Driver assistance systems and driver assistance methods

Abstract

The system accurately recalculates the remaining driving range based on the fluctuating electric power consumption / fuel efficiency, which varies depending on the characteristics of each vehicle and each driver. [Solution] The driver assistance system comprises an input unit that identifies the driver of a vehicle and inputs driver characteristic data indicating the driver's operation of the vehicle, and vehicle characteristic data related to the vehicle's electric power consumption / fuel consumption; a processing unit that performs machine learning based on the driver characteristic data and vehicle characteristic data to estimate the electric power consumption / fuel consumption associated with the driver and vehicle, and recalculates the cruising range based on the electric power consumption / fuel consumption and the vehicle's remaining charge / fuel level; and an output unit that outputs the cruising range.

Description

Technical Field

[0001] The present invention relates to a vehicle driving support system and a driving support method.

Background Art

[0002] There are multiple types of vehicles, such as fuel vehicles equipped with internal combustion engines like gasoline engines, hybrid vehicles equipped with both internal combustion engines and electric motors, and electric vehicles equipped with batteries and electric motors. In fuel vehicles, lead-acid batteries are used for power supply to electrical components. Nickel-metal hydride batteries are used for power supply to the electric motors of hybrid vehicles. Lithium-ion batteries are used for power supply to the electric motors of electric vehicles.

[0003] The cruising range of a fuel vehicle can be calculated by multiplying the remaining fuel amount by the fuel consumption (km / L). The cruising range of an electric vehicle can be calculated by multiplying the remaining battery charge by the electricity consumption (km / kWh). Based on the cruising range displayed on the meter, the driver can determine whether refueling or charging is necessary.

[0004] As transportation infrastructure in Japan, there are many gas stations such as gasoline stations, but the spread of charging stations is slower compared to gasoline stations. Also, since the charging capacity of the batteries of electric vehicles is limited, the driver needs to drive the electric vehicle while paying attention to the cruising range and the location of charging stations.

[0005] Patent Document 1 discloses a driving support system for an electric vehicle. The driving support system determines whether charging is necessary based on the planned driving distance from the current location of the electric vehicle and the drivable distance estimated from the remaining battery charge, and notifies the driver of the location information of charging stations within the range of the drivable distance.

Prior Art Documents

Patent Documents

[0006]

Patent Document 1

[0007] Patent Document 1 extracts charging stations within the driving range from pre-stored charging station location information and notifies the driver of the location of those charging stations. This is fine if the driver drives the electric vehicle around within a predetermined area, but there is a concern that when the driver drives the electric vehicle in a different area, there may be no charging stations that they have used before. In this case, the driver assistance system in Patent Document 1 has difficulty suggesting the location information of charging stations within the driving range.

[0008] Japan has a long, narrow landmass stretching from north to south, resulting in significant regional differences in outside temperature. Furthermore, atmospheric temperature fluctuates greatly between summer and winter. Therefore, the driving range calculated based on the remaining battery charge of an electric vehicle (EV) varies depending on the outside temperature. For example, during periods of low outside temperature, the actual driving distance of an EV may be shorter even with the same remaining charge, potentially leading to the EV running out of power during a trip. The battery charge in an EV is used not only to power the electric motor but also to supply power to electrical components such as air conditioners and navigation systems. While fuel-powered vehicles use batteries such as lead-acid batteries to power electrical components, the battery in an EV functions not only as a power source for propulsion but also for powering electrical components. Using the air conditioner for heating in an EV consumes a lot of power, which quickly depletes the battery charge, making it difficult to accurately calculate the driving range.

[0009] Each electric vehicle exhibits different tendencies regarding battery degradation and power consumption, and each driver has different tendencies regarding their driving style and operation of electrical components. The same applies to fuel-powered vehicles. As a result, the fuel / electricity consumption of a vehicle varies from vehicle to vehicle and / or from driver to driver. Furthermore, the vehicle's fuel / electricity consumption fluctuates depending on how the driver uses the vehicle (e.g., frequency of use). Moreover, fuel / electricity consumption fluctuates depending on the driving conditions. For example, when driving on roads with significant elevation changes or roads with many curves and traffic lights, the vehicle's fuel / electricity consumption decreases, and it also fluctuates depending on traffic congestion. Driving conditions influence the tendencies of each driver and each vehicle. Thus, because fuel / electricity consumption fluctuates according to the tendencies of each vehicle and each driver, it has been difficult to accurately calculate the driving range for each vehicle or each driver. Furthermore, it is necessary to alleviate drivers' anxiety about the possibility of running out of power or fuel by providing them with the exact remaining driving range for each vehicle or driver.

[0010] To solve the above-mentioned problems, the embodiments of the present invention aim to provide a driver assistance system and a driver assistance method that can accurately calculate the remaining driving range according to the tendencies of each vehicle and each driver. [Means for solving the problem]

[0011] A first aspect of the present invention is a driver assistance system comprising: an input unit that identifies the driver of a vehicle and inputs driver characteristic data indicating the driver's operation of the vehicle, and vehicle characteristic data related to the vehicle's electric power consumption or fuel efficiency; a processing unit that performs machine learning based on the driver characteristic data and vehicle characteristic data to estimate the electric power consumption or fuel efficiency associated with the driver and the vehicle, and recalculates the cruising range based on the electric power consumption or fuel efficiency and the vehicle's remaining charge or fuel level; and an output unit that outputs the cruising range.

[0012] A second aspect of the present invention is a driving assistance method that identifies the driver of a vehicle, inputs driver characteristic data indicating the driver's operation of the vehicle and vehicle characteristic data related to the vehicle's electric power consumption or fuel consumption, performs machine learning based on the driver characteristic data and vehicle characteristic data to estimate the electric power consumption or fuel consumption associated with the driver and the vehicle, recalculates the cruising range based on the electric power consumption or fuel consumption and the vehicle's remaining charge or fuel level, and outputs the cruising range. [Effects of the Invention]

[0013] According to embodiments of the present invention, the remaining driving range can be accurately calculated according to the tendencies of each vehicle and each driver. [Brief explanation of the drawing]

[0014] [Figure 1] This is a block diagram showing a configuration for implementing a driving range calculation task in a driver assistance system according to the first embodiment of the present invention. [Figure 2] (A) A configuration diagram showing the first example of machine learning applied to the processing unit of the driver assistance system shown in Figure 1, and (B) A configuration diagram showing the second example of machine learning applied to the processing unit of the driver assistance system shown in Figure 1. [Figure 3] This is a block diagram showing a configuration for implementing a charging location information presentation task in a driver assistance system according to a second embodiment of the present invention. [Figure 4A] This is a flowchart showing the first part of the charging location information presentation process by the driver assistance system according to the second embodiment of the present invention. [Figure 4B] This is a flowchart showing the second part of the charging location information presentation process by the driver assistance system according to the second embodiment of the present invention. [Modes for carrying out the invention]

[0015] A driver assistance system and a driver assistance method according to embodiments of the present invention will be described with reference to the accompanying drawings. Embodiments of the present invention are applicable to various vehicles such as fuel-powered vehicles, hybrid vehicles, and electric vehicles. In the following description, electric vehicles will be described, but the invention is not limited thereto. For this reason, the term "vehicle" will be used regardless of the type of vehicle.

[0016] In embodiments of the present invention, the in-vehicle system implements a driving range calculation task (first task) that calculates the driving range based on the remaining charge and energy consumption of a battery mounted on a vehicle such as an electric vehicle, and a charging location information presentation task (second task) that determines the need for charging based on the driving range and presents location information of charging stations, but is not limited to this. The in-vehicle system is equipped with at least a processor such as a CPU or GPU and semiconductor memory such as RAM or ROM. In addition, the in-vehicle system is equipped with a control mechanism which consists of multiple electronic control units (ECUs) connected by an in-vehicle network such as CAN for controlling the behavior of the vehicle (brake operation, accelerator operation, etc.), and electrical components such as a driver monitoring system with an in-vehicle camera capable of capturing images of the driver and a navigation device linked with GPS.

[0017] (1) First Embodiment In the first embodiment of the present invention, the driver assistance system uses machine learning to analyze trends related to fluctuations in the battery charge level due to the vehicle's ambient temperature and usage patterns. The driver assistance system also identifies the driver actually operating the vehicle from the driver's behavior detected by a driver monitoring system (such as an in-vehicle camera) installed in the vehicle (hereinafter referred to as "driver monitor information"), and uses machine learning to analyze each driver's tendencies when operating the vehicle. Specifically, the driver assistance system implements a machine learning model that predicts fluctuations in energy consumption, such as fuel efficiency and electricity consumption, and recalculates the vehicle's cruising range for each vehicle and each driver based on the inference results of the machine learning model.

[0018] In the first embodiment, the characteristic quantities related to the energy consumption for each vehicle and the characteristic quantities related to the energy consumption for each driver are input into a machine learning model as input variables, and the machine learning model infers output variables related to energy consumption (such as fuel consumption and electricity cost). Thereafter, the cruising range is recalculated based on the inference result of the machine learning model.

[0019] Machine learning models related to driving support and autonomous driving are created by learning and training based on a large number of datasets by IT vendors and users (such as automobile manufacturers). In the first embodiment, the main focus is on accurately calculating the cruising range of the vehicle, and it is necessary to accurately calculate the fuel consumption / electricity cost that changes according to the characteristics of each vehicle and each driver. For this reason, a machine learning model for estimating the fuel consumption / electricity cost of the vehicle is generated, and then the machine learning model is installed in the driving support system. Usually, since the machine learning model uses a large number of parameters, the scale of the machine learning model becomes large. In order to implement a driving support system equipped with a machine learning model in a vehicle, it is necessary to reduce the weight or compress the machine learning model. Alternatively, it is conceivable to optimize the number of parameters according to a specific application and reduce the scale without sacrificing the accuracy of the machine learning model and implement it in a driving support system (such as an in-vehicle computer).

[0020] On the other hand, when the scale of the machine learning model is large and insufficient for the computing power of the processor and the memory capacity of the memory installed in the vehicle, it is also possible to store the machine learning model in an external server such as a data center. In this case, the machine learning model stored in the data center may be read out by high-speed communication such as 4G or 5G by the communication device installed in the vehicle and a predetermined task may be executed. That is, it may be possible to access the machine learning model stored in the data center from the vehicle, apply desired input variables, and recalculate the cruising range based on the inference result of the machine learning model.

[0021] FIG. 1 is a block diagram showing a configuration for implementing a cruising range calculation task (first task) in a driving support system according to a first embodiment of the present invention. In the first embodiment, the driving support system is an in-vehicle system 100 mounted on a vehicle. The in-vehicle system 100 includes at least a processor and a memory, and by the processor executing a predetermined program stored in the memory, an input unit 110, a processing unit 120, and an output unit 130 are implemented. In a normal vehicle, the cruising range is simply calculated by multiplying the remaining gasoline amount by the fuel consumption, or multiplying the remaining battery charge by the electricity cost, and displayed on an in-vehicle meter or a display device.

[0022] The input unit 110 of the in-vehicle system 100 inputs vehicle characteristic data 111, driver characteristic data 112, and driver monitor information 113.

[0023] The vehicle characteristic data 111 exemplifies data indicating a tendency that affects the electricity cost for each vehicle. As the vehicle characteristic data 111, six data Va to Vf are listed. The data Va to Vf are automatically measured by various sensing devices mounted on the vehicle and input to the input unit 110 of the in-vehicle system 100. Note that the data Va to Vf indicate the situation of the vehicle that reduces the electricity cost.

[0024] Va: Daily usage situation of the vehicle. Vb: Deterioration state of the battery (such as a battery pack). Vc: Change in the remaining battery charge due to the outside air temperature. Vd: Change in the electricity cost due to the number of passengers. Ve: Change in the electricity cost due to the air pressure of the vehicle's wheels. Vf: Change in the electricity cost due to the usage situation of the air conditioner's heating and cooling in the vehicle.

[0025] Data Va indicates the frequency of use, recorded according to the time of day the vehicle is used (daily, weekly, etc.). The battery power in the vehicle is consumed not only when the vehicle is running or when electrical components are used, but the battery can also naturally discharge even when the vehicle is not in use. Therefore, the remaining charge fluctuates depending on the frequency of vehicle use, which affects the energy efficiency. If the vehicle is used repeatedly in a short period, the remaining charge decreases according to the frequency of vehicle use. On the other hand, even if the vehicle is not used for a long period, the remaining charge decreases due to natural discharge. Therefore, the driving range is recalculated according to the frequency of vehicle use.

[0026] Data Vb indicates that the battery degradation state differs depending on the vehicle, due to variations in the number of battery packs (battery modules) installed and the charge / discharge performance of each battery pack. Furthermore, the rate of change in remaining charge also differs depending on the vehicle's battery degradation state. Therefore, the cruising range is recalculated according to the vehicle's battery degradation state.

[0027] Data Vc indicates that the battery's charge and discharge characteristics change in response to changes in the vehicle's ambient temperature, resulting in different charge levels or battery degradation states even under the same vehicle usage conditions. In other words, the degree of change in charge level differs depending on the ambient temperature, regardless of vehicle usage, and consequently, the energy consumption also changes. Therefore, the driving range is recalculated according to the change in charge level due to the vehicle's ambient temperature.

[0028] Data Vd indicates that the rate at which the remaining battery charge is consumed differs depending on the number of passengers in the vehicle, and therefore the energy efficiency changes. The load on the vehicle's power source increases according to the vehicle's weight, the weight of the cargo, and the weight of the passengers. In other words, as the number of passengers increases, the remaining battery charge decreases more easily, and the energy efficiency decreases. For this reason, the driving range is recalculated according to the change in energy efficiency due to the number of passengers in the vehicle.

[0029] Data Ve indicates that the friction between the vehicle's wheels (tires, etc.) and the road surface during driving differs depending on the air pressure, which in turn affects the torque required to rotate the wheels, thus influencing the rate at which the battery charge decreases and thus changing the energy efficiency. For example, if the air pressure in the vehicle's wheels decreases, the contact friction between the wheels and the road surface increases, the torque required to rotate the wheels also increases, and consequently, the battery charge decreases more quickly. In other words, when the air pressure in the vehicle's wheels decreases, the battery charge decreases more easily, and the energy efficiency decreases. Therefore, the driving range is recalculated according to the change in energy efficiency due to the air pressure in the wheels.

[0030] Data Vf indicates that while the driver switches between heating and cooling modes on the vehicle's air conditioning system depending on the temperature and humidity inside the vehicle, the usage of the air conditioning system affects the rate at which the battery charge decreases, thus changing the energy efficiency. In other words, when the driver operates the air conditioning system to use heating or cooling, the battery charge decreases faster, and consequently, the energy efficiency decreases. Therefore, the driving range is recalculated according to the change in energy efficiency due to the usage of the vehicle's air conditioning system.

[0031] Driver characteristic data 12 exemplifies data showing tendencies that affect energy consumption for each driver. Five data points, Da through De, are included as driver characteristic data 112. Data points Da through De are automatically measured by various sensing devices and navigation systems installed in the vehicle and input to the input unit 110 of the in-vehicle system 100. Data points Da through Dc relate to changes in the vehicle's energy consumption, while data points Dd and De relate to the charging location and timing of the vehicle's battery. Da: Heating and cooling usage based on outside temperature. Db: The status of the vehicle's accelerator and brake pedals. Dc: The roads that vehicles normally travel on. Dd: The charging station that the driver normally uses. De: The driver's usual charging schedule.

[0032] Data Da indicates that the driver operates the air conditioning system to provide heating and cooling based on the vehicle's outside temperature, and that the use of heating and cooling affects the remaining battery charge, thus changing the energy efficiency. Heating consumes more energy than cooling, so the remaining battery charge decreases faster, and consequently, the energy efficiency decreases. Also, if the difference between the vehicle's outside temperature and the temperature inside the vehicle is large, the energy consumption for heating and cooling will increase. Therefore, the driving range is recalculated according to the change in energy efficiency due to the use of heating and cooling based on the vehicle's outside temperature.

[0033] Data Db indicates that the remaining battery charge changes in response to the driver's operation of the vehicle's accelerator and brake pedals, and therefore the energy consumption changes. The driver's operation of the accelerator and brake pedals is detected by a control mechanism consisting of multiple ECUs. When the driver operates the accelerator pedal, the rotation speed and torque of the wheels increase, causing the remaining battery charge to decrease more quickly and reducing energy consumption. On the other hand, when the driver operates the brake pedal, regenerative braking may be activated, where the electric motor functions as a generator to charge the battery during braking or deceleration. In this case, the amount of electricity generated by regenerative braking increases the remaining battery charge, and therefore the energy consumption may increase. For this reason, the driving range is recalculated according to the change in energy consumption caused by the driver's operation of the accelerator and brake pedals.

[0034] Data Dc indicates that the driver drives efficiently when driving on familiar roads, but not efficiently when driving on roads other than familiar ones, which affects the decrease in remaining charge and changes in energy consumption. "Familiar roads" refer to roads the driver frequently uses. These frequently used roads can be recorded by the navigation system or confirmed through applications providing map information and route search services. For example, roads driven several times a week can also be registered as frequently used roads. Therefore, the remaining range is recalculated based on changes in energy consumption, which are determined by whether or not the vehicle is traveling on frequently used roads, which are determined by factors such as the driver's location relative to their home and workplace.

[0035] Data Dd indicates the location information of the charging station that the driver normally uses when charging the vehicle's battery. Whether or not a charging station is normally used can be determined by recording the driver's usage of the same charging station in the navigation or map information. This allows the location information of the normally used charging station to be determined from the vehicle's current location, and this location information can be displayed on the vehicle's display device and used to provide the driver with charging location information.

[0036] Data De indicates the driver's usual battery charging timing when operating the vehicle. For example, if the driver tends to charge the battery when the remaining charge falls below 20% of full charge, the charging timing can be displayed on the vehicle's display device according to the remaining charge, and the driver can also be notified of the charging timing. It is conceivable that the driver may not pay attention to the remaining charge, causing the vehicle to run out of power and impeding driving. In such cases, notifying the driver of the charging timing can prevent the vehicle from running out of power due to a loss of remaining charge.

[0037] The driver monitor information 113 is used to identify the driver based on the driver's facial image and behavior detected by the driver monitoring system installed in the vehicle. When multiple drivers operate the same vehicle, the driver monitor information 113 can be used to identify the driver, and the driver characteristic data 112 associated with that driver can be selected.

[0038] The processing unit 120 implements machine learning (ML) and range recalculation (CAL). The machine learning (ML) acquires vehicle characteristic data 111, driver characteristic data 112, and driver monitor information 113 as three inputs a, b, and c, and performs predetermined calculations to calculate the energy consumption that reflects the tendencies of each vehicle and each driver. Subsequently, the processing unit 120 performs range recalculation (CAL) based on the energy consumption from the machine learning (ML) and the remaining charge of the vehicle.

[0039] The output unit 130 executes the cruising range output 131 and outputs the accurate cruising range recalculated by the cruising range recalculation CAL to a display device installed in the vehicle. Note that the destination of the cruising range output is not limited to the vehicle's display device; it may also be audible through the vehicle's speakers or wirelessly transmitted to a mobile device carried by the driver.

[0040] The processing unit 120 is equipped with a machine learning model M specifically designed for calculating fuel consumption. The machine learning model M incorporates numerous parameters and prediction algorithms, but it may also be generated through supervised learning using training data and target data.

[0041] As training data, data affecting electric efficiency and fuel consumption for each vehicle type is prepared. For example, factors affecting electric efficiency include the vehicle's external environment (such as ambient temperature), battery charge capacity, and power consumption of electrical components. Factors affecting fuel consumption include the vehicle's external environment (such as ambient temperature), gasoline tank capacity, and transmission performance (mechanical energy loss). As training data, data related to electric efficiency or fuel consumption obtained by measuring the distance traveled and the battery power consumption or gasoline fuel consumption when the vehicle is actually driven can be used. In other words, the internal parameters and calculation formulas (such as prediction algorithms) of the machine learning model M are adjusted so that the difference between the inference results output based on the training data and the training data is minimized.

[0042] Furthermore, the historical data of vehicle feature data 111 may be used as training data. Data related to electric power consumption or fuel consumption based on the driving history of each vehicle may be used as training data. The internal parameters and prediction algorithm are adjusted so that the difference between the prediction results of electric power consumption or fuel consumption for each vehicle by the machine learning model M and the training data is minimized.

[0043] In the above example, a vehicle-specific prediction algorithm was incorporated into the machine learning model M, but a driver-specific prediction algorithm may also be incorporated into the machine learning model M. For example, if multiple drivers are assigned to the same vehicle, the past history of driver characteristic data 112 for each driver may be used as training data, and the electric power consumption or fuel consumption when each driver actually drives the vehicle may be measured and used as training data. In other words, the internal parameters of the machine learning model M and the prediction algorithm are adjusted so that the difference between the inference results output based on the training data and the training data is minimized.

[0044] Two configurations are possible for the machine learning ML to be implemented in the processing unit 120 in Figure 1. Figure 2(A) shows machine learning ML1, which utilizes the inference function of the machine learning model M, and Figure 2(B) shows machine learning ML2, which utilizes the retraining (tuning) of the machine learning model M.

[0045] First, let's explain the machine learning model ML1 shown in Figure 2(A). The machine learning model M generates one output variable based on two input variables. In Figure 2(A), the vehicle feature FQ1 and driver feature FQ2 are input to the machine learning model M. The vehicle feature data 111 includes multiple data points from Va to Vf, but each data point has different units and measurement criteria. Therefore, the vehicle feature data 111 cannot be directly input to the machine learning model M, and preprocessing for machine learning is necessary. In other words, features that affect the variation in energy consumption are extracted from the vehicle feature data 111 and converted into an input format for machine learning related to vehicles to generate the vehicle feature FQ1. For example, selected data from data points Va to Vf are quantified or normalized, and a weighted average or median is calculated to determine the vehicle feature FQ1.

[0046] Similarly, the driver feature data 112 includes multiple data points Da through Dc related to fuel consumption, but since the criteria for each data point are different, the driver feature data 112 cannot be directly input into the machine learning model M. Therefore, as a preprocessing step for machine learning, features that affect the fluctuation of fuel consumption are extracted from the driver feature data 112 and converted into an input format for machine learning about drivers to generate driver feature FQ2. For example, selected data from data Da through Dc are quantified or normalized, and weighted averages or medians are calculated to obtain the driver feature FQ2. The driver feature FQ2 is generated based on the driver feature data 112 related to the driver identified by the driver monitor information 113.

[0047] In machine learning ML1, the inference function of a pre-generated machine learning model M, which is trained on a dataset related to vehicle energy consumption, is utilized. Specifically, vehicle features FQ1 and driver features FQ2 are input to the machine learning model M, and the energy consumption is calculated as an inference result of the machine learning model M.

[0048] Typically, energy consumption is calculated using a simplified method based on the vehicle's mileage and battery power consumption, which does not reflect the tendencies of individual vehicles or drivers. Therefore, the vehicle's remaining range is simply calculated by multiplying the energy consumption by the remaining battery charge.

[0049] In this embodiment, in order to accurately determine the remaining driving range, vehicle feature data 111 and driver feature data 112 are acquired, and vehicle feature quantity FQ1 and driver feature quantity FQ2 are calculated and applied to the machine learning model M. Therefore, the energy consumption output by the machine learning model M's inference reflects the current behavior of the vehicle and driver, compared to the energy consumption obtained by a simple calculation of past vehicle mileage and power consumption.

[0050] The processing unit 120 performs a cruising range recalculation CAL, accurately recalculating the cruising range based on the power consumption and remaining charge calculated by the machine learning model M's inference. The remaining charge is measured by a voltmeter and an ammeter or power meter connected to the battery.

[0051] Next, we will explain the machine learning ML2 shown in Figure 2(B). In machine learning ML2, retraining (tuning) TN is performed on the machine learning model M using the vehicle dataset DS1 and the driver dataset DS2 for retraining. The machine learning model M performs inference (e.g., prediction) using a large number of parameters, and after learning and training using a large number of datasets, inference is performed using the machine learning model M. In this embodiment of the present invention, the main focus is on reflecting the tendencies of each vehicle and each driver in machine learning, so it is conceivable to retrain the machine learning model M to reflect the vehicle feature data 111 and the driver feature data 112.

[0052] Specifically, the retraining (tuning) TN is provided with a vehicle dataset DS1 for retraining corresponding to the vehicle feature data 111, and a driver dataset DS2 for retraining corresponding to the driver feature data 112 related to the driver identified in the driver monitor information 113. The vehicle dataset DS1 shows, for example, a tendency that the electric power consumption decreases when the battery deteriorates in data Vb. The driver dataset DS2 shows, for example, a tendency that the electric power consumption decreases when the driver activates the air conditioner's heating function due to a drop in outside temperature in data Db, as the remaining charge tends to decrease.

[0053] After retraining the TN model to reflect the tendencies of each vehicle and each driver in the machine learning model M, the machine learning model M's inference calculates the energy consumption for each vehicle and each driver. This allows for a more accurate recalculation of the remaining driving range.

[0054] It is not necessary to install the machine learning model (ML) shown in Figure 1 into the in-vehicle system 100 and complete the vehicle's remaining range calculation task within the in-vehicle system 100. If the machine learning model (ML) program and data volume are large and cannot be processed by the processing power of the processor installed in the in-vehicle system 100, the machine learning model (ML) may be implemented in a data center or similar location. In this case, the in-vehicle system 100 will have a communication function with the data center and will transmit vehicle feature data 111 and driver feature data 112 related to the driver identified by driver monitor information 113 to the data center. In the data center, as shown in Figure 2(A), the machine learning model M will perform inference of fuel consumption and transmit it to the in-vehicle system 100. Alternatively, as shown in Figure 2(B), the machine learning model M will be retrained (tuned) in the data center and the fuel consumption inferred by the retrained machine learning model M will be transmitted to the in-vehicle system 100. Subsequently, the processing unit 120 of the in-vehicle system 100 performs a cruising range recalculation CAL based on the fuel consumption inferred by machine learning (ML) and the remaining charge of the vehicle's battery.

[0055] (2) Second Embodiment In the second embodiment of the present invention, the driver's driving behavior is determined as follows: for example, whether the road the vehicle is currently traveling on is a road the driver usually drives on, whether the driver has set a destination, whether the destination is within the vehicle's driving range and whether there are charging stations near the vehicle's current location or near the destination, and the driver is notified of the location information of the charging stations at a timing that matches the driver's driving situation.

[0056] Figure 3 is a block diagram showing the configuration for implementing a charging location information presentation task (second task) in a driver assistance system according to the second embodiment of the present invention. In the second embodiment, the driver assistance system is an in-vehicle system 200 mounted on a vehicle, but it is not limited thereto. The in-vehicle system 200 comprises at least a processor and memory, and the processor implements an input unit 210, a processing unit 220, and an output unit 230 by executing a predetermined program stored in the memory.

[0057] The input unit 210 accepts multiple pieces of information. For example, in destination setting 211, the driver sets the destination and the location information of the destination is set. Driving range 212 indicates the accurate driving range recalculated in the first embodiment. Current location information 213 indicates the vehicle's current location. Charging station information 214 indicates the location information of charging stations located around or near the vehicle. If multiple charging stations exist around the vehicle, multiple location information for multiple charging stations is input as charging station information 214. Note that charging stations are searched for using navigation information and map information.

[0058] The processing unit 220 performs two types of determination processes. The first determination 221 is to determine whether or not there are charging stations around the vehicle. That is, based on the vehicle's current location information 213 and charging station information 214, it determines whether or not there are charging stations within a predetermined distance from the vehicle's current location. For this reason, it is possible that multiple charging stations exist around the vehicle.

[0059] The second determination 222 of the processing unit 220 is to determine whether a charging station is necessary. That is, based on the destination information 211 and the remaining driving range 212, it determines whether a charging station is necessary if one exists near the vehicle. For example, if a charging station exists within the vehicle's driving range along the route from the vehicle's current location to the destination, the driver is offered a charging station. On the other hand, if a charging station exists within the vehicle's driving range but is off the route from the vehicle's current location to the destination, the driver is not necessarily offered a charging station. The logical structure for deciding whether or not to offer a charging station will be explained in detail with reference to the flowchart in Figure 4.

[0060] The output unit 230 makes a suggestion 231 for a charging station. If the processing unit 220's second determination 222 determines that there is a need to suggest a charging station, the output unit 230 suggests the location information of the charging station to the driver. Specifically, the location information of the charging station is displayed on the display device.

[0061] Next, the details of the charging location information presentation process, which is implemented in the processing unit 220 of the in-vehicle system 200, will be explained with reference to the flowcharts shown in Figures 4A and 4B. Figure 4A consists of steps S1-S6 and S12-S18, and Figure 4B consists of steps S7-S11.

[0062] As shown in Figure 4A, the processing unit 220 acquires current location information 213 from the input unit 210 and determines whether the vehicle's current location is a place the driver has driven before (steps S1, S2). If the vehicle's current location is not a place the driver has driven before, it determines whether the driver has set a destination 211 (step S3). If the driver has set a destination but the destination is not within the driving range, the processing unit 220 creates an appropriate charging plan and displays the route to a charging station in the vicinity of the vehicle on a display device or the like (steps S4, S5). An "appropriate charging plan" refers to a plan in which the battery is charged at least once when the distance from the vehicle's current location to the destination is greater than the driving range. For this reason, the driver is notified of the need to charge and is offered information on the location of charging stations in the vicinity of the vehicle.

[0063] On the other hand, if the destination is within the cruising range, the processing unit 220 calculates the battery's power consumption based on the distance from the vehicle's current location to the destination and the power consumption, and predicts the remaining charge when the vehicle arrives at the destination (steps S4, S6). In this case, the processing unit 220 may present the predicted remaining charge to the driver. This is to inform the driver of the need to recharge the battery after arriving at the destination.

[0064] If the vehicle has traveled to its current location before, the processing unit 220 determines whether the vehicle's current location corresponds to a road that it frequently travels on (steps S2, S7). It also determines whether a place the driver frequently visits (for example, a destination the driver visits on a daily basis) is reachable within the vehicle's driving range (step S8). If a frequently visited place exists within the vehicle's driving range, the processing unit 220 predicts the remaining battery charge upon arrival at the destination (step S9).

[0065] Subsequently, the processing unit 220 determines whether or not there is a possibility of running out of power between charges, based on the vehicle's normal charging pattern (step S10). "Normal charging pattern" refers to a habitual charging pattern, for example, where the driver charges the battery when the remaining charge level of the vehicle's battery reaches 20% of a full charge. Therefore, "possibility of running out of power before the next charge" means that, for example, when the remaining charge level of the battery approaches 20% of a full charge, the processing unit 220 determines that there is a possibility of running out of power before the next charge. The driver may also set 20% of a full charge as a threshold for deciding whether or not to charge, taking into account the possibility of running out of power.

[0066] If there is no possibility of the vehicle running out of power, the processing unit 220 searches for charging stations around the vehicle's current location and, if necessary, displays the location information of the charging stations on the display device (steps S10, S11). "Searching for charging stations around the vehicle's current location" means that the processing unit 220 searches for charging stations located around the vehicle's current location by referring to navigation information and map information.

[0067] The processing unit 220 proceeds to step S12 or later, as shown in Figure 4B, in four cases (a) through (d). These are: (a) when the driver has not set a destination 211 (step S3, NO), (b) when the vehicle's current location is not on a frequently traveled road (step S7, NO), (c) when the vehicle's current location is on a frequently traveled road, but it cannot reach a frequently visited destination within the cruising range (step S8, NO), and (d) when there is a possibility of running out of power between charges based on the vehicle's normal charging pattern (step S10, YES). In these cases, the remaining charge of the vehicle's battery decreases, and there is a possibility of running out of power before reaching the predetermined location.

[0068] In step S12, the processing unit 220 determines whether or not there is a charging station near the vehicle's current location. If there is a charging station near the vehicle's current location, the processing unit 220 displays information about the charging station on the navigation device (step S18).

[0069] If there are no charging stations near the vehicle's current location and no charging stations within the vehicle's driving range, the processing unit 220 suggests to the driver to go directly to a charging station (steps S12, S13, S14). This is because the vehicle is likely to run out of power, and the processing unit 220 is intended to prevent this. The processing unit 220 may also suggest to the driver a driving mode that improves the vehicle's energy efficiency (step S15). For example, the driver may reduce the power consumed by means other than driving the wheels by stopping the operation of electrical equipment such as the air conditioner to prevent a decrease in the battery charge level.

[0070] If a charging station exists within the vehicle's driving range and the vehicle's battery charge level is 20% or more of a full charge, the processing unit 220 displays information about the charging station on the navigation system (steps S13, S16, S18). On the other hand, if the vehicle's battery charge level is 20% or less, the processing unit 220 suggests to the driver to go directly to a charging station (steps S16, S17). This is because the vehicle's battery charge level is low and there is a possibility of running out of power.

[0071] In the above embodiment, the cruising range was recalculated according to the electricity consumption which varies for each electric vehicle and each driver, but the vehicle is not limited to electric vehicles. Even with fuel-powered vehicles such as gasoline cars, the fuel consumption varies for each vehicle and each driver, so machine learning (ML) may be performed to estimate the fuel consumption by reflecting the measurement data for each vehicle and each driver that affects the fuel consumption, and the cruising range recalculation (CAL) may be performed based on that fuel consumption and the amount of gasoline remaining as measured by the fuel gauge in the gasoline tank.

[0072] For example, in the vehicle feature data 111 of the first embodiment, data Va to Vf were used, which may affect the energy consumption of an electric vehicle. In a fuel-powered vehicle, it is conceivable to redefine data Va to Vf as data that affects fuel efficiency. As data Va, the daily usage conditions of the vehicle may be measured, taking into account that the amount of fuel (gasoline, etc.) consumed differs depending on the frequency of vehicle use, and that fuel decreases volatilically even when the vehicle is not in use. As data Vb, the deterioration state of the drive system (gasoline tank, internal combustion engine, transmission, etc.) of a fuel-powered vehicle may be measured. As data Vc, the change in the amount of fuel remaining in the gasoline tank due to the outside temperature may be measured. As data Vd, the change in fuel efficiency may reflect the change in the overall weight of the vehicle due to the number of passengers. As data Ve, the change in fuel efficiency due to the air pressure of the vehicle's wheels may be measured. When the air pressure of the wheels decreases, the frictional force between the wheels and the road surface increases, and thus fuel efficiency decreases. As data Vf, the change in fuel efficiency due to the use of the heating and cooling of the vehicle's air conditioner may be measured. In fuel-powered vehicles, the alternator converts the engine's rotational force into electrical energy to charge batteries for electrical components such as lead-acid batteries. However, the specific gravity of the battery's electrolyte changes with charging and discharging. Furthermore, the energy consumption differs depending on the drive system between the alternator and the battery, which in turn affects the overall fuel efficiency of the vehicle.

[0073] In the driver characteristic data 112 of the first embodiment, data Da to Dc were used, which may affect the energy consumption of an electric vehicle. In a fuel-powered vehicle, it is conceivable to redefine data Da to Dc as data that affects fuel efficiency. Data Da shows that energy consumption differs depending on the alternator and battery, depending on the use of heating and cooling due to the outside temperature of the vehicle, and thus affects the overall fuel efficiency of the vehicle. Data Db shows that energy consumption differs depending on the driver's operation of the accelerator and brake pedals, and thus affects fuel efficiency. Data Dc shows that on roads the vehicle normally travels, the driver tends to drive efficiently, so the fuel consumption tends to be less compared to other roads, and thus affects fuel efficiency. Data Dd is redefined as the gas station that the driver normally uses, and data De is redefined as the driver's usual refueling timing.

[0074] When the second embodiment is applied to a fuel-powered vehicle, it is necessary to modify the in-vehicle system 200 shown in Figure 3. The input unit 210 receives charging station information 214, such as the location information of a gas station. The processing unit 220 performs a first determination 221, which determines whether or not there is a gas station near the vehicle. The processing unit 220 performs a second determination 222, which determines whether or not it is necessary to suggest a refueling station. The output unit 230 suggests a gas station instead of a charging station suggestion 231. Also, in the flowchart in Figure 4, "electricity consumption" is replaced with "fuel efficiency," and "charging station" is replaced with "gas station."

[0075] Next, the purpose and effects of the present invention will be explained. First, the configuration, functions, and effects of the driver assistance systems (in-vehicle systems 100, 200) will be described.

[0076] The driver assistance system (100) comprises an input unit (110) that identifies the driver of the vehicle and inputs driver characteristic data (112) indicating the driver's operation of the vehicle, and vehicle characteristic data (111) related to the vehicle's electric energy consumption or fuel consumption; a processing unit (120) that performs machine learning (ML) based on the driver characteristic data (112) and vehicle characteristic data (111) to estimate the electric energy consumption / fuel consumption associated with the driver and the vehicle, and recalculates the cruising range (CAL) based on the electric energy consumption / fuel consumption and the vehicle's remaining charge / fuel level; and an output unit (130) that outputs the cruising range.

[0077] The driver monitoring system installed in the vehicle identifies the driver in the driver's seat. In other words, the driver is identified by driver monitoring information (113). The input unit (110) acquires driver characteristic data (112) as a driver tendency, which shows the vehicle's operation status, such as the amount and timing of operation of the accelerator pedal and brake pedal, and the operation status of the air conditioner. The input unit (110) also acquires vehicle characteristic data (111) as a tendency related to the vehicle's electricity consumption and fuel efficiency, such as the battery degradation status, the usage status of the air conditioner, and the tendency of the remaining charge to decrease depending on the number of passengers.

[0078] The processing unit (120) performs machine learning (ML) based on driver characteristic data (112) and vehicle characteristic data (111) to estimate the electric power consumption / fuel consumption associated with the driver and vehicle. Subsequently, the processing unit (120) recalculates the driving range (CAL) based on the electric power consumption / fuel consumption and the vehicle's remaining charge / fuel level. The remaining charge is measured by a power meter on the vehicle's battery, and the remaining fuel is determined by measuring the amount of fuel in the vehicle's gasoline tank. The driving range recalculated by the processing unit (120) is output via the output unit (130). Possible destinations for the driving range output include, for example, display devices such as navigation systems and vehicle meters.

[0079] Fuel efficiency (electricity / fuel consumption) fluctuates depending on the driver's driving habits and the vehicle's operating conditions. In conventional vehicles, the driving range is calculated by multiplying the vehicle's remaining charge / fuel consumption by the vehicle's remaining charge / fuel consumption, based on the vehicle's past operating conditions. In the embodiment of the present invention, the driver's tendencies regarding fuel efficiency and the vehicle's tendencies regarding fuel efficiency are learned through machine learning, and the vehicle and driver's associated fuel efficiency are estimated based on the results of this machine learning. This allows for a more accurate recalculation of the driving range based on the fuel efficiency obtained through machine learning and the vehicle's remaining charge / fuel consumption. In other words, in this embodiment, the discrepancy between the initially calculated driving range and the recalculated driving range is reduced, and the accuracy of the driving range calculation is improved. This allows the driver to confidently plan charging locations / refueling locations and timings.

[0080] The machine learning (ML) performed in the processing unit (120) may utilize a machine learning model (M) generated by learning the past history of driver characteristic data (112) and the past history of vehicle characteristic data (111). For example, the processing unit (120) may input the current driver characteristic data (112) and the current vehicle characteristic data (111) into the machine learning model (M) to estimate the electric power consumption / fuel consumption. Alternatively, the processing unit (120) may retrain (TN) the machine learning model (M) with the current driver characteristic data (112) and the current vehicle characteristic data (111). The electric power consumption / fuel consumption associated with the driver and vehicle may then be estimated using the retrained machine learning model (M).

[0081] The processing unit (120) may estimate the electric power consumption / fuel consumption associated with the driver and the vehicle for each combination of multiple conditions, including the outside temperature and humidity of the vehicle, and the season and time of day when the vehicle is being driven. Depending on the outside temperature and humidity of the vehicle, the operating status of the air conditioner installed in the vehicle changes, and as a result, the rate of decrease in charge / fuel consumption differs. If the outside temperature of the vehicle is high but the humidity is low, the driver is less likely to use the air conditioner's cooling function, and if the humidity is high, the driver is more likely to use the air conditioner's cooling function. For this reason, the electric power consumption / fuel consumption differs depending on the combination of the outside temperature and humidity of the vehicle. Japan has four seasons, so the air conditioner's cooling function is used in the summer and the air conditioner's heating function is used in the winter. Furthermore, the usage of the air conditioner and lighting (lights) also differs depending on the time of day. For example, at night, the outside temperature drops, so the air conditioner's heating function is used and the lights are turned on, which tends to decrease the electric power consumption / fuel consumption. In this way, the electric power consumption / fuel consumption differs depending on the combination of season and time of day.

[0082] The processing unit (220) estimates the charging / refueling timing based on the electric power consumption / fuel consumption and the vehicle's remaining charge / fuel level, and the output unit (230) may propose charging / refueling information to the driver, indicating the charging / refueling timing and the vehicle's charging / refueling location.

[0083] In the above, the processing unit (220) may learn the driver's tendencies and the vehicle's driving tendencies, and estimate the charging / refueling timing based on the learning results. For example, if the driver's tendency is to move to a charging location and charge the vehicle's battery when the remaining charge drops to 20% of full charge, the system may estimate the charging timing based on the point when the remaining charge drops to 20% of full charge. This allows the system to propose charging / refueling information that the driver desires. In addition, since the vehicle's remaining range is accurately recalculated, the possibility of the driver assistance system (200) proposing charging / refueling information to the driver in an inappropriate location can be reduced.

[0084] The processing unit (220) refers to pre-recorded map information to determine whether the driver is driving the vehicle on a predetermined route that is frequently used by the driver. If the vehicle is traveling on a different route, the output unit (230) may propose charging / refueling information to the driver earlier than when the vehicle is traveling along the predetermined route.

[0085] In the above, "a predetermined route frequently used by the driver" refers to a route that the vehicle frequently travels on, where the same driving history is repeatedly recorded. For example, if a driver uses that route several days a week, it will be recorded as a predetermined route frequently used by the driver in the vehicle's computer.

[0086] When a vehicle travels on a route different from its designated route and has no prior driving history, it is assumed that the driver may not be aware of charging / refueling locations. In such cases, the driver may have difficulty finding a charging / refueling location at the appropriate time. To alleviate driver anxiety regarding charging / refueling and ensure that the vehicle can be driven safely without running out of battery or gas, the timing of suggesting charging / refueling information is advanced when the vehicle travels on a route without prior driving history. For example, when the vehicle travels on its designated route, charging / refueling information is suggested when the battery level reaches 20% of a full charge. On the other hand, when the vehicle travels on a route without prior driving history, charging / refueling information may be suggested when the battery level reaches 30% of a full charge.

[0087] The processing unit (220) acquires information about charging / refueling locations that the driver has used before, and when charging / refueling at another charging / refueling location where the vehicle has not been used before, the output unit (230) may propose charging / refueling information about the other charging / refueling location to the driver earlier than when charging / refueling at a charging / refueling location that the driver has used before.

[0088] Drivers may travel on various routes, including routes they frequently travel, routes they travel infrequently (e.g., only a few times), or routes they have never traveled before. If a driver knows the locations of charging / refueling stations regardless of the route, they can reach them without getting lost. However, when a driver travels on a route where they are unfamiliar with charging / refueling stations, they need to search for stations they have never used before. For example, a driver might stop the vehicle and search for charging / refueling stations by referring to map information such as a navigation system. In this case, it might be beneficial to suggest charging / refueling information earlier to alleviate the driver's anxiety and allow them to confidently search for unfamiliar charging / refueling stations.

[0089] Even on routes that the driver frequently uses and has a driving history, there may be charging / refueling stations that the driver has used before and charging / refueling stations that they have not used before. Therefore, the charging / refueling stations suggested by the driver assistance system (200) are not necessarily charging / refueling stations that the driver has used before. Even on routes that the driver has a driving history, the driver needs to choose between charging / refueling stations that they have used before and charging / refueling stations that they have not used before. In this case, the driver needs to appropriately select charging / refueling stations and plan accordingly. For this reason, the driver assistance system (200) has accelerated the timing of suggesting charging / refueling information regarding charging / refueling stations that the driver has not used before. Furthermore, if the driver assistance system (200) repeatedly suggests charging / refueling stations that the driver has not used before on routes with a driving history, it may set an upper limit on the number of suggestions, such as limiting the number of suggestions to a few times.

[0090] In the processing unit (220), if the driver has set a destination for the vehicle, and a threshold has been set in advance in relation to the remaining driving range, and a predetermined charging / refueling station is located closest to the destination on the route from the vehicle's current location to the destination, and there are no other charging / refueling stations within the distance related to the threshold for the vehicle's driving range, the output unit (230) may inform the driver that the predetermined charging / refueling station may be the final charging / refueling point.

[0091] In the above, the driver may set the vehicle's destination by specifying it in a navigation system, for example. In that case, the navigation system will automatically calculate the route from the vehicle's current location to the destination. Furthermore, the threshold related to the remaining driving range may be set to 20 km if the remaining driving range is 50 km. In other words, the driver may set the threshold so as to satisfy "remaining driving range (km) > threshold (km)".

[0092] The designated charging / refueling location closest to the destination refers to, for example, the charging / refueling location closest to the destination among multiple charging / refueling locations that may exist along the route from the vehicle's current location to the destination when the destination is set in the navigation system. When the vehicle passes this charging / refueling location, there are no other charging / refueling locations on the route to the destination. In addition, regardless of the destination, if there are no other charging / refueling locations within a threshold distance smaller than the vehicle's remaining range, the above charging / refueling location may become the final charging / refueling point, and the driver is notified accordingly.

[0093] The driver can understand that the final charging / refueling opportunity may be at a location close to the destination, rather than at the usual charging / refueling timing based on the remaining driving range. Therefore, the driver can charge / refuel the vehicle with peace of mind. Furthermore, in this embodiment, the accuracy of estimating the vehicle's remaining driving range has been improved, so the driver can set the driving range threshold more precisely.

[0094] Next, I will briefly explain the driver assistance methods. In the driver assistance method, the driver of the vehicle is identified, driver characteristic data (112) indicating the driver's operation of the vehicle and vehicle characteristic data (111) related to the vehicle's electric power consumption / fuel consumption are input, machine learning (ML) is performed based on the driver characteristic data (112) and vehicle characteristic data (111) to estimate the electric power consumption / fuel consumption associated with the driver and the vehicle, the driving range is recalculated (CAL) based on the electric power consumption / fuel consumption and the vehicle's remaining charge / fuel level, and the driving range is output. This driver assistance method corresponds to a driver assistance system (100).

[0095] In the above, based on the electric power consumption / fuel consumption and the vehicle's remaining charge / fuel level associated with the driver and vehicle, the timing of charging / refueling may be estimated, and charging / refueling information indicating the timing of charging / refueling and the location of charging / refueling for the vehicle may be proposed to the driver. This corresponds to a driver assistance system (200).

[0096] While several embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These novel embodiments can be carried out in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]

[0097] 100 In-vehicle systems 110 Input Section 111 Vehicle Feature Data 112 Driver Characteristics Data 113 Driver Monitor Information 120 Processing Unit 130 Output section 131 Range Output 200 In-vehicle systems 210 Input section 211 Destination setting 212 cruising range 213 Current Location Information 214 Charging Station Information 220 processing units 221 First determination (determination of charging stations around the vehicle) 222 Second Judgment (Determination of the necessity of proposing a charging station) 230 Output section 231 Proposal for a charging stand CAL Range Recalculation DS1 Vehicle Dataset for Retraining DS2 Driver Dataset for Retraining FQ1 Vehicle Features FQ2 Driver Characteristics M Machine Learning Model ML, ML1, ML2 Machine Learning TN retraining (tuning)

Claims

1. An input unit that identifies the driver of a vehicle and inputs driver characteristic data indicating the driver's operation of the vehicle, and vehicle characteristic data related to the vehicle's electric power consumption or fuel efficiency, A processing unit that performs machine learning based on the driver characteristic data and the vehicle characteristic data to estimate the electric power consumption or fuel consumption associated with the driver and the vehicle, and recalculates the cruising range based on the electric power consumption or fuel consumption and the remaining charge or fuel level of the vehicle, An output unit that outputs the aforementioned cruising range, A driver assistance system characterized by being equipped with the following.

2. The driver assistance system according to claim 1, characterized in that the processing unit is equipped with a machine learning model generated by learning the past history of the driver characteristic data and the past history of the vehicle characteristic data, and the current driver characteristic data and the current vehicle characteristic data are input to the machine learning model to estimate the electric power consumption or the fuel consumption.

3. The driver assistance system according to claim 1, characterized in that the processing unit retrains a machine learning model generated by learning the past history of the driver characteristic data and the past history of the vehicle characteristic data with the current driver characteristic data and the current vehicle characteristic data, and estimates the electric power consumption or fuel consumption associated with the driver and the vehicle using the machine learning model after the retraining.

4. The driving support system according to claim 1, characterized in that the processing unit estimates the electric power consumption or fuel consumption associated with the driver and the vehicle for each combination of multiple conditions consisting of the outside temperature and humidity of the vehicle, and the season and time of day when the vehicle is driven.

5. The driver assistance system according to claim 1, characterized in that the processing unit estimates a charging timing or refueling timing based on the electric power consumption or fuel consumption associated with the driver and the vehicle and the remaining charge or fuel level of the vehicle, and the output unit proposes charging / refueling information to the driver indicating the charging timing or refueling timing and the charging location or refueling location of the vehicle.

6. The driver assistance system according to claim 5, characterized in that the processing unit determines whether the driver is driving the vehicle on a predetermined route that is frequently used by the driver by referring to pre-recorded map information, and when the vehicle is traveling on another route, the output unit proposes the charging / refueling information to the driver at an earlier timing than when the vehicle is traveling along the predetermined route.

7. The driver assistance system according to claim 5, characterized in that the processing unit acquires information regarding the charging station or refueling station that the driver has used before, and when the vehicle is charged or refueled at another charging station or refueling station that the driver has not used before, the output unit proposes the charging / refueling information regarding the other charging station or refueling station to the driver at an earlier timing than when the vehicle is charged or refueled at the charging station or refueling station that the driver has used before.

8. The driver assistance system according to claim 5, wherein the processing unit has set a destination for the vehicle by the driver, and a threshold has been set in advance in relation to the remaining driving range, and a predetermined charging station or predetermined refueling station is located closest to the destination on the route from the vehicle's current location to the destination, and there are no other charging stations or other refueling stations within the distance related to the threshold of the vehicle's driving range, the output unit informs the driver that the predetermined charging station or predetermined refueling station may be the final charging / refueling point.

9. The system identifies the driver of the vehicle, inputs driver characteristic data indicating the driver's operation of the vehicle, and inputs vehicle characteristic data related to the vehicle's electric power consumption or fuel efficiency. Machine learning is performed based on the driver characteristic data and the vehicle characteristic data to estimate the electric power consumption or fuel consumption associated with the driver and the vehicle. The driving range is recalculated based on the aforementioned electricity consumption or fuel consumption and the remaining charge or fuel level of the vehicle. A driving assistance method characterized by outputting the cruising range.

10. Based on the electric power consumption or fuel consumption associated with the driver and the vehicle, and the remaining charge or fuel level of the vehicle, the charging timing or refueling timing is estimated. The driving assistance method according to claim 9, characterized in that it proposes charging / refueling information to the driver indicating the charging timing or refueling timing and the charging location or refueling location of the vehicle.

Images