Range extender control method and electronic equipment

Abstract

The invention provides a range extender control method and electronic equipment, and the method comprises the steps: obtaining the road condition data of an advanced driving assistance system in a vehicle and the navigation data of a navigation system, detecting a long-term congestion event and a short-term congestion event in a road in front of the vehicle according to the road condition data and the navigation data, and controlling the range extender according to the detected long-term congestion event and the short-term congestion event. According to the method, the congestion state of a road ahead is determined, operation of a range extender in the vehicle is controlled according to the congestion state and an event detection result, and range extending control based on the road congestion condition is achieved. The specific adjustment strategy of the range extender is determined by combining the congestion state and the event detection result of the long-term and short-term events, frequent start and stop and high-load operation of the range extender in the congestion environment can be avoided, the service life of the range extender is prolonged, and noise and emission of the range extender are reduced.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle control, and particularly relates to a control method for a range extender and an electronic device. Background Art

[0002] The battery endurance problem of pure electric vehicles has become an important factor restricting their current development. As an important auxiliary power source for new energy vehicles, a range extender can convert fuel into electricity when the vehicle's battery power is insufficient, providing additional electric energy for the new energy vehicle.

[0003] However, the current operation strategy of the range extender usually only refers to the current battery power and does not consider the impact of the noise and gas emissions generated by the start of the range extender on the interior and surrounding environment of the vehicle, especially in congested sections of urban traffic. In addition, in a congested environment, the frequent start and stop of the vehicle will cause a large amount of unnecessary energy consumption of the range extender, reducing the overall energy efficiency of the vehicle and accelerating the wear of the range extender, shortening its service life. Summary of the Invention

[0004] In view of the above defects or deficiencies in the prior art, the present application aims to provide a control method for a range extender and an electronic device to reduce the noise impact and gas emission impact generated by the range extender in a congested environment, improve the comfort of passengers, reduce the adverse impact on the surrounding environment and pedestrians, improve the overall environmental quality of urban traffic, and moreover, improve the overall energy efficiency of the vehicle and ensure the service life of the range extender.

[0005] An embodiment of the present application provides a control method for a range extender, the method comprising: Obtaining road condition data detected by an advanced driver assistance system in the vehicle, and obtaining navigation data provided by a navigation system in the vehicle; Based on the road condition data and the navigation data, detecting long-term congestion events and short-term congestion events in the front road of the vehicle, and determining the congestion state of the front road; Based on the congestion state and the event detection result, controlling the operation of the range extender in the vehicle.

[0006] Optionally, detecting long-term congestion events and short-term congestion events in the front road of the vehicle based on the road condition data and the navigation data, comprising: Based on the front road size, the number of surrounding vehicles, the speed of surrounding vehicles, the traffic light state, and road signs in the road condition data, determining whether there is a short-term congestion event in the front road; Based on the event types marked in the navigation data and the corresponding event duration, determining whether there is a long-term congestion event in the front road.

[0007] Optionally, based on the road conditions data including the size of the road ahead, the number of surrounding vehicles, the speed of surrounding vehicles, the traffic light status, and road signs, determining whether there is a short-term congestion event on the road ahead, including: Based on the size of the road ahead and the number of surrounding vehicles, determining the density of surrounding vehicles and the vehicle density of each lane, and based on the speed of surrounding vehicles, determining the speed distribution of surrounding vehicles and the speed distribution of each lane; Constructing a sub-vector for each lane based on the vehicle density and speed distribution of each lane, and constructing a comprehensive vector according to the sub-vectors of each lane, the density of surrounding vehicles, the speed distribution of surrounding vehicles, the traffic light status, and road signs; Based on the comprehensive vector and a pre-trained short-term event detection model, determining whether there is a short-term congestion event on the road ahead.

[0008] Optionally, based on the comprehensive vector and a pre-trained short-term event detection model, determining whether there is a short-term congestion event on the road ahead, including: Inputting the comprehensive vector into the short-term event detection model to identify candidate short-term events on the road ahead and their corresponding short-term event types; Based on the short-term event type, determining the predicted waiting time of the candidate short-term event, and based on the predicted waiting time and the density of surrounding vehicles, determining whether the candidate short-term event is a short-term congestion event.

[0009] Optionally, based on the road conditions data and the navigation data, determining the congestion status of the road ahead, including: Determining a first predicted congestion status based on the road conditions data and a second predicted congestion status based on the navigation data; Fusing the first predicted congestion status and the second predicted congestion status to obtain the congestion status of the road ahead.

[0010] Optionally, fusing the first predicted congestion status and the second predicted congestion status to obtain the congestion status of the road ahead, including: Based on the sensing distance of the advanced driver assistance system and the data update frequency of the navigation system, determining a first weight corresponding to the first predicted congestion status and a second weight corresponding to the second predicted congestion status; Based on the first weight and the second weight, fusing the first predicted congestion status and the second predicted congestion status to obtain the congestion status of the road ahead.

[0011] Optionally, based on the congestion status and the event detection result, controlling the operation of the range extender in the vehicle, including: Construct a current state space corresponding to the vehicle based on the congestion state and the event detection result; Input the current state space into a pre-trained decision agent to obtain a target control action with the maximum reward in the current state space, and control the range extender according to the target control action; Wherein, the reward is used to evaluate the benefit of executing a control action in the current state space.

[0012] Optionally, the reward is calculated by a reward function, and the reward function includes an environmental reward sub-function; The environmental reward sub-function is used to: give a positive reward to a control action of turning off the range extender when the input state space includes a short-term congestion event; and, when the input state space includes a long-term congestion event and the congestion state is less than or equal to a preset congestion level threshold, give a positive reward to a control action of running at a low power; and, when the input state space includes a long-term congestion event and the congestion state is greater than the preset congestion level threshold, give a positive reward to a control action of turning off the range extender; and, give a negative reward to a control action with a power greater than a preset power threshold; and, give a negative reward to a control action with the number of start-stop times greater than a preset number threshold within a set time.

[0013] Optionally, controlling the operation of the range extender in the vehicle based on the congestion state and the event detection result includes: Obtain the remaining power of the vehicle; Construct a current state space corresponding to the vehicle based on the remaining power, the congestion state, and the event detection result; Input the current state space into a pre-trained decision agent to obtain a target control action with the maximum reward in the current state space, and control the range extender according to the target control action; Wherein, the reward is used to evaluate the benefit of executing a control action in the current state space.

[0014] Optionally, the reward function further includes a passenger comfort reward sub-function; The passenger comfort reward sub-function is used to: give a positive reward to a control action of turning off the range extender or running at a low power.

[0015] An embodiment of the present application further provides an electronic device, and the electronic device includes: A processor and a memory; The processor is used to execute the steps of the range extender control method provided in any embodiment of the present application by calling a program or an instruction stored in the memory.

[0016] The embodiments of the present application also provide a computer-readable storage medium, which stores programs or instructions that cause a computer to execute the steps of the range extender control method provided in any embodiment of the present application.

[0017] In summary, the present application proposes a range extender control method. This method obtains the road condition data detected by the advanced driver assistance system in the vehicle and the navigation data provided by the navigation system. Then, based on the road condition data and the navigation data, it detects long-term congestion events and short-term congestion events in the vehicle's forward road, and determines the congestion state of the forward road. According to the congestion state and the event detection results, it controls the operation of the range extender in the vehicle, implementing a range extension control strategy based on road congestion. This method can combine the advanced driver assistance system and the navigation system to accurately determine road congestion, and based on the congestion state and the event detection results of long-term and short-term events, determine the specific adjustment strategy of the range extender. It can avoid the frequent start-stop and high-load operation of the range extender in a congested environment, extend the service life of the range extender, reduce the noise and emissions of the range extender, improve the comfort of the occupants and the overall environmental quality of urban traffic, and also improve the overall energy efficiency of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 is a flowchart of a range extender control method provided by an embodiment of the present application; Figure 2 is a schematic structural diagram of a range extender control device provided by an embodiment of the present application; Figure 3 is a schematic structural diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] The following will further elaborate on the present application in conjunction with the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related invention and are not intended to limit the invention. Additionally, it should be noted that for the sake of description, only parts related to the invention are shown in the drawings.

[0021] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The following will detail the present application with reference to the drawings and embodiments.

[0022] As mentioned in the background art, in view of the problems in the prior art, the present application proposes a range extender control method. Figure 1 FIG. is a flowchart of a range extender control method provided by an embodiment of the present application. The method provided by the embodiment of the present application is applicable to a range-extended electric vehicle, and controls the operation of the range extender by analyzing the congestion situation. This method can be executed by a range extender control device.

[0023] See Figure 1 , the range extender control method specifically includes: S110, obtain the road condition data detected by the advanced driving assistance system in the vehicle, and obtain the navigation data provided by the navigation system in the vehicle.

[0024] Among them, the advanced driving assistance system (Advanced Driving Assistance System, ADS) can integrate a high-precision camera and radar sensors (millimeter-wave radar, lidar, or ultrasonic radar), and can detect road condition data in real time. The road condition data may include the size of the road ahead, the number of surrounding vehicles, the speed of surrounding vehicles, the traffic light status, and road signs, etc.

[0025] Specifically, after obtaining the road condition data detected by the ADS, the road condition data can be preprocessed, and the preprocessing includes at least one of sensor calibration, data cleaning, and data synchronization. Among them, sensor calibration can be the precise calibration of sensors such as cameras, millimeter-wave radars, lidars, and ultrasonic radars; data cleaning can be to remove invalid or incorrect data to improve data quality; data synchronization can be to perform time synchronization on data from different sensors to ensure the consistency and accuracy of the data. For example, a sync signal generator or a global clock source (such as a GPS clock) can be used to ensure that all sensors collect data at the same moment.

[0026] Among them, the navigation system can include a built-in GPS module and a map application, and parameters such as an API (Application Programming Interface) key, a request frequency, and a data format can be preset to ensure that the navigation data obtained from the navigation system meets the requirements of subsequent algorithm processing.

[0027] Specifically, the navigation data may include route planning data (starting point, ending point, waypoints, and recommended routes, etc.), speed distribution of each section, congestion status, estimated travel time, event information (traffic accidents, road construction, or temporary control, etc.), and historical data (traffic flow statistics and average speed in the past period of time).

[0028] Exemplarily, real-time data collection can be performed. For example, a timer is preset to call the API interface of the navigation system every set period (such as 5 s) to obtain the latest navigation data. The obtained navigation data is stored in a local database or cloud storage, and the original data is retained for subsequent analysis. In specific implementation, the frequency of obtaining navigation data can be set according to actual needs. In addition, periodic data collection can also be performed. For example, a timer is preset to batch collect historical data and periodically updated data of the navigation system at regular intervals, and the newly collected data is merged with the existing data to form a complete traffic information database.

[0029] In the embodiment of the present application, for the navigation data obtained from the navigation system, the navigation data can be provided by a third-party platform and the reliability of the data is relatively high. To further ensure the accuracy of the data, the navigation data can also be preprocessed, including data cleaning, data standardization, and data synchronization.

[0030] Among them, data cleaning can include removing invalid data, that is, filtering out obviously incorrect or unreasonable data points from the navigation data, such as abnormally high speed values ( > ), negative distances (d < 0), etc. Data cleaning can also include filling in missing data. The linear interpolation method ( , is the data to be filled, , are two adjacent existing data points, , are respectively , corresponding detection times, can be a value randomly obtained between ~ ) can be used, and the missing time period or road segment information can be filled in combination with historical data. Data cleaning can also include duplicate removal processing. Since the navigation data is collected and read periodically, it may cause duplicate recording of data at a certain moment. Therefore, duplicate data in the navigation data can be removed to ensure the uniqueness and accuracy of the data.

[0031] Among them, data standardization can be to unify all speed units to meters per second (m / s), unify distance units to meters (m), unify time units to (s), and normalize different types of numerical features to ensure comparison on the same scale: ; In the formula, is the original data, , are respectively the maximum or minimum values, It is the result after normalizing the original data.

[0032] Among them, data synchronization can include aligning the timestamps of navigation data with the timestamps of road condition data detected by ADS to avoid misjudgment caused by time differences. The following formula can be used to adjust the timestamps: ; In the formula, is the timestamp of navigation data, is the time difference compensation amount, is the adjusted timestamp.

[0033] Data synchronization can also include converting the longitude and latitude coordinates (lat, lon) in navigation data into coordinates (x, y) in the vehicle's local coordinate system. The following geographic coordinate conversion formula can be used: ; In the formula, R is the radius of the earth, , are the longitude and latitude in the longitude and latitude coordinates respectively, and are the longitude and latitude of the reference point, , are the coordinates in the vehicle's local coordinate system.

[0034] S120. Based on the road condition data and navigation data, detect long-term congestion events and short-term congestion events in the vehicle's front road, and determine the congestion status of the front road.

[0035] Specifically, after obtaining the road condition data of ADS and the navigation data of the navigation system, considering that ADS can detect the traffic status within a certain range around the vehicle, therefore, it is possible to detect whether there are short-term congestion events in the vehicle's front road through the road condition data of ADS. And considering that the navigation system can provide traffic conditions in a relatively far range, therefore, it is possible to detect whether there are long-term congestion events in the vehicle's front road through the navigation data.

[0036] Among them, long-term congestion events can refer to events that cause long-term congestion of vehicles, such as traffic accidents, road construction, temporary control, etc.; short-term congestion events can refer to events that cause short-term congestion of vehicles, such as vehicles stopping in front of traffic lights, temporary parking in commercial areas, temporary parking at school gates, queuing to enter parking garages, etc.

[0037] It should be noted that in the embodiments of the present application, the purpose of detecting long-term congestion events and short-term congestion events in the front road is as follows: The vehicle congestion time caused by short-term congestion events is relatively short. If the operation of the range extender is controlled only based on the congestion state, it may cause the range extender to start and stop frequently in the case of short-term congestion, thereby generating noise and emissions, affecting driving comfort and the surrounding environment. Therefore, the operation of the range extender can be controlled according to the detection results of long-term congestion events and short-term congestion events, combined with the congestion state, to achieve the purpose of precise control of the range extender.

[0038] In a specific implementation manner, based on road condition data and navigation data, the long-term congestion events and short-term congestion events in the front road of the vehicle are detected, including the following steps: Step 11: Based on the front road size, the number of surrounding vehicles, the speed of surrounding vehicles, the traffic light state, and road signs in the road condition data, determine whether there is a short-term congestion event in the front road; Step 12: Based on the event type and the corresponding event duration marked in the navigation data, determine whether there is a long-term congestion event in the front road.

[0039] Among them, the front road size may include the regional length and regional width of the front road. The regional length can be determined by the detection range of the sensors in the ADS, and the regional width can be the road width of the current passable lane of the vehicle. For the traffic light state, one-hot encoding can be used for processing. For example, , where: ; In the formula, T represents the traffic light state, , , are the red light state quantity, yellow light state quantity, and green light state quantity respectively. If it is a red light, , and the remaining state quantities are 0; if it is a yellow light, , and the remaining state quantities are 0; if it is a green light, , and the remaining state quantities are 0.

[0040] Road signs may include ground arrows, speed limit signs, etc., and can be obtained by the camera of the ADS to identify the signs in the sky or on the ground. For example, , where, is the road sign, is the image of the road sign, is the recognition function, and a classification model based on deep learning can be used to implement the detection of road signs.

[0041] In step 11, it is possible to detect whether there is a short-term congestion event in the road ahead based on the size of the road ahead, the number of surrounding vehicles, the speed of surrounding vehicles, the traffic light status, and road signs.

[0042] Regarding step 11 above, in one example, based on the size of the road ahead, the number of surrounding vehicles, the speed of surrounding vehicles, the traffic light status, and road signs in the road condition data, to determine whether there is a short-term congestion event in the road ahead, the following steps are included: Step 111: Based on the size of the road ahead and the number of surrounding vehicles, determine the density of surrounding vehicles and the vehicle density of each lane, and based on the speed of surrounding vehicles, determine the speed distribution of surrounding vehicles and the speed distribution of each lane; Step 112: Construct sub-vectors for each lane based on the vehicle density and speed distribution of each lane, and construct a comprehensive vector according to the sub-vectors of each lane, the density of surrounding vehicles, the speed distribution of surrounding vehicles, the traffic light status, and road signs; Step 113: Based on the comprehensive vector and a pre-trained short-term event detection model, determine whether there is a short-term congestion event in the road ahead.

[0043] Among them, in step 111, the density of surrounding vehicles can be calculated first according to the size of the road ahead and the number of surrounding vehicles, for example: ; In the formula, is the density of surrounding vehicles, N is the number of surrounding vehicles, S is the width of the area, L is the length of the area, represents the area of the region. Similarly, the vehicle density of each lane can also be calculated with reference to this formula.

[0044] To further ensure the accuracy of the density of surrounding vehicles and the vehicle density of each lane, the calculated density of surrounding vehicles and the vehicle density of each lane can also be adjusted based on the vehicle spacing in the road condition data.

[0045] In addition to determining the vehicle density, the speed distribution of surrounding vehicles and the speed distribution of each lane can also be determined according to the speed of surrounding vehicles, so as to identify whether there is a deceleration or stagnation phenomenon. Among them, the speed of surrounding vehicles can be detected by a millimeter-wave radar or can also be determined in combination with the video stream collected by a camera. Among them, the speed distribution can include the average speed and the speed variance.

[0046] Exemplarily, a vehicle speed list can be formed first according to the speed of surrounding vehicles. For example, the vehicle speed list can be expressed as: ; Among them, is the speed of the i-th vehicle (m / s), and N is the number of surrounding vehicles. Based on the vehicle speed list, the average speed of surrounding vehicles and the variance of surrounding vehicle speeds can be calculated to form the speed distribution of surrounding vehicles, and the average speed and speed variance of each lane can be calculated to form the speed distribution of each lane. Exemplarily, the average speed and the speed variance can be calculated respectively using the following formulas: ; To ensure comparison of different features on the same scale, min-max normalization (Min-Max Scaling) can be performed on numerical features: ; where is the original feature value, , are the minimum and maximum values of this feature respectively, is the value of the original feature after normalization. After normalization, the normalized value can also be Z-score standardized: , where , are the mean and standard deviation of this feature respectively, is the value after standardization.

[0047] For road signs, a pre-trained classification model (such as CNN) can be used to extract features and convert the classification results into a numerical representation form. Specifically in implementation, the category index or embedding vector can be selected according to actual needs to represent the road sign category.

[0048] Among them, in step 112, lane sub-vectors of each lane can be constructed first according to the vehicle density and speed distribution of each lane. For example, the lane sub-vector of the i-th lane is represented as , is the vehicle density of the i-th lane, represents the speed distribution of the i-th lane ( is the average speed, is the speed variance). And the lane sub-vectors of all lanes can be concatenated into an overall lane vector, that is, the overall lane vector .

[0049] Furthermore, an integrated vector can be formed according to the overall lane vector, lane sub-vectors of each lane, surrounding vehicle density, surrounding vehicle speed distribution, traffic light status, and road signs, such as: ; where is the integrated vector, They are the normalized density of surrounding vehicles, average speed, and speed variance respectively; represents the speed distribution of surrounding vehicles (i.e., the list of vehicle speeds), is the one-hot encoding of the traffic light state, is the numerical representation form of road signs; is the normalized lane sub-vector.

[0050] After obtaining the comprehensive vector, further, in step 113, the comprehensive vector can be input into a pre-trained short-term event detection model to determine whether there is a short-term congestion event on the road ahead, that is, whether there is a temporary traffic stop or deceleration phenomenon on the road ahead due to specific reasons. Short-term congestion events include, but are not limited to, vehicles stopping in front of traffic lights, temporary parking at the entrance of a business district, temporary parking at the entrance of a school, vehicles queuing up to enter a parking garage, etc.

[0051] Exemplarily, the comprehensive vector can be input into the short-term event detection model, and the short-term event detection model extracts speed distribution features, traffic light states, and landmark detection features based on the comprehensive vector. Among them, the landmark detection feature L can be expressed as: ; Further, after the short-term event detection model extracts the above features, it can detect whether there is a short-term congestion event through the above features.

[0052] Through the above steps 111 - 113, the features of each lane can be analyzed respectively, and then combined with the features of each lane and surrounding features, a comprehensive vector is constructed for the model to detect short-term congestion events, which can ensure the accuracy of event detection.

[0053] In the embodiments of the present application, considering that there may be cases where a long-term congestion is caused by a short-term event, and such cases are not short-term congestion events. Therefore, in order to further improve the detection accuracy of short-term congestion events, it is also possible to further determine whether the event is a short-term congestion event based on the result output by the model and in combination with the predicted waiting time of the event.

[0054] Optionally, based on the comprehensive vector and a pre-trained short-term event detection model, determining whether there is a short-term congestion event on the road ahead includes: Inputting the comprehensive vector into the short-term event detection model to identify candidate short-term events on the road ahead and the corresponding short-term event types; determining the predicted waiting time of the candidate short-term event based on the short-term event type, and determining whether the candidate short-term event is a short-term congestion event based on the predicted waiting time and the density of surrounding vehicles.

[0055] Specifically, first, the comprehensive vector can be input into the short-term event detection model to obtain the detection result output by the model. If the model detects the existence of a short-term congestion event, the model can output it as a candidate short-term event and output the corresponding short-term event type. Further, the predicted waiting time of the candidate short-term event can be determined according to the short-term event type.

[0056] Exemplarily, if the short-term event type is that the vehicle in front stops at a traffic light, it can first be determined whether the traffic light countdown can be extracted from the road condition data detected by the ADS. If the extraction is successful, the predicted waiting time can be determined according to the traffic light countdown and the number of vehicles in front. If the extraction fails, the historical switching time of the same type of traffic light can be collected from the cloud, and according to the current state of the traffic light recognized by the ADS camera and the corresponding switched time, the predicted waiting time can be determined according to the switched time corresponding to the current state and the historical switching time, such as: ; wherein, is the historical switching time (i.e., the start time of the next green light), is the switched time corresponding to the current state, is the predicted waiting time for the vehicle in front to stop at the traffic light.

[0057] If the short-term event type is other events except that the vehicle in front stops at a traffic light, the average waiting time can be calculated according to the historical data of the same event type, and then the average waiting time can be adjusted according to the real-time traffic state perceived by the ADS to obtain the predicted waiting time. Such as: ; In the formula, is the predicted waiting time, is the average waiting time, is the weight of the historical data, is the waiting time estimated according to the real-time traffic state.

[0058] After determining the predicted waiting time of the candidate short-term event according to the short-term event type, further, it can be determined whether the candidate short-term event is a short-term congestion event according to the predicted waiting time and the surrounding vehicle density. For example, if the predicted waiting time is less than the set time threshold and the surrounding vehicle density is less than the set density threshold, it is determined that the candidate short-term event is a short-term congestion event. Such as: If t wait < t threshold and ρ < ρ threshold , it is determined that the candidate short-term event is a short-term congestion event, wheret wait To predict the waiting time, ρ is the density of surrounding vehicles, t threshold is the set time threshold (such as 60s, which can be adjusted according to actual needs), ρ threshold is the set density threshold (such as 0.5 vehicles / meter, which can be adjusted according to actual needs).

[0059] In the above optional implementation, the candidate short-term events and corresponding short-term event types in the front road are identified through the short-term event detection model, and then the predicted waiting time is obtained according to the short-term event type. By combining the predicted waiting time and the density of surrounding vehicles, it is determined whether the candidate short-term event is a short-term congestion event, which further improves the detection accuracy of short-term congestion events, and can avoid misidentifying events such as long-term congestion at traffic lights and long-term congestion at the entrance of commercial areas as short-term congestion events, thereby improving the accuracy of range extender control.

[0060] In addition to detecting whether there is a short-term congestion event in the front road based on the road condition data of ADS, in step 12, it is also possible to detect whether there is a long-term congestion event in the front road according to the navigation data. Specifically, based on the event type marked in the navigation data and the corresponding event duration, it is determined whether there is a long-term congestion event in the front road. Among them, for the traffic light switching time, the current state and the expected switching time of the traffic light can be directly obtained from the navigation data as the event duration. For other events, the expected waiting time can be directly obtained from the navigation data as the event duration. Of course, the event duration can also be updated according to the speed distribution information provided by the navigation system and combined with the actual driving situation.

[0061] Exemplarily, if the event type is a long-term event, such as road construction, major traffic accidents, etc., and the event duration exceeds the preset time threshold, it is determined that the front road is a long-term congestion event.

[0062] Through the above steps 11 - 12, short-term congestion events in the front road can be detected based on road condition data, and long-term congestion events in the front road can be detected based on navigation data. By combining the characteristics of small detection range and high accuracy of ADS, short-term congestion events can be identified, and by combining the characteristics of large detection range of navigation data, long-term congestion events can be identified, ensuring the accuracy of event detection.

[0063] In the embodiments of the present application, in addition to detecting long-term congestion events and short-term congestion events in the road ahead based on road condition data and navigation data, the congestion state of the road ahead can also be determined based on road condition data and navigation data. Among them, the congestion state can be the degree of road congestion described in numerical form, such as the value range is [0, 1], 0 indicates smooth traffic, and 1 indicates severe congestion.

[0064] Exemplarily, the congestion state can be calculated according to the density of surrounding vehicles, the density of vehicles in each lane, the speed distribution of surrounding vehicles, and the speed distribution of each lane. Or, determine the congestion state of the road ahead based on each feature in the comprehensive vector; as shown in the following formula: ; In the formula, is the congestion state determined based on road condition data, , , , , , are the predicted congestion scores corresponding to vehicle density, the predicted congestion scores corresponding to vehicle speed, the predicted congestion scores corresponding to speed variance, the predicted congestion scores corresponding to traffic light status, the predicted congestion scores corresponding to road signs, and the predicted congestion scores corresponding to lane features, respectively.

[0065] Specifically, , is the weight corresponding to vehicle density, is the normalized density of surrounding vehicles; , is the weight corresponding to vehicle speed, is the standardized vehicle speed and the ratio of the maximum speed limit , ; , is the weight corresponding to speed variance, is the normalized speed variance.

[0066] In addition, , is the weight corresponding to traffic light status, is the state quantity corresponding to traffic light status (the aforementioned traffic light status T can be converted into a value within the range of [0, 1], such as 0 indicates red light, 0.5 indicates yellow light, and 1 indicates green light, to obtain the corresponding state quantity); , is the weight corresponding to road signs, is the factor corresponding to road signs, , is the function, It is possible to return corresponding factors according to different road signs ; , is the weight corresponding to the lane feature, is the result of averaging the vehicle density, or the speed, or the speed variance in the normalized lane sub-vector Li′ of all lanes.

[0067] In addition to determining the congestion status based on the road condition data of ADS, it is also possible to determine the congestion status of the road ahead according to the navigation data. For example, according to whether there is congestion and the degree of congestion (mild congestion, moderate congestion, severe congestion) provided in the navigation data, the degree of congestion is converted into a numerical value (i.e., a value within the range of [0,1]) to obtain the congestion status. Exemplarily, the navigation data can be represented by a vector: ; In the formula, is the vector represented by the navigation data, are the starting point and the ending point respectively, represents the set of waypoints, is the recommended route, respectively represent the set of speed distributions of each road segment, the set of congestion degrees of each road segment, and the set of estimated travel times of each road segment, is the traffic accident identifier (e.g., 0 represents no traffic accident, 1 represents a traffic accident), respectively represent the normalized road construction identifier (e.g., 0 represents no road construction, 1 represents road construction), the temporary control identifier (e.g., 0 represents no temporary control, 1 represents temporary control), the traffic flow statistics and the average speed over a past period of time.

[0068] In the embodiments of the present application, in order to improve the accuracy of the congestion status, it is possible to determine the congestion status by integrating the road condition data and the navigation data to fuse multi-source data to obtain a more accurate congestion situation.

[0069] In a specific implementation manner, based on the road condition data and the navigation data, determining the congestion status of the road ahead includes the following steps: Step 21: Determine the first predicted congestion status based on the road condition data and determine the second predicted congestion status based on the navigation data; Step 22: Fuse the first predicted congestion status and the second predicted congestion status to obtain the congestion status of the road ahead.

[0070] Among them, in step 21, the first predicted congestion status can be calculated according to the surrounding vehicle density, the vehicle density of each lane, the surrounding vehicle speed distribution, and the speed distribution of each lane; or, the first predicted congestion status of the road ahead can be determined based on each feature in the comprehensive vector.

[0071] Moreover, according to whether there is congestion and the degree of congestion provided in the navigation data, the second predicted congestion state can be determined. For example, the degree of congestion (slight congestion, moderate congestion, severe congestion) is converted into a value within the range of [0, 1] to obtain the second predicted congestion state.

[0072] Furthermore, in step 22, the first predicted congestion state and the second predicted congestion state can be fused to obtain the congestion state of the road ahead.

[0073] Through the above steps 21 - 22, the congestion state can be predicted respectively through the road condition data and the navigation data, and then the prediction results of the road condition data and the navigation data are fused to obtain a more accurate congestion state, which can further ensure the accuracy of the range - extender control.

[0074] Among them, for the fusion of the first predicted congestion state and the second predicted congestion state, the two can be fused based on a preset weight. In the embodiments of the present application, in order to further improve the accuracy of the congestion state, the weight can also be dynamically calculated according to the real - time perception range of the ADS or the data update frequency of the navigation system, and then fused.

[0075] Regarding the above step 22, in one example, fusing the first predicted congestion state and the second predicted congestion state to obtain the congestion state of the road ahead includes the following steps: Step 221: Determine the first weight corresponding to the first predicted congestion state and the second weight corresponding to the second predicted congestion state based on the perception distance of the advanced driving assistance system and the data update frequency of the navigation system; Step 222: Based on the first weight and the second weight, fuse the first predicted congestion state and the second predicted congestion state to obtain the congestion state of the road ahead.

[0076] Among them, in step 221, considering that the smaller the perception distance of the ADS, the more accurate the detected data, and the more frequent the data update of the navigation system, the more accurate the detected data. Therefore, the distance allocation weight between the ADS and the navigation system can be first determined according to the perception distance of the ADS; then, according to the confidence level of the navigation system, the confidence allocation weight between the ADS and the navigation system is determined; finally, the distance allocation weight and the confidence allocation weight are fused to obtain the first weight and the second weight.

[0077] For example, the closer the perception distance of the ADS is, the greater the distance allocation weight of the ADS, and then the distance allocation weight of the navigation system is obtained. For example, when the perception distance is a short distance (such as 0 - 500 meters), the distance allocation weights of the ADS and the navigation system are 0.8 and 0.2 respectively; when the perception distance is a medium distance (such as 500 - 2000 meters), the distance allocation weights of the ADS and the navigation system are 0.5 and 0.5 respectively; when the perception distance is a long distance (such as 2000 meters), the distance allocation weights of the ADS and the navigation system are 0.2 and 0.8 respectively.

[0078] The higher the data update frequency of the navigation system is, the higher the confidence level of the navigation system is, the greater the confidence allocation weight of the navigation system is, and then the confidence allocation weight of the ADS is obtained. For example, when the data update frequency of the navigation system is greater than the set frequency threshold and the navigation data is accurate / road condition data is fuzzy, the confidence allocation weight of the navigation system is increased. For example, the confidence allocation weights of the ADS and the navigation system are 0.4 and 0.6 respectively; when the data update frequency of the navigation system is less than the set frequency threshold, or there are errors in the navigation data / road condition data is clear and consistent, the confidence allocation weight of the ADS is increased. For example, the confidence allocation weights of the ADS and the navigation system are 0.6 and 0.4 respectively.

[0079] After obtaining the confidence allocation weights and distance allocation weights of the ADS and the navigation system, further, the confidence allocation weight and distance allocation weight of the ADS can be fused to obtain the first weight, and the confidence allocation weight and distance allocation weight of the navigation system can be fused to obtain the second weight. It should be noted that the sum of the first weight and the second weight is 1, and the first weight and the second weight can be obtained through normalization processing, as shown in the following formula: ; ; In the formula, and are the distance allocation weight and confidence allocation weight of the ADS respectively, and are the distance allocation weight and confidence allocation weight of the navigation system respectively. and are the first weight and the second weight respectively.

[0080] Further, in step 222, the first predicted congestion state and the second predicted congestion state can be fused according to the first weight and the second weight to obtain the congestion state of the road ahead. As shown in the following formula: ; In the formula, and They are the first predicted congestion state and the second predicted congestion state (i.e., the predicted congestion state determined based on the congestion level provided by the navigation data), is the fused congestion state, 、 are the first weight and the second weight respectively.

[0081] Through the above steps 221 - step 222, it is possible to fuse the congestion states obtained from the ADS and the navigation system in real time according to the sensing distance of the advanced driver assistance system and the data update frequency of the navigation system, further ensuring the accuracy of the analysis of road congestion conditions, and thus further ensuring the reliability of the range extender control.

[0082] S130. Based on the congestion state and the event detection result, control the operation of the range extender in the vehicle.

[0083] Among them, the event detection result can be no event, including short - term congestion events, including long - term congestion events, or including short - term congestion events and long - term congestion events. For example, if the event detection result contains 1, it means there is a short - term congestion event; if it contains 0, it means there is a long - term congestion event; and if it is an empty set, it means there is no event.

[0084] Specifically, after obtaining the congestion state and the event detection result, a control instruction for the range extender can be generated according to the event detection result and the congestion state, so as to control the operation of the range extender. Among them, the control instruction can be a signal for controlling the start and stop of the range extender (such as ADS_REStartStop, 0 means off, 1 means on), or a signal for adjusting the power of the range extender (such as ADS_REPowerAdjust, with a value range of 0 - 1, representing the power percentage).

[0085] Exemplarily, in the case where the event detection result contains short - term congestion events, the range extender can be controlled to turn off; in the case where the event detection result contains long - term congestion events, if the congestion state is less than or equal to the preset congestion level threshold, the range extender can be controlled to operate at a low power to minimize the impact of noise and emissions on the surrounding environment; if the congestion state is greater than the preset congestion level threshold, the range extender can be controlled to turn off to minimize the impact of noise and emissions to the greatest extent.

[0086] In the embodiments of the present application, an intelligent agent can also be pre - trained through a reward function. The congestion state and the event detection result are input into the intelligent agent, and the corresponding control strategy is output through the intelligent agent. Among them, the intelligent agent can be understood as a decision - making model obtained through reinforcement learning.

[0087] In a specific implementation manner, controlling the operation of the range extender in the vehicle based on the congestion state and the event detection result includes: Based on the congestion state and the event detection result, construct the current state space corresponding to the vehicle; input the current state space into a pre-trained decision-making agent to obtain the target control action with the maximum reward in the current state space, and control the range extender according to the target control action.

[0088] Among them, the reward is used to evaluate the benefit of executing the control action in the current state space; the control action can include the on / off control action of the range extender and the power control action when the range extender is turned on.

[0089] In one example, the reward is calculated by a reward function, and the reward function includes an environmental reward sub-function; the environmental reward sub-function is used to: give a positive reward to the control action of turning off the range extender when the input state space contains a short-term congestion event; and, when the input state space contains a long-term congestion event and the congestion state is less than or equal to a preset congestion level threshold, give a positive reward to the control action of operating at low power (range extender turned on); and, when the input state space contains a long-term congestion event and the congestion state is greater than the preset congestion level threshold, give a positive reward to the control action of turning off the range extender; and, give a negative reward to the control action with a power greater than a preset power threshold (range extender turned on); and, give a negative reward to the control action with the number of start / stop times greater than a preset number threshold within a set time.

[0090] Specifically, a sample state space and a reward function can be pre-constructed, and the reward function is used to evaluate the quality of executing each control action in the input state space. In order to minimize the emissions and noise of the range extender under congestion conditions, the reward function can include an environmental reward sub-function. By inputting the sample state space into the agent, the agent then calculates the rewards corresponding to each control action according to the reward function and adjusts its internal parameters, repeating this process until the iteration stop condition is reached to complete the training of the decision-making agent.

[0091] Among them, the environmental reward sub-function can encourage the agent to turn off the range extender under short-term congestion events; for example, give a positive reward when the range extender is turned off under short-term congestion events : ; In the formula, is the weight coefficient, controlling the intensity of this reward term, is the start / stop signal of the range extender, which can be a binary variable (0 means off, 1 means on); is the positive reward obtained by calculation when the range extender is turned off in the case of a short-term congestion event.

[0092] The environmental reward sub - function can also encourage the agent to control the range extender in combination with the congestion level during long - term congestion events. For example, during long - term congestion events, when the congestion state is less than or equal to the preset congestion level threshold, the agent is encouraged to keep the range extender running at low power, and a positive reward is given for the low - power operation of the range extender. : ; In the formula, is the weight coefficient, which controls the intensity of this reward item. is the power adjustment signal of the range extender, and its range is [0, 1]; is the positive reward obtained by calculation when the range extender runs at low power in the case of long - term congestion events and the congestion state does not exceed the preset congestion level threshold.

[0093] The environmental reward sub - function can also encourage the agent to turn off the range extender when the congestion state is greater than the preset congestion level threshold during long - term congestion events, and a positive reward is given for turning off the range extender. : ; In the formula, is the weight coefficient, which controls the intensity of this reward item. is the power adjustment signal of the range extender, and its range is [0, 1]; is the positive reward obtained by calculation when the range extender is turned off in the case of long - term congestion events and the congestion state is greater than the preset congestion level threshold.

[0094] The environmental reward sub - function can also punish the agent for frequently starting and stopping the range extender, or punish the range extender for running at too high a power. For example, a negative reward is given for control actions with the number of start - stop times greater than the preset number threshold within a set time. And a negative reward is given for control actions with a power greater than the preset power threshold. : ; ; In the formula, 、 are weight coefficients, which respectively control the intensities of the start - stop times reward item and the power reward item; is the start or stop signal of the range extender at time t. is the start or stop signal of the range extender at time t - 1, and T is the number of detection periods included in the set time. can reflect the number of start - stop times of the range extender within the set time; is the power adjustment signal of the range extender, and its range is [0, 1]. is the preset power threshold, and its range is [0, 1]. It can reflect the degree to which the operating power of the range extender exceeds the preset power threshold; is the negative reward obtained by calculation if the number of start-stop times within the set time is greater than the preset number threshold, is the negative reward obtained by calculation if the operating power of the range extender is greater than the preset power threshold.

[0095] By adding up each reward item of the above environmental reward sub-function (which can be combined with preset weights), the environmental reward can be obtained. Furthermore, the decision-making agent can adjust its internal parameters based on the rewards corresponding to each control action to find the control actions that can optimize the environmental rewards in each state space and complete the learning. In other words, the learning and training process of the decision-making agent is a process of obtaining higher rewards for the action space.

[0096] After the decision-making agent completes the training, the congestion state and event detection results can be used as the current state space and input into the decision-making agent to obtain the target control actions that can optimize the rewards in the current state space, such as turning off the range extender, starting the range extender, adjusting the operating power of the range extender, etc., so as to control the range extender according to the target control actions.

[0097] The decision-making agent trained by the environmental reward sub-function determines the control strategy of the range extender, which can try to keep the range extender turned off during short-term congestion events. During long-term congestion events, if the severity of congestion is high, try to keep the range extender turned off; if the severity of congestion is low, try to keep the range extender running at a low power. Moreover, it can also try to avoid the range extender running at too high a power and frequent start-stop, and can greatly reduce the vibration, emissions and noise caused by the operation of the range extender in the congested environment, improve the comfort of the occupants and reduce the impact on the environment.

[0098] In the embodiment of the present application, in addition to controlling the range extender according to the congestion situation of the environment, in order to ensure the stable driving of the vehicle, the control strategy of the range extender can also be determined according to the remaining power of the battery pack.

[0099] In a specific implementation manner, based on the congestion state and event detection results, controlling the operation of the range extender in the vehicle includes: Obtain the remaining power of the vehicle; based on the remaining power, congestion state and event detection results, construct the current state space corresponding to the vehicle; Input the current state space into the pre-trained decision-making agent to obtain the target control action with the maximum reward in the current state space, and control the range extender according to the target control action; wherein, the reward is used to evaluate the benefit of executing the control action in the current state space; In the above embodiments, a current state space including the remaining battery power, the congestion state, and the event detection result can be constructed, so that the input of the decision-making agent includes not only the congestion state and the event detection result, but also the remaining battery power of the vehicle. Further, the input is sent to the decision-making agent to obtain the target control action with the maximum reward in the current state space.

[0100] Among them, the reward function can include a power management reward sub-function in addition to the environmental reward sub-function. Correspondingly, the reward function required for training the decision-making agent can include a power management reward sub-function in addition to the environmental reward sub-function.

[0101] During the process of training the agent, by inputting the sample state space into the agent, the agent then calculates the rewards corresponding to each control action according to the reward function (including the power management reward and the environmental management reward), and adjusts the internal parameters according to the rewards. Repeat this process until the iteration stop condition is reached to complete the training of the decision-making agent.

[0102] Among them, the power management reward sub-function is used to: give a positive reward to the control action of starting the range extender when the remaining battery power in the input state space is lower than the preset battery power threshold; and, give a positive reward to the control action of turning off the range extender or operating at low power when the remaining battery power in the input state space is greater than or equal to the preset battery power threshold.

[0103] Specifically, the power management reward sub-function can encourage the agent to keep the range extender turned off or operating at low power when the remaining battery power is sufficient. For example, when the remaining battery power is greater than or equal to the preset battery power threshold, a positive reward is given to the range extender being turned off or operating at low power. The magnitude of this reward can be positively correlated with the magnitude of the remaining battery power. The greater the difference between the remaining battery power and the preset battery power threshold, the greater the reward.

[0104] The power management reward sub-function can also encourage the agent to start the range extender to supplement power when the remaining battery power is insufficient. For example, when the remaining battery power is lower than the preset battery power threshold, a negative reward is given, and a positive reward is given to starting the range extender. The magnitude of this reward can be positively correlated with the magnitude of the remaining battery power. The greater the difference between the preset battery power threshold and the remaining battery power, the greater the reward.

[0105] Add the above environmental reward and power management reward (which can be combined with a preset weight), and then the decision-making agent can adjust the internal parameters based on the rewards corresponding to each control action to find the control action that can optimize the environmental reward in each state space and complete the learning.

[0106] After the decision-making agent is trained, the congestion state, event detection results, and remaining power can be used as the current state space and input into the decision-making agent to obtain the target control action that can optimize the reward in this current state space, thereby controlling the range extender according to the target control action.

[0107] The decision-making agent trained through the environmental reward sub-function and the power management reward sub-function determines the control strategy of the range extender and can also try to start the range extender when the power is insufficient and keep it running at a low power or turn it off when the power is sufficient, thereby ensuring that the vehicle has sufficient energy as much as possible.

[0108] In the embodiments of the present application, in addition to the environmental reward and the power management reward, a passenger comfort reward can also be set to improve the comfort of the occupants.

[0109] Optionally, the reward function further includes a passenger comfort reward sub-function; the passenger comfort reward sub-function is used to: give a positive reward for the control action of turning off the range extender or running at a low power.

[0110] Specifically, the passenger comfort reward sub-function can encourage the agent to try to keep the range extender turned off or running at a low power. For example, a positive reward is given when the range extender is turned off or running at a low power. : ; In the formula, is the weight coefficient, which controls the intensity of this reward item. is the power adjustment signal of the range extender, and the range is [0,1]; is the positive reward obtained by calculation if the range extender is turned off or running at a low power.

[0111] Add the above rewards (the preset weights can be combined). Then, the decision-making agent can adjust the internal parameters based on the rewards corresponding to each control action to find the control action that can optimize the environmental reward in each state space and complete the learning. After the decision-making agent is trained, the current state space can be input into the decision-making agent to obtain the target control action that can optimize the reward in this current state space, thereby controlling the range extender according to the target control action.

[0112] The decision-making agent trained through the environmental reward sub-function and the passenger comfort reward sub-function determines the control strategy of the range extender and can also try to keep the range extender turned off or running at a low power to reduce vibration and noise and improve the comfort of the passengers.

[0113] In the embodiments of the present application, by setting the above reward function, the agent can learn how to select the optimal control actions in different state spaces. For example, when the battery power is sufficient, the agent will preferentially choose to turn off the range extender or operate it at low power to reduce noise and emissions; when the battery power is insufficient, the agent will start the range extender in time to supplement the power to ensure the endurance; when there is a long-term congestion event and light to moderate congestion, the agent will maintain low-power operation, and when there is a long-term congestion event and severe congestion, the range extender will be turned off. The agent will also try to avoid frequent start-stop of the range extender or setting too high power to improve the passenger comfort and extend the equipment life. The design of these reward functions ensures that the agent can make reasonable decisions in a complex and changeable driving environment, so as to achieve the best use effect of the range extender.

[0114] The range extender control method provided by the embodiments of the present application obtains the road condition data detected by the advanced driver assistance system in the vehicle, and obtains the navigation data provided by the navigation system. Then, according to the road condition data and the navigation data, it detects the long-term congestion events and short-term congestion events in the front road of the vehicle, and determines the congestion state of the front road. According to the congestion state and the event detection results, it controls the operation of the range extender in the vehicle to implement the range extension control strategy based on the road congestion situation. This method can combine the advanced driver assistance system and the navigation system to accurately determine the road congestion situation, and combine the congestion state and the event detection results of the long-term and short-term events to determine the specific adjustment strategy of the range extender. It can avoid the frequent start-stop and high-load operation of the range extender in a congested environment, extend the service life of the range extender, reduce the noise and emissions of the range extender, improve the passenger comfort and the overall environmental quality of urban traffic, and can also improve the overall energy efficiency of the vehicle and achieve the optimization of energy consumption and emissions.

[0115] Figure 2 It is a schematic structural diagram of a range extender control device provided by the embodiments of the present application. The device includes a data acquisition module 210, a congestion determination module 220, and a control module 230, where: The data acquisition module 210 is configured to obtain the road condition data detected by the advanced driver assistance system in the vehicle, and obtain the navigation data provided by the navigation system in the vehicle; The congestion determination module 220 is configured to detect long-term congestion events and short-term congestion events in the front road of the vehicle based on the road condition data and the navigation data, and determine the congestion state of the front road; The control module 230 is configured to control the operation of the range extender in the vehicle based on the congestion state and the event detection results.

[0116] On the basis of the above embodiments, optionally, the congestion determination module 220 includes a short-term event detection unit and a long-term event detection unit, where: A short-term event detection unit, configured to determine whether there is a short-term congestion event in the forward road based on the forward road size, the number of surrounding vehicles, the speed of surrounding vehicles, the traffic light status, and road signs in the road condition data; A long-term event detection unit, configured to determine whether there is a long-term congestion event in the forward road based on the event type marked in the navigation data and the corresponding event duration.

[0117] Based on the above embodiments, optionally, the short-term event detection unit is specifically configured to: Based on the forward road size and the number of surrounding vehicles, determine the surrounding vehicle density and the vehicle density of each lane, and based on the speed of the surrounding vehicles, determine the surrounding vehicle speed distribution and the speed distribution of each lane; construct a sub-vector for each lane based on the vehicle density and speed distribution of each lane, and construct a comprehensive vector according to the sub-vectors of each lane, the surrounding vehicle density, the surrounding vehicle speed distribution, the traffic light status, and the road signs; based on the comprehensive vector and a pre-trained short-term event detection model, determine whether there is a short-term congestion event in the forward road.

[0118] Based on the above embodiments, optionally, the short-term event detection unit is further configured to input the comprehensive vector into the short-term event detection model to identify candidate short-term events and corresponding short-term event types in the forward road; determine the predicted waiting time of the candidate short-term event based on the short-term event type, and determine whether the candidate short-term event is a short-term congestion event based on the predicted waiting time and the surrounding vehicle density.

[0119] Based on the above embodiments, optionally, the congestion determination module 220 includes a status analysis unit, and the status analysis unit is configured to determine a first predicted congestion status based on the road condition data and a second predicted congestion status based on the navigation data; fuse the first predicted congestion status and the second predicted congestion status to obtain the congestion status of the forward road.

[0120] Based on the above embodiments, optionally, the status analysis unit is further configured to determine a first weight corresponding to the first predicted congestion status and a second weight corresponding to the second predicted congestion status based on the perception distance of the advanced driver assistance system and the data update frequency of the navigation system; fuse the first predicted congestion status and the second predicted congestion status based on the first weight and the second weight to obtain the congestion status of the forward road.

[0121] Based on the above embodiments, optionally, the control module 230 is specifically configured to construct a current state space corresponding to the vehicle based on the congestion state and the event detection result; input the current state space into a pre-trained decision agent to obtain a target control action with the maximum reward in the current state space, and control the range extender according to the target control action; wherein, the reward is used to evaluate the benefit of executing the control action in the current state space.

[0122] Based on the above embodiments, optionally, the reward is calculated by a reward function, and the reward function includes an environmental reward sub-function; the environmental reward sub-function is used for: giving a positive reward to the control action of turning off the range extender when the input state space includes a short-term congestion event; and, giving a positive reward to the control action of running at a low power when the input state space includes a long-term congestion event and the congestion state is less than or equal to a preset congestion level threshold; and, giving a positive reward to the control action of turning off the range extender when the input state space includes a long-term congestion event and the congestion state is greater than the preset congestion level threshold; and, giving a negative reward to the control action with a power greater than a preset power threshold; and, giving a negative reward to the control action with the number of start-stop times greater than a preset number threshold within a set time.

[0123] Based on the above embodiments, optionally, the control module 230 is specifically configured to obtain the remaining power of the vehicle; construct a current state space corresponding to the vehicle based on the remaining power, the congestion state, and the event detection result; input the current state space into a pre-trained decision agent to obtain a target control action with the maximum reward in the current state space, and control the range extender according to the target control action; wherein, the reward is used to evaluate the benefit of executing the control action in the current state space.

[0124] Based on the above embodiments, optionally, the reward function further includes a passenger comfort reward sub-function; the passenger comfort reward sub-function is used for: giving a positive reward to the control action of turning off the range extender or running at a low power.

[0125] The battery balancing device provided in the embodiments of the present application can execute the steps in the range extender control method provided in the method embodiments of the present application, and the implementation steps and beneficial effects are not described herein again.

[0126] Figure 3 It is a schematic structural diagram of an electronic device provided in the embodiments of the present application. As Figure 3 shown, the electronic device 400 includes one or more processors 401 and a memory 402.

[0127] The processor 401 can be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and can control other components in the electronic device 400 to perform desired functions.

[0128] The memory 402 can include one or more computer program products, and the computer program products can include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory can include, for example, random access memory (RAM) and / or cache memory, etc. The non-volatile memory can include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions can be stored on the computer-readable storage media, and the processor 401 can run the program instructions to implement the range extender control method of any embodiment of the present application described above and / or other desired functions. Various contents such as initial external parameters, thresholds, etc. can also be stored in the computer-readable storage media.

[0129] In one example, the electronic device 400 can further include: an input device 403 and an output device 404, and these components are interconnected through a bus system and / or other forms of connection mechanisms (not shown). The input device 403 can include, for example, a keyboard, a mouse, etc. The output device 404 can output various information to the outside, including warning prompt information, braking force, etc. The output device 404 can include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0130] Of course, for simplicity, Figure 3 only some of the components related to the present application in the electronic device 400 are shown, and components such as buses, input / output interfaces, etc. are omitted. In addition, according to specific application scenarios, the electronic device 400 can further include any other appropriate components.

[0131] In addition to the above methods and devices, the embodiments of the present application can also be computer program products, which include computer program instructions, and when the computer program instructions are run by a processor, the processor is caused to execute the steps of the range extender control method provided by any embodiment of the present application.

[0132] The computer program product can be written in any combination of one or more programming languages for programming code to perform the operations of the embodiments of the present application. The programming languages include object-oriented programming languages such as Java, C++, etc., and also include conventional procedural programming languages such as the "C" language or similar programming languages. The programming code can be executed entirely on the user's computing device, partially on the user's device, executed as a stand-alone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0133] In addition, an embodiment of the present application can also be a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are run by a processor, the processor is caused to execute the steps of the range extender control method provided by any embodiment of the present application.

[0134] The computer-readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, include but is not limited to an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0135] It should be noted that the terms used in the present application are only for describing specific embodiments and do not limit the scope of the present application. As shown in the specification and claims of the present application, unless the context clearly indicates otherwise, words such as "a", "an", "one", and / or "the" are not specifically singular and may also include the plural. The term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, or device. Without further limitation, an element defined by the statement "including a..." does not exclude the presence of additional identical elements in the process, method, or device including the element.

[0136] It should also be noted that the orientation or positional relationship indicated by terms such as "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present application. Unless otherwise clearly specified and defined, terms such as "installed", "connected", "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0137] In this article, specific examples are used to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only for helping to understand the method and its core idea of the present application. The above is only the preferred implementation manner of the present application. It should be noted that due to the limitation of literal expression, and objectively there are infinite specific structures. For those of ordinary skill in the technical field, without departing from the principle of the present application, several improvements, retouches or changes can also be made, or the above technical features can be combined in an appropriate manner; these improvements, retouches, changes or combinations, or directly applying the concept and technical solution of the invention to other occasions without improvement, should all be regarded as the protection scope of the present application.

Claims

1. A range extender control method, characterized in that: include: Acquiring road condition data detected by an advanced driver assistance system in a vehicle, and acquiring navigation data provided by a navigation system in the vehicle; Based on the road condition data and the navigation data, detecting long-term congestion events and short-term congestion events in the road ahead of the vehicle, and determining the congestion state of the road ahead; Based on the congestion status and event detection results, the operation of the range extender in the vehicle is controlled.

2. The range extender control method according to claim 1, characterized in that: Based on the road condition data and the navigation data, detecting a long-term congestion event and a short-term congestion event on the road ahead of the vehicle includes: Based on the road size ahead, the number of surrounding vehicles, the speed of surrounding vehicles, the status of traffic lights, and road signs in the road condition data, determining whether there is a short-term congestion event on the road ahead; Based on the event type marked in the navigation data and the corresponding event duration, it is determined whether there is a long-term congestion event in the road ahead.

3. The range extender control method according to claim 2, characterized in that: Based on the road size ahead, the number of surrounding vehicles, the speed of surrounding vehicles, the status of traffic lights, and road signs in the road condition data, determining whether there is a short-term congestion event on the road ahead, including: Determine the surrounding vehicle density and the vehicle density of each lane based on the size of the road ahead and the number of surrounding vehicles, and determine the surrounding vehicle speed distribution and the speed distribution of each lane based on the surrounding vehicle speed; Constructing each lane sub-vector based on the vehicle density and speed distribution of each lane, and constructing a comprehensive vector based on each lane sub-vector, the surrounding vehicle density, the surrounding vehicle speed distribution, the traffic light state and the road sign; Based on the comprehensive vector and a pre-trained short-term event detection model, it is determined whether there is a short-term congestion event in the road ahead.

4. The range extender control method according to claim 3, characterized in that: Based on the comprehensive vector and the pre-trained short-term event detection model, determining whether there is a short-term congestion event in the road ahead includes: Inputting the comprehensive vector into the short-term event detection model to identify candidate short-term events and corresponding short-term event types in the road ahead; The predicted waiting time of the candidate short-term event is determined based on the short-term event type, and whether the candidate short-term event is a short-term congestion event is determined based on the predicted waiting time and the surrounding vehicle density.

5. The range extender control method according to claim 1, characterized in that: Determining the congestion state of the road ahead based on the road condition data and the navigation data includes: determining a first predicted congestion state based on the traffic data, and determining a second predicted congestion state based on the navigation data; The first predicted congestion state and the second predicted congestion state are merged to obtain the congestion state of the road ahead.

6. The range extender control method according to claim 5, characterized in that: The first predicted congestion state and the second predicted congestion state are merged to obtain the congestion state of the road ahead, including: Determining, based on the perception distance of the advanced driver assistance system and the data update frequency of the navigation system, a first weight corresponding to the first predicted congestion state and a second weight corresponding to the second predicted congestion state; Based on the first weight and the second weight, the first predicted congestion state and the second predicted congestion state are fused to obtain the congestion state of the road ahead.

7. The range extender control method according to claim 1, characterized in that: Based on the congestion state and the event detection result, controlling the operation of the range extender in the vehicle includes: Based on the congestion state and the event detection result, construct a current state space corresponding to the vehicle; Inputting the current state space into a pre-trained decision-making agent to obtain a target control action with the maximum reward under the current state space, and controlling the range extender according to the target control action; The reward is used to evaluate the benefit of executing the control action in the current state space.

8. The range extender control method according to claim 7, characterized in that: The reward is calculated by a reward function, wherein the reward function includes an environment reward sub-function; The environmental reward subfunction is used to: when the input state space includes a short-term congestion event, give a positive reward for the control action of shutting down the range extender; and, when the input state space includes a long-term congestion event and the congestion state is less than or equal to a preset congestion level threshold, give a positive reward for the control action of low-power operation; and, when the input state space includes a long-term congestion event and the congestion state is greater than the preset congestion level threshold, give a positive reward for the control action of shutting down the range extender; Also, a negative reward is given for a control action whose power is greater than a preset power threshold; and a negative reward is given for a control action whose start and stop times within a set time are greater than a preset number threshold.

9. The range extender control method according to claim 1, characterized in that: Based on the congestion state and the event detection result, controlling the operation of the range extender in the vehicle includes: Obtaining the remaining power of the vehicle; Based on the remaining power, the congestion state and the event detection result, construct a current state space corresponding to the vehicle; Inputting the current state space into a pre-trained decision-making agent to obtain a target control action with the maximum reward under the current state space, and controlling the range extender according to the target control action; The reward is used to evaluate the benefit of executing the control action in the current state space.

10. The range extender control method according to claim 7, characterized in that: The reward function also includes a passenger comfort reward sub-function; The passenger comfort reward sub-function is used to give a positive reward for a control action of shutting down the range extender or operating at low power.

11. An electronic device, characterized in that: The electronic device comprises: Processor and memory; The processor is used to execute the steps of the range extender control method according to any one of claims 1 to 10 by calling the program or instruction stored in the memory.

Images