between the predicted driving trajectory and the actual driving trajectory based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period; and adjusting the parameters of the original trajectory prediction model based on the matching rate.

[0047] In this embodiment, if the matching rate between the predicted driving trajectory of the road object in the second historical time period output by the original trajectory prediction model and the actual driving trajectory of the road object in the second historical time period is less than a first threshold, the parameters of the original trajectory prediction model are adjusted and the number of training rounds of the original trajectory prediction model is increased until the matching rate between the predicted driving trajectory of the road object in the second historical time period and the actual driving trajectory is no less than the first threshold. In this case, the current original trajectory prediction model is determined as the final trajectory prediction model. The first threshold can be preset, for example, the first threshold can be set to 85%, which is not limited here.

[0048] For example, after the original trajectory prediction model outputs a predicted driving trajectory of a road object within a second historical time period, it compares the predicted driving trajectory with the actual driving trajectory within the second historical time period to obtain a matching rate between the predicted driving trajectory and the actual driving trajectory of the road object within the second historical time period. If the matching rate is less than 85%, it indicates that the deviation between the predicted driving trajectory of the road object output by the original trajectory prediction model and the actual driving trajectory of the road object is large, and the accuracy of the original trajectory prediction model is low. In this case, the original trajectory prediction model can adaptively adjust parameters and increase the number of training rounds. After each training, the output predicted driving trajectory can be compared with the actual driving trajectory until the matching rate between the predicted driving trajectory of the road object output by the original trajectory prediction model and the actual driving trajectory of the road object is no less than 85%. This indicates that the predicted driving trajectory output by the original trajectory prediction model is closely consistent with the actual driving trajectory of the road object, that is, the prediction accuracy of the original trajectory prediction model is high. At this point, the current original trajectory prediction model can be used as the final trajectory prediction model. Subsequently, the vehicle control device can use this trajectory prediction model to predict the driving trajectory of road objects within the vehicle's information collection range.

[0049] As an optional embodiment, step S103 controls the driving state of the vehicle based on the predicted driving trajectory of the road object, including: determining the driving strategy of the vehicle based on the predicted driving trajectory of the road object, wherein the driving strategy is used to characterize the driving trajectory of the vehicle in a future time period; and controlling the driving state of the vehicle based on the driving strategy, wherein the driving state of the vehicle includes any one of left turn, right turn, left lane change, right lane change, cruising, emergency stop, U-turn and unknown state.

[0050] In this embodiment, after the control device determines the predicted driving trajectory of a road object within the vehicle's information collection range, it can determine the vehicle's driving strategy based on the determined predicted driving trajectory of the road object. For example, if the predicted driving trajectory of the road object directly in front of the vehicle is an emergency stop, then to avoid a rear-end collision, the control device can control the vehicle to an emergency stop based on the predicted driving trajectory of the road object in front of the vehicle, thereby avoiding the occurrence of a collision accident. In other words, the control device can determine the vehicle's driving strategy based on the predicted driving trajectory of the road object within the vehicle's information collection range, and then control the vehicle's driving state based on the driving strategy.

[0051] This embodiment obtains environmental information of at least one road object collected within the current time period, wherein the road object is a road object within the information collection range of the vehicle; inputs the obtained environmental information into a trajectory prediction model for prediction to obtain the driving trajectory of the road object in the future time period; and controls the driving state of the vehicle based on the predicted driving trajectory of the road object. In other words, the vehicle control method provided by the embodiment of the present invention can predict the driving trajectory of the road object based on the environmental information obtained for at least one road object within the information collection range of the current vehicle, and then control the driving state of the current vehicle based on the predicted driving trajectory of the road object. Since the predicted driving trajectory of the road object is predicted based on the environmental information obtained in real time, the predicted driving trajectory of the road object is more accurate. Therefore, controlling the driving state of the vehicle based on the predicted driving trajectory of the road object can improve the safety of vehicle driving and solve the technical problem of low accuracy in vehicle driving trajectory prediction.

[0052] Example 2

[0053] The technical solutions of the embodiments of the present invention are described below with reference to preferred implementation methods.

[0054] Currently, autonomous driving systems need to deal with many unexpected scenarios, such as overtaking by other vehicles, vehicle conflicts at intersections, pedestrians crossing the road, etc. When faced with these complex traffic scenarios, autonomous vehicles usually make decisions based on instantaneous states. However, this method of making decisions based on instantaneous states has poor adaptability to environmental development and changes, and is prone to delays when dealing with some complex scenarios, leading to collision risks. In this case, how to improve the accuracy of vehicle trajectory prediction becomes particularly important.

[0055] To overcome these issues, a related technique has proposed a rule-based method for predicting the trajectory of road objects. This method uses a behavioral rule library based on formal rules, traffic regulations, and common driving knowledge to predict the trajectory of road objects. However, due to the limited scenarios covered by the behavioral rule library, this method cannot accurately predict the trajectory of road objects in some unexpected situations.

[0056] However, an embodiment of the present invention proposes using a trajectory prediction model to predict the driving trajectory of a road object. This method uses a pre-trained trajectory prediction model to predict the driving trajectory of a road object. The method obtains environmental information about the road object, processes the acquired environmental information, and obtains a feature vector. The feature vector is then input into the trajectory prediction model for prediction, thereby obtaining a predicted driving trajectory of the road object. Because the predicted driving trajectory of the road object is predicted based on the environmental information of the road object, the predicted driving trajectory of the road object is more consistent with the actual situation. In other words, the predicted driving trajectory of the road object is more accurate. Based on this, the driving strategy of the vehicle determined according to the driving trajectory of the road object can better improve the safety of vehicle driving, thereby solving the technical problem of low accuracy in vehicle driving trajectory prediction.

[0057] Next, the training method of the trajectory prediction model provided by the embodiment of the present invention is further introduced with examples. The method may include the following steps:

[0058] The first step is to obtain the historical environment information and historical trajectory information of the road object in the historical time period.

[0059] The vehicle's radar and / or image acquisition device can continuously acquire, at preset time intervals, environmental information of road objects within the vehicle's information acquisition range and driving trajectory information. This environmental information may include at least one of the road object's speed information, scene information, traffic light information ahead, speed and distance information of surrounding road objects, and map information of the road object's location. This trajectory information includes a driving estimate of the road object. Taking a particular road object as an example, the control device can acquire, from the multiple pieces of environmental information of the road object acquired by the vehicle's radar and / or image acquisition device, multiple pieces of environmental information of the road object acquired at preset time intervals within a historical time period and the driving trajectory information of the road object within the historical time period.

[0060] In the second step, the acquired environmental information is processed into a feature vector in the range of [0,1] for training the trajectory prediction model.

[0061] After acquiring the environmental information and driving trajectory information of a road object within a historical time period, the control device may divide the historical time period into two adjacent time periods, referred to as a first historical time period and a second historical time period, with the second historical time period being located after the first historical time period. Based on a target processing algorithm, the control device may process the multiple sets of environmental information acquired for the road object within the first historical time period into feature vectors within the range [0, 1], thereby obtaining multiple feature vectors. Furthermore, the control device may determine the actual driving trajectory of the road object within the second historical time period from the historical driving trajectory information of the road object within the historical time period.

[0062] The third step is to train the trajectory prediction model.

[0063] After determining multiple feature vectors corresponding to multiple sets of environmental information of the road object in the first historical time period and the actual driving trajectory of the road object in the second historical time period, the control device can input the multiple feature vectors in the first time period into an original trajectory prediction model and train the original trajectory prediction model. The original trajectory prediction model outputs a predicted driving trajectory of the road object in the second historical time period based on the multiple feature vectors in the first historical time period, and compares the predicted driving trajectory with the actual driving trajectory in the second time period to determine the accuracy of the driving trajectory predicted by the original trajectory prediction model.

[0064] For example, Figure 2 FIG is a schematic diagram of a trajectory prediction model according to an embodiment of the present invention. Figure 2 As shown, the trajectory prediction model includes a bidirectional GRU neural network model and a feedforward neural network model. The control device can calculate the feature vector X corresponding to the environmental information collected at preset time intervals in the first historical time period. t-1 、X t 、X t+1 Input into the bidirectional GRU neural network model, which includes forward GRU and backward GRU. Among them, the forward GRU and backward GRU can process the data separately and then output the hidden layer information in two directions. Among them, the hidden layer information output by the forward GRU can be used To express, the hidden layer information output by the backward GRU can be used After the bidirectional GRU neural network model outputs the hidden layer information in two directions, the hidden layer information in the two directions can be input into the feedforward neural network model, which includes an input layer, hidden layer 1, hidden layer 2, and an output layer. After the hidden layer information in the two directions is processed by the input layer, hidden layer 1, hidden layer 2, and output layer of the feedforward neural network model, the behavior state T of the road object at time t-1, t, and t+1 can be output. t-1 、T t 、T t+1 , then, the trajectory prediction model can be based on the predicted behavior state T of the road object at each moment t-1 、T t 、T t+1 , determining a predicted driving trajectory of the road object, and then comparing the predicted driving trajectory of the road object in the second historical time period with the pre-input actual driving trajectory of the road object in the second historical time period to determine a matching rate between the predicted driving trajectory and the actual driving trajectory. If the matching rate between the predicted driving trajectory and the actual driving trajectory is less than a first threshold, the trajectory prediction model can adaptively adjust parameters, increase the number of training rounds, and compare the predicted driving estimate output after each training with the actual driving trajectory. If the matching rate between the predicted driving trajectory and the actual driving trajectory is less than the first threshold, training is continued. If the matching rate between the predicted driving trajectory and the actual driving trajectory is not less than the first threshold, the trajectory prediction model that has completed this training is determined as the trajectory prediction model to be ultimately used.

[0065] This embodiment of the present invention proposes a trajectory prediction model training process. This process trains an original trajectory prediction model based on historical environmental information about road objects acquired over a historical time period to produce a trajectory prediction model. Because this trajectory prediction model can predict the trajectory of road objects based on this environmental information, it can be applied to a variety of road scenarios, extending its scope of application.

[0066] The following is a further example of a method for controlling a vehicle's driving state based on a predicted driving trajectory of a road object determined by a trajectory prediction model provided by an embodiment of the present invention. The method may include the following steps:

[0067] The first step is to determine the predicted driving trajectory of the road object based on the trained trajectory prediction model.

[0068] In this embodiment, the control device may obtain a set of environmental information about road objects within the vehicle's information collection range during the current time period and process this set of environmental information using a target processing algorithm to generate a feature vector that satisfies the trajectory prediction model input. After determining the feature vector corresponding to the set of environmental information about the road objects during the current time period, the control device may input the feature vector into the trajectory prediction model for prediction. The trajectory prediction model may then output a predicted trajectory for the road object in the future time period.

[0069] The second step is to control the vehicle's driving state based on the predicted driving trajectory of the road object.

[0070] After determining the predicted driving trajectory of a road object within the vehicle's information collection range, the control device may determine the vehicle's driving state based on the determined predicted driving trajectory of the road object. The control device may determine a driving strategy for the vehicle based on the predicted driving trajectory of the road object and control the vehicle to drive according to the determined driving strategy.

[0071] In this embodiment of the present invention, a process is proposed for determining a vehicle's driving state based on a predicted driving trajectory of a road object determined by a trajectory prediction model. Based on a trained trajectory prediction model, the driving trajectory of a road object within the vehicle's information collection range is predicted, and the vehicle's driving state is then controlled based on the predicted driving trajectory of the road object. Because the trajectory prediction model can predict the driving trajectory of a road object in a future time period based on environmental information about the road object, where this environmental information is acquired in real time, the predicted driving trajectory of the road object based on this environmental information is more accurate. The vehicle's driving strategy is then determined based on the predicted driving trajectory of the road object, thereby improving vehicle driving safety and resolving the technical issue of low vehicle trajectory prediction accuracy.

[0072] Example 3

[0073] According to an embodiment of the present invention, a vehicle control device is further provided. It should be noted that the vehicle control device can be used to execute the vehicle control method in Example 1.

[0074] Figure 3 FIG. 1 is a schematic diagram of a vehicle control device according to an embodiment of the present invention. Figure 3 As shown, the vehicle control device 300 may include: an acquisition module 301 , a prediction module 302 and a control module 303 .

[0075] An acquisition module 301 is configured to acquire environmental information of at least one road object collected during a current time period, wherein the road object is a road object within an information collection range of the vehicle;

[0076] Prediction module 302, configured to input environmental information into a trajectory prediction model to obtain a predicted driving trajectory of the road object in a future time period, wherein the trajectory prediction model is pre-trained based on the original trajectory prediction model based on the historical environmental information and historical driving trajectory information of the road object in a historical time period;

[0077] The control module 303 is used to control the driving state of the vehicle based on the predicted driving trajectory of the road object.

[0078] Optionally, the prediction module 302 may include: an adding unit for adding feature values ​​to the environmental information to obtain a feature vector corresponding to the environmental information; and a processing unit for inputting the feature vector into a trajectory prediction model for processing to obtain a predicted driving trajectory.

[0079] Optionally, the historical time period includes a first historical time period and a second historical time period, wherein the second historical time period is a time period after the first historical time period and adjacent to the first historical time period. The device 300 may include: an input module for inputting environmental information of the road object in the first historical time period into the original trajectory prediction model for prediction, to obtain a predicted driving trajectory of the road object in the second historical time period; an adjustment module for adjusting the parameters of the original trajectory prediction model based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period, to obtain a trajectory prediction model.

[0080] Optionally, the adjustment module may include: a first determination unit, used to determine the actual driving trajectory of the road object in a second historical time period based on the historical driving trajectory information of the road object in the historical time period; a second determination unit: used to determine the matching rate between the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period based on the predicted driving trajectory and the actual driving trajectory; and an adjustment unit: used to adjust the parameters of the original trajectory prediction model based on the matching rate.

[0081] Optionally, the device 300 may include: a processing module, configured to adjust the parameters of the original trajectory prediction model in response to the matching rate between the predicted driving trajectory and the actual driving trajectory being less than a first threshold, increase the number of training rounds of the original trajectory prediction model, until the matching rate between the predicted driving trajectory and the actual driving trajectory is not less than the first threshold, and determine the current original trajectory prediction model as the trajectory prediction model.

[0082] Optionally, the control module 303 may include: a third determination unit, used to determine the vehicle's driving strategy based on the predicted driving trajectory of the road object, wherein the driving strategy is used to characterize the vehicle's driving trajectory in a future time period; a control unit, used to control the vehicle's driving state based on the driving strategy, wherein the driving state includes one of the following: left turn, right turn, lane change to the left, lane change to the right, cruising, emergency stop, U-turn and unknown state.

[0083] In this embodiment, the acquisition module is used to acquire environmental information of at least one road object collected in the current time period, wherein the road object is a road object within the information collection range of the vehicle; the prediction module is used to input the environmental information into the trajectory prediction model for prediction to obtain the predicted driving trajectory of the road object in the future time period; the control module is used to control the driving state of the vehicle based on the predicted driving trajectory of the road object. Since the predicted driving trajectory of the road object is predicted based on the environmental information of the road object actually obtained, the predicted driving trajectory is relatively accurate. Controlling the driving state of the vehicle based on the predicted driving trajectory of the road object can improve the safety of vehicle driving and solve the technical problem of low accuracy in vehicle driving trajectory prediction.

[0084] Example 4

[0085] According to an embodiment of the present invention, a computer-readable storage medium is further provided. The storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the vehicle control method in Example 1.

[0086] Example 5

[0087] According to an embodiment of the present invention, a processor is further provided, which is used to run a program, wherein the vehicle control method in embodiment 1 is executed when the program is run.

[0088] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0089] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0090] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of modules can be a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0091] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected to achieve the purpose of the present embodiment according to actual needs.

[0092] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0093] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc. Various media that can store program codes.

[0094] The above are only preferred embodiments of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A vehicle control method, characterized in that: include: Acquiring environmental information of at least one road object collected within a current time period, wherein the road object is a road object within an information collection range of the vehicle; Inputting the environmental information into a trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in a future time period, wherein the trajectory prediction model is obtained by pre-training an original trajectory prediction model based on historical environmental information and historical driving trajectory information of the road object in a historical time period; controlling a driving state of the vehicle based on the predicted driving trajectory of the road object; The historical time period includes a first historical time period and a second historical time period, wherein the second historical time period is a time period after the first historical time period and adjacent to the first historical time period. The method further includes: inputting the environmental information of the road object in the first historical time period into the original trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in the second historical time period; and adjusting the parameters of the original trajectory prediction model based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period to obtain the trajectory prediction model.

2. The method according to claim 1, characterized in that Inputting the environmental information into a trajectory prediction model for prediction to obtain a predicted driving trajectory of the road object in a future time period includes: Adding a characteristic value to the environmental information to obtain a characteristic vector corresponding to the environmental information; The feature vector is input into the trajectory prediction model for processing to obtain the predicted driving trajectory.

3. The method according to claim 1, characterized in that The adjusting the parameters of the original trajectory prediction model based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period includes: determining the actual driving trajectory of the road object in the second historical time period based on the historical driving trajectory information of the road object in the historical time period; determining a matching rate between the predicted driving trajectory and the actual driving trajectory of the road object within the second historical time period; Based on the matching rate, parameters of the original trajectory prediction model are adjusted.

4. The method according to claim 3, characterized in that The method further comprises: In response to the matching rate between the predicted driving trajectory and the actual driving trajectory being less than a first threshold, the parameters of the original trajectory prediction model are adjusted, and the number of training rounds of the original trajectory prediction model is increased until the matching rate between the predicted driving trajectory and the actual driving trajectory is not less than the first threshold, and the current original trajectory prediction model is determined as the trajectory prediction model.

5. The method according to claim 1, wherein The controlling the driving state of the vehicle based on the predicted driving trajectory of the road object includes: determining a driving strategy of the vehicle based on the predicted driving trajectory of the road object, wherein the driving strategy is used to characterize the driving trajectory of the vehicle in the future time period; Based on the driving strategy, the driving state of the vehicle is controlled, wherein the driving state includes one of the following: left turn, right turn, left lane change, right lane change, cruising, emergency stop, U-turn and unknown state.

6. A vehicle control device, characterized in that: include: an acquisition module, configured to acquire environmental information of at least one road object collected within a current time period, wherein the road object is a road object within an information collection range of the vehicle; a prediction module, configured to input the environmental information into a trajectory prediction model for prediction, thereby obtaining a predicted driving trajectory of the road object in a future time period, wherein the trajectory prediction model is obtained by pre-training an original trajectory prediction model based on historical environmental information and historical driving trajectory information of the road object in a historical time period; a control module, configured to control a driving state of the vehicle based on the predicted driving trajectory of the road object; Wherein, the historical time period includes a first historical time period and a second historical time period, wherein the second historical time period is a time period after the first historical time period and adjacent to the first historical time period, and the device is further used to perform the following steps: inputting the environmental information of the road object in the first historical time period into the original trajectory prediction model for prediction to obtain the predicted driving trajectory of the road object in the second historical time period; based on the predicted driving trajectory and the actual driving trajectory of the road object in the second historical time period, adjusting the parameters of the original trajectory prediction model to obtain the trajectory prediction model.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 5.

8. A processor, characterized in that: The processor is configured to run a program, wherein the program executes the method according to any one of claims 1 to 5 when run by the processor.

9. A vehicle, characterized in that: The vehicle is used to perform the method according to any one of claims 1 to 5.

Images