Consumption power prediction analysis and improvement suggestion system for battery electric vehicle using lllm

Abstract

This disclosure relates to a power consumption prediction analysis and improvement prompt system for battery electric vehicles using LLM, comprising: an estimation unit that, based on the current driving state and past driving history, uses LLM to estimate the power consumption of the battery in a battery-powered electric vehicle on a predetermined route, and uses LLM to estimate improvement measures to minimize the estimated power consumption; and a prompting unit that prompts the driver with the estimated improvement measures.

Description

Technical Field

[0001] This invention relates to the technical field of systems applicable to battery electric vehicles, including, for example, battery electric vehicles (so-called BEVs), which predict and analyze the power consumption of the battery and, if necessary, suggest improvements to that power consumption to enable reaching the destination. Background Technology

[0002] As a technology associated with this system, a range notification device has been developed as a navigation device for battery-electric vehicles. This device displays the remaining range of the vehicle based on the remaining battery power and notifies the user whether to return home or to a designated charging station (see WO14 / 188652). Summary of the Invention

[0003] However, based on the aforementioned background technology, simply displaying information related to remaining range or whether the destination can be reached in a navigation device may not be acceptable to users or drivers. The accuracy and precision of such displayed information are also questionable. In particular, without detailed explanations of power consumption and the impact of various factors, it is unclear which aspect poses the greatest problem. Furthermore, there is a technical problem in providing alerts when the destination cannot be reached.

[0004] The objective of this invention is to provide a power consumption prediction analysis and improvement prompt system for battery-electric vehicles using LLM, which can predict power consumption with more accurate prediction precision and prompt power consumption improvement measures in natural language that humans can understand.

[0005] To address the aforementioned issues, one aspect of the present invention, a power consumption prediction analysis and improvement prompting system for battery-electric vehicles using LLM, comprises: an estimation unit that, based on information related to the driver and driving of the battery-electric vehicle, uses LLM to estimate the power consumption of the battery in the battery-electric vehicle on a predetermined route to its destination, and uses LLM to estimate improvement measures that minimize the estimated power consumption, wherein the information includes the current driving status and past driving history of the battery-electric vehicle, and the past driving history of battery-electric vehicles of the same model as the battery-electric vehicle; and a prompting unit that prompts the driver with the estimated improvement measures.

[0006] According to one aspect of the system of the present invention, by using LLM, it is possible to predict power consumption with greater accuracy through high-dimensional learning that includes text, and to suggest improvements to power consumption in natural language that humans can understand.

[0007] The effects of this invention become more apparent through the following description of the embodiments of the invention. Attached Figure Description

[0008] The features, advantages, and technical and industrial significance of exemplary embodiments of the present invention will now be described with reference to the accompanying drawings, in which the same reference numerals denote the same elements, and are attached as follows:

[0009] Figure 1 This is a block diagram illustrating the overall structure of the system in the implementation method;

[0010] Figure 2 This is a flowchart illustrating an example of processing in a system implementing an embodiment. Detailed Implementation

[0011] First, refer to Figure 1 This document describes the overall structure of a battery-electric vehicle power consumption prediction analysis and improvement prompt system (hereinafter referred to as "analysis and improvement prompt system" where appropriate) using LLM (Large Language Models) according to an implementation method.

[0012] The LLM used in the analysis and estimation of this embodiment, or in the language prompts, can be a so-called unimodal LLM or a multimodal LLM. This embodiment is constructed as a system for performing LLM analysis or AI analysis based on LLM learning data. The LLM learning data appropriately includes, for example, various information such as: the driving status of the battery electric vehicle, past driving history and various information of various elements, various information of the driving path or map in the predetermined route, information on the past driving history of battery electric vehicles of the same or similar model as the battery electric vehicle, various personal information of the user or driver of the battery electric vehicle (i.e., the user or driver of the battery electric vehicle and other battery electric vehicles), information as traffic regulations and common knowledge of traffic regulations, and information as general common knowledge.

[0013] Specifically, as detailed below, the system is designed to provide highly accurate information regarding remaining range and the feasibility of reaching the destination, based on battery level, driving method, and driving operations, in a format easily understood by the user or driver. Furthermore, by explaining the detailed consumption of power and the impact of each factor in this way, the user or driver can clearly understand which aspect is most problematic. It also provides guidance on measures to take if the destination cannot be reached.

[0014] Furthermore, in addition to traditional AI learning systems such as teacher-led, teacherless, or reinforcement learning, this type of AI learning or LLM learning can also employ new technologies such as generative AI or LLM that have recently been put into practical use or are currently under development or will be developed in the future. For example, this type of AI learning or LLM learning can include neural networks that perform efficient learning through representation learning, transfer learning, feature selection, fine-tuning or hyperparameter tuning, ensemble learning, etc.

[0015] like Figure 1 As shown, the analysis and improvement prompting system of this embodiment is configured to include an on-board unit 101 and a server unit 200 mounted on a battery-electric vehicle 100. The on-board unit 101 and the server unit 200 are housed in a communication network 10, such as the Internet or a dedicated network line. Similarly, multiple or numerous other battery-electric vehicles 100 are housed in the communication network 10. Furthermore, the communication network 10 includes an external related knowledge collection unit 301, which collects information obtainable outside the battery-electric vehicle 100. This information can be used to perform fine-tuning or over-tuning in the analysis and improvement prompting system to impart domain knowledge (i.e., "external related knowledge"). The external related knowledge collection unit 301 may be at least partially located within the server unit 200 or within a device equipped with the server unit 200, or it may be located within the on-board unit 101 or inside the vehicle.

[0016] A database (DB) 300, which stores various data including data used in the analysis and improvement system, is connected to a server unit 200. The DB 300 can also be connected to the server unit 200 or the vehicle-mounted unit 101 via a communication network 10. The server unit 200 includes various computer-mounted devices and computer equipment for centralized or distributed processing; in other words, the analysis and improvement system is configured to utilize the large-scale data from the DB 300 for centralized or distributed processing.

[0017] exist Figure 1 In this context, the battery electric vehicle 100 includes a battery 150, for example, configured as a BEV. Alternatively, the battery electric vehicle 100 can also be a battery-powered vehicle such as a HEV (Hbrid Electric Vehicle), PHEV (Plug-in HEV), or FCEV (Fuel Cell EV).

[0018] The vehicle unit 101 is configured to include: a sensor unit 102, including various sensors installed at predetermined locations inside the vehicle; a processing unit 103, including a computer; a communication unit 104, including a modem or the like configured to communicate with the outside of the vehicle via a communication network 10; and an interface unit 106 configured to interact with a user or driver inside the vehicle via sound or images.

[0019] As one of the detection functions, the sensor unit 102 detects the remaining amount of battery 150 and sends it to the processing unit 103. In addition, the sensor unit 102 also detects the current driving status of the battery electric vehicle 100 and various information 102a of the driver of the battery electric vehicle 100, and delivers it to the processing unit 103 as CAN (Controlled Area Network) data, etc.

[0020] The processing unit 103 includes a CPU, memory, etc., for controlling the sensor unit 102, the communication unit 104, and the interface unit 106. It transmits various information concerning the driver and driving of the battery-electric vehicle 100, including the predetermined route of the battery-electric vehicle 100, the current driving status of the battery-electric vehicle 100, and past driving history, in a predetermined format from the communication unit 104 to the server unit 200. Furthermore, it is configured to display improvement measures, etc., represented by improvement measure data received from the server unit 200 via the communication unit 104 and processed on the server unit 200, to the user or driver via the interface unit 106.

[0021] Under the control of the processing unit 103, the communication unit 104 transmits data collected in the battery-electric vehicle 100 for power consumption prediction analysis and improvement suggestions to the server unit 200 via the communication network 10. Furthermore, it is also configured to receive data related to the power consumption prediction analysis of the battery-electric vehicle 100 generated in the server unit 200 using LLM and improvement suggestions via the communication network 10.

[0022] The interface unit 106 is configured to allow input of the destination of the battery-electric vehicle 100, and input of conditions for selecting a predetermined route to the destination, via voice input or predetermined operations on an image. Regarding the selection of the predetermined route (i.e., the navigation function), it can be configured to be performed entirely or partially by the processing unit 103, or it can be configured to be performed partially or entirely by the processing unit 202 on the server unit 200 side (in other words, a browser function is specifically implemented on the vehicle unit 101 side). The interface unit 106 is also configured to output data obtained from the server unit 200 side related to the results of power consumption prediction analysis and improvement prompts via voice output or via a predetermined format on an image.

[0023] exist Figure 1 In this unit, the server unit 200 includes: a communication unit 201, including a modem or the like that can communicate with each battery-electric vehicle 100 and the external related knowledge collection unit 301 via the communication network 10; a processing unit 202, including a computer that can perform LLM-based power consumption estimation processing, which will be detailed later; and a prompting unit 203 that can generate prompt data that indicates the improvement measures corresponding to the estimation results of the processing unit 202.

[0024] Under the control of the processing unit 202, the communication unit 201 receives data collected in the battery electric vehicle 100, which is needed for power consumption prediction analysis and improvement prompts, via the communication network 10. Under the control of the processing unit 202, the communication unit 201 receives external correlation knowledge collected by the external correlation knowledge collection unit 301 via the communication network 10 as part of the data needed for power consumption prediction analysis and improvement prompts. The communication unit 201 is also configured to transmit data generated by the processing unit 202 and the prompting unit 203, which is relevant to power consumption prediction analysis and improvement prompts for the battery electric vehicle 100, the object of this analysis, to the battery electric vehicle 100 via the communication network 10.

[0025] The processing unit 202 is configured to: with respect to the battery electric vehicle 100, which is the object of this analysis, based on information related to the driver and driving of the battery electric vehicle 100, use LLM to estimate the battery power consumed on a predetermined route to the destination, and use LLM to estimate improvement measures to make the estimated power consumption close to a minimum. The information related to the driver and driving of the battery electric vehicle 100 includes the current driving status and past driving history of the battery electric vehicle 100, as well as the past driving history of other battery electric vehicles 100 of the same model as the battery electric vehicle 100.

[0026] The prompting unit 203 is configured to generate prompt data in a predetermined format corresponding to the interface unit 106 installed inside the battery electric vehicle 100, indicating that the proposed improvement measures are to be prompted to the driver or user of the battery electric vehicle 100, and then deliver this data to the communication unit 201. In this embodiment, the "prompting unit" is configured as including the prompting unit 203 located on the server unit 200 side and the interface unit 106 located on the vehicle unit 101 side, with the vehicle unit 101 side primarily performing browser functions regarding the prompting function.

[0027] DB300 is configured to include a large-scale storage device that performs high-speed data input / output. This storage device stores various data received on the server unit 200 side via the communication network 10, especially various data required for estimation processing using LLM, estimation results or related data of intermediate processes generated in the processing unit 202, prompt data generated in the prompting unit 203, etc.

[0028] Next, besides Figure 1 In addition to the block diagram, it also references Figure 2 The flowchart illustrates an example of the processing in the analysis and improvement prompt system of this embodiment (in particular, the processing performed using LLM in the processing unit 202 of the server unit 200).

[0029] exist Figure 2 First, using the browser function of the navigation device in the interface unit 106 of the vehicle unit 101, the user inputs the "destination" for this trip. Then, a route search to the destination is performed by the processing unit 103 or by the processing unit 202 via the communication network 10 (step S1).

[0030] Next, the processing unit 202 in the server unit 200 determines whether there is a travel history for each road segment of the searched predetermined route (step S2). This determination function can also be partially or entirely performed by the processing unit 103 on the vehicle unit 101 side. The data related to the travel history is basically stored in the DB300. Additionally, a portion of the data related to the travel history can also be stored in the memory of the processing unit 103 on the vehicle unit 101 side or in a separate vehicle memory.

[0031] If the determination result of step S2 is no driving history (step S2: No), the "current driving status" is directly obtained by the sensor unit 102 and processing unit 103 on the vehicle unit 101 side and handed over to the processing unit 202 on the server unit 200 side (step S3). After that, the processing of the LLM-based estimated improvement measures and the like in the processing unit 202 is carried out (step S4 and onwards).

[0032] On the other hand, if the determination result of step S2 is that there is a driving history (step S2: yes), then the processing unit 202 obtains "the past driving history of battery electric vehicles (BEVs) of the same model as the battery electric vehicle 100 for each road segment" from the DB300, etc. (step S5). Here, "battery electric vehicles of the same model" includes not only BEVs that are exactly the same model or of the same type as its own BEV, but also BEVs with pre-set common elements or similar elements. For example, if two vehicles have the same powertrain or the same navigation number, these vehicles can be regarded as the same model. Furthermore, for minor differences between two vehicles, after applying correction through LLM to make them able to be treated as vehicles of the same model, the corrected driving history can be used for processing related to power consumption (step S6 and later).

[0033] Next, the processing unit 202 extracts the history equivalent to "minimum power consumption" from the obtained past driving history (step S6). Afterwards, the processing unit 202 performs the process of estimating improvement measures using LLM (step S4 and later). Here, in order to estimate improvement measures based on past driving history using LLM (step S4 and later), the relationship curves and numerical data between accumulated elements and power consumption for each road segment are all transformed into documents using LLM, and then vectorized.

[0034] Simultaneously with, before, or after the processing in steps S1 to S6 described above, the external correlation knowledge collection unit 301 acquires external correlation knowledge (step S11), and the processing unit 202 performs fine-tuning or over-tuning to impart domain knowledge (step S12). Afterwards, the processing unit 202 performs presumed improvement measures using LLM (step S4 onwards). In this way, by using the LLM fine-tuned using domain knowledge related to battery electric vehicles, such as BEV domain knowledge, remedial measures can be generated by comparing it with optimal driving settings.

[0035] Next, the processing unit 202 performs an estimation process (step S4) to identify the difference between the driving state obtained in step S3 and the driving state extracted in step S6. The identification of the difference using LLM here is based not only on various data obtained from the vehicle unit 101 and various data obtained or extracted from the database, but also on the domain knowledge assigned in step S12.

[0036] Next, data related to personal preferences that was previously stored in DB300 is retrieved from DB300 as driver-specific data of the battery electric vehicle 100, and this data is transmitted to the processing unit 202 (step S7).

[0037] Next, based on the various data acquired or generated in steps S3, S4, S6, S7, etc., the processing unit 202 generates power consumption improvement measures by using LLM estimation (step S8). Here, as improvement measures, LLM is used to estimate improvement measures that bring power consumption close to the minimum.

[0038] The LLM performed in steps S4 and S8 above uses a large amount of information, i.e., a large amount of verbalized information, as a large amount of text data. This information appropriately includes, for example, the driving status of the battery electric vehicle 100, past driving history and various information of various elements, various information of the driving path or map in the predetermined route, information of the past driving history of battery electric vehicles of the same or similar model as the battery electric vehicle 100, various personal information of the user or driver of the battery electric vehicle 100 (i.e., the user or driver of the battery electric vehicle 100 and other battery electric vehicles 100), information as traffic regulations and common sense involved in traffic regulations (e.g., driving on the left, speed limits in residential areas, etc.), and information as general common sense (e.g., it is dark at night, traffic jams are common during rush hour or holidays, landmarks near the predetermined route, etc.).

[0039] However, in these steps S4, S8, etc., a multimodal LLM that can perform AI learning not only based on verbal information but also on non-verbal information can also be adopted. That is, the data used in the processing of steps S4, S8, etc. includes, but is not limited to, text data.

[0040] In the processing steps S4, S8, etc., described above, a large amount of text data is used, and the LLM is fine-tuned. As a result, it can be applied to various natural language processing (NLP) tasks such as text classification and sentiment analysis, information extraction, document summarization, text generation, and question-and-response.

[0041] In step S8, the prompting unit 203 further generates prompt data to indicate the improvement measures generated in the LLM, and the interface unit 106 on the vehicle unit 101 side outputs the "improvement measures" via sound or image output. Furthermore, the processing unit 202 and the prompting unit 203 also output audio or image to the prompt data indicating the "basis" for the improvement measures via the interface unit 106. Alternatively, the generation of prompt data for sound or image output of the improvement measures in the prompting unit 203 can be performed using the LLM. That is, here, various natural language processing tasks such as text classification and sentiment analysis, information extraction, document summarization, text generation, and question-and-answer can be matched to provide AI-based sound or image prompts to users or drivers.

[0042] Next, determine if there is feedback (step S9). If there is (step S9: yes) Y, return to step S7 and repeat the subsequent processing, prompting the "improvement measures" that have been updated through AI learning (step S8).

[0043] The suggested improvement measures include, for example, "On this highway, to reduce power consumption, please stay in the left lane and maintain a constant speed of 80 km / h," "Due to low battery levels, please avoid sudden acceleration or deceleration before the next charge," and "Charge at the charging station 25 km ahead along the road." Finally, at the interface 106 on the vehicle unit 101 side, these prompts are displayed to the driver via AI voice or AI image output.

[0044] In this embodiment, it is preferable to provide the rationale behind the improvement measures along with the suggestions. This is preferable from the perspective of gaining the user's or driver's approval, or in other words, from the perspective of ensuring the driver complies with the improvement measures. Examples of the rationale provided include: "(Based on) the vehicle's energy consumption is optimal at 80 km / h," "(Based on) in the future, repeated acceleration and deceleration will not maintain the battery level until the next charging station," and "(Based on) if there is a charging station 25 km ahead, you can drive as usual without any problem."

[0045] After processing related to the prompting of improvement measures, if the determination result in step S9 is that there is no feedback "Step S9: No", then the series of processes ends.

[0046] As detailed above, according to this embodiment, the past driving history of battery electric vehicles such as BEVs of the same model is extracted for each road segment (steps S5 and S6), the relationship between driving status (air conditioning, vehicle speed, vehicle weight, in-vehicle temperature, etc.) and power consumption is input into the LLM (step S3), and then the difference between the current setting and the driving status with minimum power consumption is output from the LLM (step S4), and the improvement measures for saving power (i.e., improvement measures to make power consumption close to the minimum value) are prompted by language or sound (steps S7 and S8).

[0047] Furthermore, based on user feedback (step S9), suggestions for improvement measures that further consider personal preferences can be provided in subsequent steps (steps S7 and S8). In this way, LLM-specific feedback mechanisms such as RLHF (i.e., reinforcement learning from human feedback) can be used to reflect personal preferences. Therefore, feedback on each element affecting power consumption can be collected, and improvement measures that are more helpful to the individual can be regenerated.

[0048] According to this embodiment, not only numerical data as in the prior art or background art, but also textual data such as papers and related literature can be extracted using LLM, and power consumption can be calculated or estimated at a higher dimension. Furthermore, according to this embodiment, by using LLM, results, evidence, and improvement measures can be provided to users simultaneously in natural language or voice, thereby improving the acceptance and satisfaction of the response.

[0049] Postscript

[0050] The following notes are also disclosed regarding the above-described embodiments.

[0051] Postscript 1

[0052] The analysis and improvement prompting system described in Appendix 1 of the present invention is characterized by comprising: an estimation unit that, based on information related to the driver and driving of the battery-electric vehicle equipped with a battery, uses LLM to estimate the battery power consumed in the battery-electric vehicle on a predetermined route to the destination, and uses LLM to estimate improvement measures to minimize the estimated power consumption, wherein the information includes the current driving status and past driving history of the battery-electric vehicle, and the past driving history of battery-electric vehicles of the same model as the battery-electric vehicle; and a prompting unit that prompts the driver with the estimated improvement measures.

[0053] According to the analysis and improvement suggestion system described in Appendix 1, by using LLM (Limited Learning Model), it is possible to predict power consumption with greater accuracy through high-dimensional learning that includes text or even more text. Furthermore, it can utilize natural language that humans can understand to explain the factors that have the greatest impact on power consumption and corresponding improvement measures. Additionally, the collected driving history can be provided as reference information to other vehicles.

[0054] Appendix 2

[0055] The analysis and improvement suggestion system described in Appendix 2 of the present invention is characterized in that, in the analysis and improvement suggestion system described in Appendix 1, the estimation unit identifies the difference between the current driving state and the driving state with the minimum power consumption in the past driving history through the LLM, thereby estimating the improvement measures.

[0056] According to the analysis and improvement suggestion system described in Note 2 of the present invention, by identifying the difference between the current driving state and the driving state with the minimum power consumption in the past driving history through LLM, it is possible to more efficiently estimate improvement measures to make power consumption approach the minimum.

[0057] Appendix 3

[0058] The analysis and improvement suggestion system described in Appendix 3 of the present invention is characterized in that: in the power consumption prediction analysis and improvement suggestion system described in Appendix 1 or 2, the estimation unit acquires external correlation knowledge other than the information of the current driving state and the past driving history, and estimates the improvement measures based on the domain knowledge related to the battery electric vehicle by using the acquired external correlation knowledge and fine-tuning the LLM.

[0059] The analysis and improvement suggestion system described in Appendix 3 of the present invention not only bases its predictions on information related to the current driving state and past driving history, but also infers improvement measures by using an LLM finely tuned with external correlation knowledge, taking into account domain knowledge, thus enabling more accurate predictions.

[0060] Appendix 4

[0061] The analysis and improvement suggestion system described in Appendix 4 of the present invention is characterized in that: in the power consumption prediction analysis and improvement suggestion system described in any one of Appendices 1 to 3, the estimation unit obtains information about the personal preferences of the driver of the battery electric vehicle, and uses an LLM-specific feedback mechanism on the obtained information about personal preferences to reflect personal preferences and estimate improvement measures through LLM.

[0062] The analysis and improvement prompting system described in Appendix 4 of the present invention not only estimates improvement measures based on information related to the current driving status and past driving history, but also in a manner that reflects personal preferences through LLM, thus enabling the prompting of highly accurate improvement measures corresponding to the driver or user in a more acceptable form.

[0063] Appendix 5

[0064] The analysis and improvement prompting system described in Appendix 5 of the present invention is characterized in that, in the power consumption prediction analysis and improvement prompting system described in any one of Appendices 1 to 4, the prompting unit will prompt the information indicating the basis for estimating the improvement measures together with the improvement measures.

[0065] The analysis and improvement prompting system described in Appendix 5 of the present invention not only prompts for improvement measures, but also for the basis of the improvement measures, thus enabling highly accurate prompting of improvement measures in a form that is easier for drivers or users to understand.

[0066] Appendix 6

[0067] The analysis and improvement suggestion method described in Appendix 6 of the present invention is characterized by comprising the following steps: based on information related to the driver and driving of the battery-electric vehicle equipped with a battery, estimating the battery power consumed in the battery-electric vehicle on a predetermined route to the destination using LLM, and estimating improvement measures to minimize the estimated power consumption using LLM, wherein the information includes the current driving status and past driving history of the battery-electric vehicle, as well as the past driving history of battery-electric vehicles of the same model as the battery-electric vehicle; and suggesting the estimated improvement measures to the driver.

[0068] According to Appendix 6 of the present invention, the analysis and improvement suggestion method, similar to the analysis and improvement suggestion system described in Appendix 1, can predict power consumption with more accurate prediction precision by using high-dimensional learning that includes text or even more text, and can also use natural language that humans can understand to explain the factors and improvement measures that have the greatest impact on power consumption.

[0069] The present invention can be modified appropriately without departing from the spirit or idea of ​​the invention as read in its entirety from the claims and description, and the analysis and improvement prompting system and method accompanying such modifications are also included in the technical concept of the present invention.

Claims

1. A power consumption prediction, analysis, and improvement suggestion system for battery-electric vehicles using LLM (Large-Scale Language Modeling), characterized in that, have: The estimation unit, based on information related to the driver and driving of the battery-electric vehicle equipped with a battery, uses LLM (Limited Linear Modeling) to estimate the battery power consumed by the battery on a predetermined route to the destination, and uses LLM to estimate improvement measures to minimize the estimated power consumption. The information includes the current driving status and past driving history of the battery-electric vehicle, as well as the past driving history of battery-electric vehicles of the same model as the battery-electric vehicle; and The prompting unit informs the driver of the proposed improvement measures.

2. The power consumption prediction, analysis, and improvement alert system for battery-electric vehicles using LLM as described in claim 1, characterized in that, The estimation unit uses the LLM to identify the difference between the current driving state and the driving state with the minimum power consumption in the past driving history, thereby estimating the improvement measures.

3. The power consumption prediction analysis and improvement prompt system for battery-electric vehicles using LLM according to claim 1 or 2, characterized in that, The estimation unit acquires external correlation knowledge beyond information related to the current driving state and the past driving history. By using the acquired external correlation knowledge, the LLM is finely adjusted, and the improvement measures are estimated based on domain knowledge related to the battery electric vehicle.

4. The power consumption prediction, analysis, and improvement alert system for battery-electric vehicles using LLM according to any one of claims 1 to 3, characterized in that, The estimation unit acquires information about the personal preferences of the driver of the battery-electric vehicle, and uses an LLM-specific feedback mechanism on the acquired personal preference information to reflect the personal preferences in a way that LLM estimates the improvement measures.

5. The power consumption prediction analysis and improvement prompt system for battery-electric vehicles using LLM according to any one of claims 1 to 4, characterized in that, The prompt section will display information indicating the basis for presuming the improvement measure along with the improvement measure itself.

6. A method for predicting and analyzing power consumption and providing improvement suggestions for battery-electric vehicles using LLM, characterized in that, Includes the following steps: Based on information related to the driver and driving of a battery-electric vehicle equipped with a battery, LLM (Limited Linear Modeling) is used to estimate the battery power consumed on a predetermined route to the destination in the battery-electric vehicle, and LLM is used to estimate improvement measures to minimize the estimated power consumption. The information includes the current driving status and past driving history of the battery-electric vehicle, as well as the past driving history of battery-electric vehicles of the same model as the battery-electric vehicle; and The driver is then prompted with the proposed improvement measures.

Images