Refueling recommendation device and method based on AI-Agent

Abstract

The invention belongs to the related technical field of automobiles, and discloses an AI-Agent-based refueling recommendation device and method.The refueling recommendation device comprises a vehicle-mounted electronic terminal and a background server, and the vehicle-mounted electronic terminal is provided with a GNSS high-precision positioning module, a long-distance communication module, a voice interaction module, an algorithm module and a front-end database; the background server is provided with a large model algorithm module, a voice recognition module, a voice synthesis module, an API (Application Program Interface) and a front-end data interface; all functional modules of the vehicle-mounted electronic terminal and the background server jointly form the refueling recommendation system of the AI-Agent general architecture. The refueling recommendation system comprises a large language model, a memory unit, a planning unit, a tool unit and an action unit. The method is based on precise space-time and owner demand matching housekeeper type Agent service, is based on a housekeeper type active recommendation and passive query mode favored by a user, and is combined with the driving destination comprehensive route recommendation of this time instead of position recommendation.

Description

Technical Field

[0001] The present invention belongs to the technical field of automobiles, and more specifically, to an AI-Agent-based refueling recommendation device and method. Background Art

[0002] Refueling is a common consumption behavior in daily life. Refueling reminders and recommendations have become business types with relatively rigid demand and high recommendation conversion rate. There are many websites or mobile apps that provide refueling recommendation services, and the recommendation methods and strategies they adopt are also different.

[0003] Patent CN110704757A discloses a method and device for recommending a refueling strategy: generating a navigation route for a vehicle based on a geographic coordinate information group, screening gas stations of preset types along the navigation route based on gas station type information and navigation route information, and selecting a gas station that the vehicle can reach with its current fuel volume and has the lowest price for refueling.

[0004] Patent CN111949897A discloses a gas station recommendation method based on multivariate optimization: extract all gas stations along the driving route to form a gas station preparation set, each gas station includes attribute data of multiple dimensions, determine an alternative set from the gas station preparation set according to a preset method and recommend it to the user, and determine the target gas station according to the user's selection result. The above patent optimizes the selection of gas stations along the route when the driving route is known, which is suitable for long-distance driving application scenarios, but not for daily commuting application scenarios.

[0005] Patent CN111815344A discloses a method, electronic device and storage medium for recommending automobile refueling: obtaining the gas station locations and gas station discount information of multiple gas stations; for each gas station, determining the preferential benefits of refueling the car at the gas station based on the gas station discount information of the gas station, and determining the fuel cost of the car to reach the gas station based on the gas station location of the gas station; ranking the gas stations based on the comprehensive preferential benefits and fuel costs, and making refueling recommendations based on the ranking results.

[0006] Existing technologies basically collect information of one or more dimensions of candidate gas stations and sort them according to certain rules when users actively initiate refueling requests. They lack relevant information about users' vehicles and do not have the ability to make active recommendations. To make active recommendations, it is necessary to include information about gas stations, vehicles, user route habits, and user consumption preferences. It is unrealistic to rely on users to input all this information. A system or device is needed to collect and connect all relevant information.

[0007] To this end, the present invention provides an AI-Agent-based refueling recommendation device and method. Summary of the invention

[0008] In order to overcome the deficiencies of the prior art, the present invention provides a refueling recommendation method and device based on AI-Agent.

[0009] The purpose of the present invention can be achieved by the following technical solutions:

[0010] A refueling recommendation device and method based on AI-Agent, the refueling recommendation device comprising an on-board electronic terminal and a backend server, the on-board electronic terminal being provided with a GNSS high-precision positioning module, a long-distance communication module, a voice interaction module, an algorithm module and a front-end database, the backend server being provided with a large model algorithm module, a voice recognition module, a voice synthesis module, an API interface and a front-end data interface;

[0011] The functional modules of the vehicle-mounted electronic terminal and the backend server together constitute a refueling recommendation system of the AI-Agent general architecture, and the refueling recommendation system includes a large language model, a memory unit, a planning unit, a tool unit, and an action unit;

[0012] The refueling recommendation method is based on the refueling recommendation device and the refueling recommendation system, and specifically includes the following steps:

[0013] S1: Use various modules of the vehicle electronic terminal to collect various data required by the refueling recommendation system;

[0014] S2: Analyze the above data using the algorithm module deployed in the vehicle electronic terminal and the large model algorithm module of the backend server, and initiate a refueling recommendation when the preset conditions are met;

[0015] S3: Calling API interfaces based on various data to query external data required for this refueling recommendation;

[0016] S4: using a large language model deployed on a backend server to synthesize all data using a sorting algorithm to generate an optimal recommendation plan, wherein the optimal recommendation plan is based on the principle of minimizing the total cost of refueling;

[0017] S5: broadcasting the recommended solution in the form of voice through the voice interaction module of the vehicle-mounted electronic terminal;

[0018] S6: Based on user feedback, execute the refueling recommendation plan, or modify the current recommendation plan and optimize the relevant parameters of the refueling recommendation system.

[0019] As a further preferred technical solution of the present invention, the data collected by various modules of the vehicle-mounted electronic terminal include user consumption preferences, user's frequently used locations, user's car gasoline grade, fuel tank capacity, fuel consumption, current cruising range, current location and destination, and current voice interaction data;

[0020] The current location is obtained in real time by the GNSS high-precision positioning module;

[0021] Frequently used locations of users are acquired through historical trajectory data generated by the GNSS high-precision positioning module;

[0022] The gasoline grade, fuel tank capacity and fuel consumption of the user's car are obtained through the voice interaction module or derived from the user's car model data in the front-end database;

[0023] The current cruising range is calculated by the algorithm module based on the user's last refueling situation, historical trajectory, and fuel consumption.

[0024] User consumption preferences are obtained based on user feedback on historical refueling recommendation plans.

[0025] As a further preferred technical solution of the present invention, the preset conditions for initiating a refueling recommendation include:

[0026] A. Issue refueling recommendation request through the voice interaction module of the vehicle electronic terminal;

[0027] B. The current cruising range is lower than the set threshold 1, and the navigation route contains a gas station with a high cost-effectiveness;

[0028] C. Before the oil price adjustment window period, and the current cruising range is lower than the set threshold 2.

[0029] As a further preferred technical solution of the present invention, the external data includes: a list of surrounding gas stations, price inquiries for each gas station, navigation routes for each gas station, and traffic congestion conditions.

[0030] As a further preferred technical solution of the present invention, the optimal recommendation solution is generated by using the large language model to perform the following steps using a sorting algorithm:

[0031] First, obtain the alternative gas stations in the list of surrounding gas stations, then calculate the time cost and travel cost of each gas station for refueling, and calculate the expected fuel cost of this refueling based on the user's refueling amount and the gas station's fuel price, and then calculate the total cost of this refueling. Finally, select one or more gas stations with the lowest total cost of this refueling to generate the best recommendation plan.

[0032] As a further preferred technical solution of the present invention, the total refueling cost is calculated by the following formula:

[0033] Total refueling cost = time cost × user preference coefficient + distance cost + expected refueling cost.

[0034] As a further preferred technical solution of the present invention, based on user feedback, executing a refueling recommendation plan includes including using the recommended gas station location as an intermediate destination in the user's navigation path and applying for coupons or other discount offers of the recommended gas station for the user.

[0035] As a further preferred technical solution of the present invention, based on user feedback, modifying the current recommendation solution includes:

[0036] A. Modify to recommend gas stations with lower fuel prices;

[0037] B. Modified to recommend a gas station that is closer;

[0038] C. Modified to recommend gas stations within the designated area;

[0039] D. Postpone the recommendation time and make the recommendation when the oil volume is lower;

[0040] According to user feedback, the relevant parameters for optimizing the refueling recommendation system include:

[0041] A. Optimize user preference coefficient;

[0042] B. Optimize the range threshold;

[0043] C. Modify the user's gasoline grade, fuel tank capacity and fuel consumption.

[0044] As a further preferred technical solution of the present invention, the vehicle-mounted electronic terminal transmits data with the background server through the long-distance communication module, and the background server receives the data from the vehicle-mounted electronic terminal through the front-end data interface;

[0045] The long-distance wireless communication includes one of 4G or 5G communication.

[0046] As a further preferred technical solution of the present invention, the large language model is either privately deployed on a background server, or a third-party large model is called using an API interface. The large language model serves as the core controller of the AI-Agent and manages and schedules each unit to perform operations.

[0047] The memory unit includes long-term memory and short-term memory. The long-term memory includes the user's consumption preferences, the user's frequently used places, the user's car gasoline grade, fuel tank capacity and fuel consumption. The short-term memory includes the current cruising range, current location and destination, current voice interaction content and the National Development and Reform Commission's oil price adjustment plan.

[0048] The planning unit is responsible for initiating refueling recommendations when preset conditions are met, and modifying refueling recommendation plans based on user feedback;

[0049] The tool unit is responsible for calling the API interface to query the external data required for refueling recommendation;

[0050] The action unit uses the voice interaction module of the vehicle-mounted electronic terminal to perform refueling recommendations on the user through voice interaction.

[0051] As described above, the AI-Agent-based refueling recommendation device and method provided by the present invention has the following beneficial effects:

[0052] 1. The refueling recommendation method of the present invention is different from the refueling recommendation product architecture of the Internet era, which only provides location information when customers inquire about refueling. Instead, it actively understands customer needs, makes recommendations at the time when refueling is needed, and comprehensively combines all internal and external data to recommend the lowest cost solution. It is a butler-style Agent service based on precise time and space matching with car owners' needs.

[0053] 2. The refueling recommendation method of the present invention is a new product in the AI ​​era. It adopts the best interactive mode during vehicle driving - voice interaction, which perfectly solves the problem of manual input, greatly facilitates car owners, and improves driving safety.

[0054] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0056] Figure 1 A schematic block diagram of an AI-Agent-based refueling recommendation device and method for the present invention;

[0057] Figure 2 This is a system module architecture diagram of an AI-Agent-based refueling recommendation device and method applied for by the present invention. DETAILED DESCRIPTION

[0058] The following is a description of the implementation of the present invention by means of specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0059] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions that the present invention can be implemented, so they have no substantial technical significance. Any modification of the structure, change in the proportional relationship, or adjustment of the size should still fall within the scope of the technical content disclosed by the present invention without affecting the effects that the present invention can produce and the purposes that can be achieved. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description, and are not used to limit the scope of the implementation of the present invention. The change or adjustment of the relative relationship should also be regarded as the scope of the implementation of the present invention without substantial changes in the technical content. The specific structure can be described with reference to the drawings of the patent application.

[0060] The present invention provides a refueling recommendation device and method based on AI-Agent, please refer to Figure 1 to Figure 2 As shown, the refueling recommendation device includes an on-board electronic terminal and a backend server, combined with Figure 1 As shown, the vehicle-mounted electronic terminal is provided with a GNSS high-precision positioning module, a long-distance communication module, a voice interaction module, an algorithm module and a front-end database, and the back-end server is provided with a large model algorithm module, a voice recognition module, a voice synthesis module, an API interface and a front-end data interface;

[0061] The vehicle-mounted electronic terminal transmits data with the background server through the long-distance communication module, and the background server receives data from the vehicle-mounted electronic terminal through the front-end data interface;

[0062] The long-distance wireless communication includes one of 4G or 5G communication.

[0063] Combination Figure 2 As shown, the functional modules of the vehicle-mounted electronic terminal and the backend server together constitute a refueling recommendation system of the AI-Agent general architecture, and the refueling recommendation system includes a large language model, a memory unit, a planning unit, a tool unit, and an action unit;

[0064] The large language model is either privately deployed on a background server or uses an API interface to call a third-party large model. The large language model serves as the core controller of the AI-Agent and manages and schedules each unit to perform operations.

[0065] The memory unit includes long-term memory and short-term memory. The long-term memory includes the user's consumption preferences, the user's frequently used places, the user's car gasoline grade, the fuel tank capacity and fuel consumption. The short-term memory includes the current cruising range, the current location and destination, the current voice interaction content and the National Development and Reform Commission's oil price adjustment plan. It should be noted that: the long-term memory data involving user privacy information is stored in the front-end database of the vehicle terminal, and is only called when the large language model adopts the sorting algorithm. It is not stored on the background server and will not be sent to a third party; the short-term memory data not involving user privacy information is stored on the background server and is cleared after the refueling recommendation plan is executed;

[0066] The planning unit is responsible for initiating refueling recommendations when preset conditions are met, and modifying refueling recommendation plans based on user feedback;

[0067] The tool unit is responsible for calling the API interface to query the external data required for refueling recommendation and is deployed on the background server;

[0068] The action unit uses the voice interaction module of the vehicle-mounted electronic terminal to perform refueling recommendations on the user through voice interaction.

[0069] The refueling recommendation method is based on the refueling recommendation device and the refueling recommendation system, and specifically includes the following steps:

[0070] S1: Utilize various modules of the vehicle-mounted electronic terminal to collect various data required by the refueling recommendation system. The data collected by various modules of the vehicle-mounted electronic terminal include user consumption preferences, user's frequently used locations, user's car gasoline grade, fuel tank capacity, fuel consumption, current cruising range, current location and destination, and current voice interaction data, wherein:

[0071] The current location is obtained in real time by the GNSS high-precision positioning module;

[0072] Frequently used locations of users are acquired through historical trajectory data generated by the GNSS high-precision positioning module;

[0073] The gasoline grade, fuel tank capacity and fuel consumption of the user's car are obtained through the voice interaction module or derived from the user's car model data in the front-end database;

[0074] The current cruising range is calculated by the algorithm module based on the user's last refueling situation, historical trajectory, and fuel consumption.

[0075] User consumption preferences, obtained based on user feedback on historical refueling recommendation plans;

[0076] It should be noted that among all the data collected by the vehicle electronic terminal, the current voice interaction data is transmitted to the background server through the long-distance communication module, and the background server then performs voice recognition and semantic analysis through the large model algorithm module and the voice recognition module; other data is stored in the front-end database of the vehicle electronic terminal, and is only uploaded to the background server to participate in the calculation of the current refueling recommendation algorithm after the refueling recommendation is initiated, and is not stored in the background server;

[0077] S2: Analyze the above data using the algorithm module deployed in the vehicle electronic terminal and the large model algorithm module of the background server, and initiate a refueling recommendation when the preset conditions are met. The preset conditions for initiating a refueling recommendation include: A. The refueling recommendation requirement is issued through the voice interaction module of the vehicle electronic terminal; B. The current cruising range is lower than the set threshold value 1, and the navigation path contains a gas station with a high cost performance; C. Before the oil price adjustment window period, and the current cruising range is lower than the set threshold value 2;

[0078] Among them, threshold 1 is a lower threshold. When the demand for refueling is lower than this threshold, the demand is more urgent. Threshold 1 is usually set between 10%-20% of the total capacity of the fuel tank. For example, if the total capacity of the fuel tank is 50 liters, then threshold 1 can be set to 5-10 liters. Threshold 2 is a higher threshold. When the demand for refueling is lower than this threshold, refueling can be considered. Threshold 2 is usually set between 25%-35% of the total capacity of the fuel tank. For example, if the total capacity of the fuel tank is 50 liters, then threshold 2 can be set to 12.5-17.5 liters.

[0079] Among them, the preset condition A is determined by the large model algorithm module of the background server and initiates the refueling recommendation, while the preset conditions B and C are determined by the algorithm module of the vehicle electronic terminal and initiate the refueling recommendation;

[0080] S3: Based on various data, the API interface is called to query the external data required for this refueling recommendation, and the external data includes: a list of surrounding gas stations, price inquiries for each gas station, navigation routes for each gas station, and traffic congestion conditions.

[0081] S4: using a large language model deployed on a backend server to synthesize all data using a sorting algorithm to generate an optimal recommendation plan, wherein the optimal recommendation plan is based on the principle of minimizing the total cost of refueling;

[0082] The optimal recommendation solution is generated by using the large language model and a sorting algorithm to perform the following steps:

[0083] First, obtain each candidate gas station in the list of surrounding gas stations, then calculate the time cost and travel cost of each gas station, and calculate the expected fuel cost of this refueling based on the user's refueling amount and the gas station's fuel price, and then calculate the total cost of this refueling. Finally, select one or more gas stations with the lowest total cost of this refueling to generate the best recommendation plan;

[0084] The total refueling cost is calculated by the following formula:

[0085] Total refueling cost = time cost × user preference coefficient + distance cost + expected refueling cost.

[0086] in:

[0087] The user preference coefficient is a coefficient that indicates whether the user values ​​time cost or money cost more. In the initial stage, multiple gas stations with different time costs and distance costs can be recommended for the user to choose. In the later stage, the user's preference is evaluated based on the user's historical choices and quantified into the user preference coefficient.

[0088] Time cost is the time required for refueling after considering the navigation path and traffic congestion. If the current destination is known, the time cost is the additional time after adding the gas station to the original planned path;

[0089] The distance cost is the gasoline cost consumed by the user during the refueling journey after comprehensively considering the navigation route and traffic congestion. If the current destination is known, the time cost is the gasoline cost consumed by the additional distance after adding the gas station to the original planned route.

[0090] The expected fuel cost for refueling is the product of the expected fuel volume and the fuel price;

[0091] S5: broadcasting the recommended solution in the form of voice through the voice interaction module of the vehicle-mounted electronic terminal;

[0092] S6: Execute the refueling recommendation plan according to the user's feedback, or modify the current recommendation plan and optimize the relevant parameters of the refueling recommendation system;

[0093] Based on the user's feedback, executing the refueling recommendation solution includes including the recommended gas station location as an intermediate destination in the user's navigation path and applying for a coupon or other discount offer of the recommended gas station for the user.

[0094] Based on user feedback, modifications to the current recommendations include:

[0095] A. Modify to recommend gas stations with lower fuel prices;

[0096] B. Modified to recommend a gas station that is closer;

[0097] C. Modified to recommend gas stations within the designated area;

[0098] D. Postpone the recommendation time and make the recommendation when the oil volume is lower;

[0099] According to user feedback, the relevant parameters for optimizing the refueling recommendation system include:

[0100] A. Optimize user preference coefficient;

[0101] B. Optimize the range threshold;

[0102] C. Modify the user's gasoline grade, fuel tank capacity and fuel consumption.

[0103] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A refueling recommendation device and method based on AI-Agent, characterized in that: The refueling recommendation device includes an on-board electronic terminal and a backend server. The on-board electronic terminal is provided with a GNSS high-precision positioning module, a long-distance communication module, a voice interaction module, an algorithm module and a front-end database. The backend server is provided with a large model algorithm module, a voice recognition module, a voice synthesis module, an API interface and a front-end data interface. The functional modules of the vehicle-mounted electronic terminal and the backend server together constitute a refueling recommendation system of the AI-Agent general architecture, and the refueling recommendation system includes a large language model, a memory unit, a planning unit, a tool unit, and an action unit; The refueling recommendation method is based on the refueling recommendation device and the refueling recommendation system, and specifically includes the following steps: S1: Use various modules of the vehicle electronic terminal to collect various data required by the refueling recommendation system; S2: Analyze the above data using the algorithm module deployed in the vehicle electronic terminal and the large model algorithm module of the backend server, and initiate a refueling recommendation when the preset conditions are met; S3: Calling API interfaces based on various data to query external data required for this refueling recommendation; S4: using a large language model deployed on a backend server to synthesize all data using a sorting algorithm to generate an optimal recommendation plan, wherein the optimal recommendation plan is based on the principle of minimizing the total cost of refueling; S5: broadcasting the recommended solution in the form of voice through the voice interaction module of the vehicle-mounted electronic terminal; S6: Based on user feedback, execute the refueling recommendation plan, or modify the current recommendation plan and optimize the relevant parameters of the refueling recommendation system.

2. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: The data collected by various modules of the vehicle-mounted electronic terminal include user consumption preferences, user's frequently used places, user's car gasoline grade, fuel tank capacity, fuel consumption, current cruising range, current location and destination, and current voice interaction data; The current location is obtained in real time by the GNSS high-precision positioning module; Frequently used locations of users are acquired through historical trajectory data generated by the GNSS high-precision positioning module; The gasoline grade, fuel tank capacity and fuel consumption of the user's car are obtained through the voice interaction module or derived from the user's car model data in the front-end database; The current cruising range is calculated by the algorithm module based on the user's last refueling situation, historical trajectory, and fuel consumption. User consumption preferences are obtained based on user feedback on historical refueling recommendation plans.

3. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: The preset conditions for initiating a refueling recommendation include: A. Issue refueling recommendation request through the voice interaction module of the vehicle electronic terminal; B. The current cruising range is lower than the set threshold 1, and the navigation route contains a gas station with a high cost-effectiveness; C. Before the oil price adjustment window period, and the current cruising range is lower than the set threshold 2.

4. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: The external data includes: a list of surrounding gas stations, price inquiries for each gas station, navigation routes for each gas station, and traffic congestion conditions.

5. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: The optimal recommendation solution is generated by using the large language model and a sorting algorithm to perform the following steps: First, obtain the alternative gas stations in the list of surrounding gas stations, then calculate the time cost and travel cost of each gas station for refueling, and calculate the expected fuel cost of this refueling based on the user's refueling amount and the gas station's fuel price, and then calculate the total cost of this refueling. Finally, select one or more gas stations with the lowest total cost of this refueling to generate the best recommendation plan.

6. The AI-Agent-based refueling recommendation device and method according to claim 5, characterized in that: The total refueling cost is calculated by the following formula: Total refueling cost = time cost × user preference coefficient + distance cost + expected refueling cost.

7. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: Based on the user's feedback, executing the refueling recommendation solution includes including the recommended gas station location as an intermediate destination in the user's navigation path and applying for a coupon or other discount offer of the recommended gas station for the user.

8. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: Based on user feedback, modifications to the current recommendations include: A. Modify to recommend gas stations with lower fuel prices; B. Modified to recommend a gas station that is closer; C. Modified to recommend gas stations within the designated area; D. Postpone the recommendation time and make the recommendation when the oil volume is lower; According to user feedback, the relevant parameters for optimizing the refueling recommendation system include: A. Optimize user preference coefficient; B. Optimize the range threshold; C. Modify the user's gasoline grade, fuel tank capacity and fuel consumption.

9. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: The vehicle-mounted electronic terminal transmits data with the background server through the long-distance communication module, and the background server receives data from the vehicle-mounted electronic terminal through the front-end data interface; The long-distance wireless communication includes one of 4G or 5G communication.

10. The AI-Agent-based refueling recommendation device and method according to claim 1, characterized in that: The large language model is either privately deployed on a background server or uses an API interface to call a third-party large model. The large language model serves as the core controller of the AI-Agent and manages and schedules each unit to perform operations. The memory unit includes long-term memory and short-term memory. The long-term memory includes the user's consumption preferences, the user's frequently used places, the user's car gasoline grade, fuel tank capacity and fuel consumption. The short-term memory includes the current cruising range, current location and destination, current voice interaction content and the National Development and Reform Commission's oil price adjustment plan. The planning unit is responsible for initiating refueling recommendations when preset conditions are met, and modifying refueling recommendation plans based on user feedback; The tool unit is responsible for calling the API interface to query the external data required for refueling recommendation; The action unit uses the voice interaction module of the vehicle-mounted electronic terminal to perform refueling recommendations on the user through voice interaction.

Images