Movement performance determination device of vehicle
The vehicle motion performance determination device uses electrical state detection and machine learning to accurately assess and enhance driving performance by adjusting the vehicle's electrical state, addressing the complexity of the electrical state-performance correlation.
Patent Information
- Application Number
- JP2023075149
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-04-28
- Publication Date
- 2024-11-08
AI Technical Summary
The correlation between the electrical state of a vehicle body and its dynamic performance is complex, making it difficult to estimate and improve driving performance effectively.
A vehicle motion performance determination device that uses electrical state detection means and a machine learning algorithm to determine the degree of high driving performance based on detected electrical state values, allowing for accurate evaluation and adjustment of the vehicle's electrical state to enhance maneuverability.
Enables quick evaluation of driving performance corresponding to the vehicle's electrical state, eliminating trial-and-error processes and providing valuable information for vehicle design and adjustment to improve aerodynamic performance.
Smart Images

Figure 2024159277000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an apparatus for determining the dynamic performance of a vehicle such as an automobile, and more particularly to an apparatus for determining the dynamic performance of a vehicle that changes depending on the electrical state of the vehicle. [Background technology]
[0002] As shown in Patent Document 1 by the applicants of the present application, a phenomenon is observed in which shaking and vibration are suppressed in areas where electrons are supplied on the inside and outside surfaces of a vehicle, and shaking and vibration are permitted in areas where electrons are removed. Therefore, the applicants of the present application have proposed a device that utilizes this phenomenon to control the aerodynamic characteristics of a vehicle by actively supplying or removing electrons by connecting the positive and negative poles of a power source to any part of the vehicle. According to such a vehicle aerodynamic characteristic control device, it is possible to improve the running performance or driving performance by appropriately controlling the distribution of electric potential or current in the vehicle. In relation to the electric potential or current on a vehicle, Patent Document 2 discloses a method for determining an earth point to which an electric power system earth cable is connected and an earth point to which a signal system earth cable is connected among multiple earth points on the vehicle. Patent Document 3 discloses a vehicle current sensor that measures the current value in a current path connected to the vehicle battery. Patent Document 4 discloses a technology for ensuring the safety of electric wires in a vehicle. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Patent Publication No. 2022-143555 [Patent Document 2] Patent Publication 2015-147487 [Patent Document 3] Patent Publication 2018-155625 [Patent Document 4] U.S. Patent Publication 2012 / 0134062 Summary of the Invention [Problem to be solved by the invention]
[0004] As shown in the above Patent Document 1, it has been found that the distribution of electric potential or current at the inner and outer surface parts of a vehicle (hereinafter referred to as the "electrical state of the vehicle body") is correlated with the vehicle's dynamic performance (running performance, drivability). Therefore, it is expected that the vehicle's dynamic performance can be estimated by detecting the electrical state of the vehicle body. If the vehicle's dynamic performance can be estimated, it will be possible to control the electrical state of the vehicle body so as to improve the vehicle's dynamic performance based on the estimation result. In this regard, in the development research of the present inventors, it has been found that, although the correlation between the electrical state of the vehicle body and the vehicle's dynamic performance is somewhat complicated, a discriminator configured according to a machine learning algorithm using electric potentials or currents (electrical state values) detected at various parts of a vehicle driven under various electrical states as input data and the vehicle's dynamic performance in each electrical state as correct answer data can determine the vehicle's dynamic performance based on the electrical state value. This knowledge is utilized in the present invention.
[0005] SUMMARY OF THE PRESENT EMBODIMENTS It is therefore an object of the present invention to provide an apparatus for detecting an electrical state of a vehicle body and judging the vehicle's dynamic performance based on the detected value. [Means for solving the problem]
[0006] According to the present invention, the above problem is solved by a vehicle dynamics performance determination device, an electrical condition detection means for detecting an electrical condition value in at least one portion of the vehicle; a driving performance level determination means for determining a level of driving performance of the vehicle based on the electric state value detected by the electric state detection means; Including, The driving performance level determination means uses electrical state values detected in a vehicle running under various electrical states and the degree of driving performance obtained in each of the electrical states as learning data, and includes a discrimination means that has learned according to a machine learning algorithm so as to output the degree of driving performance corresponding to an electrical state value when the electrical state value in the learning data is input, and is configured to determine that the degree of driving performance output by the discrimination means for the electrical state value obtained by the electrical state detection means is the degree of driving performance when that electrical state value was obtained. This is achieved by:
[0007] In the above configuration, the "electrical state value" may be a potential value or a current value at at least one portion of the vehicle. Therefore, the "electrical state detection means" may be a voltage sensor or a current sensor provided at at least one portion of the vehicle. The "at least one portion" where the electrical state value is detected may be selected as appropriate, and may be, for example, the front or rear of the vehicle body, a pillar, etc. The "degree of high driving performance of the vehicle" may be set arbitrarily, and may be set, for example, in a plurality of stages ranging from a "good" state to a "poor" state of driving performance. As described above, the "driving performance level determination means" includes a discrimination means (classifier) configured to use the electric state values detected in a vehicle running under various electric states and the degree of driving performance obtained in each electric state as learning data according to an arbitrary machine learning algorithm, and to output the degree of driving performance corresponding to the electric state value when the electric state value in the learning data is input, and is configured to output the corresponding "degree of driving performance" when the electric state value detected from time to time is input to the discrimination means while the vehicle is running. For example, linear discrimination, quadratic discrimination, neural network, support vector machine, etc. can be used as the machine learning algorithm constituting the discrimination means. The learning data may be collected, for example, by performing a test run in which various electric states are given to the vehicle as exemplified in Patent Document 1, and the electric state value at that time is measured while measuring or determining the degree of driving performance. Specifically, the degree of high driving performance may be recorded, for example, by a sensory evaluation of the driving performance by the driver during a test drive, or instead (or in addition), it may be judged based on index values representing the vehicle's driving performance, such as the magnitude of fluctuation in the steering angle while driving, the range of fluctuation or delay time of the trajectory until the vehicle starts driving straight after steering, and the speed of response of acceleration / deceleration to depression of the accelerator pedal or brake pedal.
[0008] According to the above-mentioned configuration of the device of the present invention, the level of the vehicle's dynamic performance can be quickly evaluated according to the electric state of the vehicle while the vehicle is running. The evaluation result is useful information for further improving the vehicle's dynamic performance by applying a voltage or current to each part of the vehicle to change the electric state. It is also expected that the evaluation result using the device of the present invention will be useful information for vehicle design.
[0009] In the above-mentioned device of the present invention, at least one portion for measuring the electric state value may be appropriately selected experimentally. More specifically, a portion for which the detected voltage value or current value changes as much as possible when different electric states are given to the vehicle, for example, by applying voltage or current in various ways to each portion of the vehicle, may be selected. This makes it possible to accurately identify the difference in the electric state of the vehicle by the detected electric state value, and therefore it is expected that the degree of high driving performance can be more clearly identified. Specifically, for example, the algorithm of the ReliefF method may be applied to the voltage values or current values measured by sensors provided at multiple portions in various electric states, and the order of the sensors that have the largest change in voltage value or current value in response to the change in the electric state may be determined, and the portion provided with the sensor with the highest order may be selected as the portion for measuring the electric state value.
[0010] The device of the present invention described above may be used together with or as a part of the vehicle aerodynamic characteristic control device described in Patent Document 1. According to the device of the present invention, the degree of the vehicle's driving performance in any electrical state can be determined, and the vehicle's electrical state may be changed in any manner to improve the driving performance by referring to the determination result. Effect of the Invention
[0011] Thus, the above-mentioned device of the present invention can grasp the level of driving performance corresponding to the electric state of the vehicle, so that it is expected that the trial and error process when controlling or adjusting the electric state of the vehicle to improve the aerodynamic performance can be shortened, and the electric state that improves the driving performance can be obtained more quickly. Also, as already mentioned, the results obtained by the device of the present invention provide useful information when adjusting the electric state of the vehicle to improve the driving performance in the design of the vehicle. For example, the electric state of the vehicle changes depending on the position of the earth point of the auxiliary equipment, and the information obtained by the device of the present invention can detect the driving performance according to the arrangement of the earth point of the auxiliary equipment, so that it can be advantageously used when searching for an earth point that gives a higher driving performance.
[0012] Other objects and advantages of the present invention will become apparent from the following description of preferred embodiments of the invention. [Brief description of the drawings]
[0013] [Figure 1] Fig. 1(A) is a schematic plan view of a vehicle equipped with a vehicle dynamics determination device according to the present embodiment, and Fig. 1(B) is a block diagram showing a schematic configuration of the vehicle dynamics determination device according to the present embodiment. [Diagram 2] Fig. 2(A) is a schematic plan view of a vehicle in which sensors for measuring electrical conditions are arranged in various parts of the vehicle body. Fig. 2(B) is a diagram showing importance scores (the higher the sensitivity to changes in the electrical conditions of the vehicle body, the higher the value. See the main text for details) determined by the ReliefF method for the sensors installed in the vehicle shown in Fig. 2(A). [Diagram 3]Fig. 3(A) is a graph showing the distribution of electrical state values (voltage values) measured by two sensors (sensors installed at locations highly sensitive to changes in the electrical state of the vehicle body) when various electrical states are applied to the vehicle body, and a discrimination surface for discriminating the driving performance code (good ⇔ bad) determined based on the distribution. Fig. 3(B) is a graph showing the distribution of electrical state values (voltage values) measured by two sensors when an electrical state different from that given in Fig. 3(A) is applied. [Explanation of symbols]
[0014] 10...vehicle, 12FL,FR,RL,RR...tires, 20...power source, 21f,R...terminals, 30...electrical state control device, 40...determination display unit, 50...determination unit, 70fl to rr...electrical state sensors, s11 to s17...electrical state sensors BEST MODE FOR CARRYING OUT THEINVENTION
[0015] Device configuration The vehicle dynamics determination device according to the present embodiment may be mounted on a vehicle 10 such as a normal automobile as shown in FIG. 1(A). In the vehicle 10, as shown in the figure, sensors 70fl-rr are provided at least one appropriately selected portion of the interior and exterior surfaces of the vehicle body, such as a pillar, the front and rear ends of the vehicle body, for example, to detect a potential value or a current value (electrical state value) at the selected portion, and the detection value of the sensor is transmitted to the determination unit 50. Specifically, the determination unit 50 may be a normal type of computer device that operates according to a program, and is configured to operate as a discriminator that outputs a "dynamics performance code" representing the degree of the vehicle's dynamics as a determination result when an electrical state value from the sensors 70fl-rr is input according to an algorithm of supervised machine learning of any type, as will be described later, and the determination result may be displayed on a display unit 40 of any type. In addition, in the vehicle 10, a means for donating or removing electrons may be mounted on any portion of the vehicle body, as described in Patent Document 1. Specifically, as such a means, as shown in the figure, a power source 20 such as a battery may be mounted on the vehicle, and electrode terminals 21fl-rr of the power source may be arranged at appropriately selected locations, and electrons may be provided or removed at selected locations or regions of the vehicle body. Then, the electric state control device 30 may be configured to control, in an arbitrary manner, to which of the electrode terminals 21fl-rr should a voltage or current be applied from the power source 20 in order to improve driving performance, based on the determination result of the determination unit 50. The electric state control device 30 may be composed of a normal type of computer device that operates according to a program, and a drive circuit or actuator that controls the output of the power source.
[0016] In the determination unit 50, as shown in FIG. 1B, according to the algorithm of the supervised machine learning model, for an electrical state value detected at at least one part of the vehicle, a motion performance code is output, which indicates the degree of high motion performance of the vehicle when the electrical state value is detected. Here, the "motion performance" may be the accuracy or responsiveness of the vehicle's behavior to the driving operation of the vehicle (for example, turning performance, straight running performance, acceleration / deceleration performance), and specifically, may be the driver's feeling when driving, and the higher the degree of high motion performance is, the easier the driver feels to drive, the higher the degree of high motion performance may be determined. The motion performance code may be set to distinguish the degree of high motion performance into a plurality of stages from a high state to a low state, and specifically, may be an index that expresses the motion performance by dividing the state from a "good" state to a "poor" state into at least two stages.
[0017] In the learning of the determination unit 50 (identifier) that outputs a motion performance code from an electrical state value detected in a vehicle, first, the vehicle is test-driven with various electric potential distributions or current distributions (electrical states) applied to the vehicle, while detecting electrical state values (V1, V2, ... I1, ...) in at least one part of the vehicle, and recording the motion performance code (S) at that time, and a pair (V1, V2, ... I1, ...; S) of the detected electrical state value and the recorded motion performance code is collected as learning data. In one embodiment, the motion performance code may be determined by the moment-to-moment sensory evaluation of the driver while driving. In that case, specifically, for example, the motion performance code may be determined by determining which of a plurality of stages (good ⇔ bad) the driver feels the ease of driving at each moment is. In another aspect, in a vehicle, in addition to detecting electrical state values, index values representing the vehicle's behavior, such as the magnitude of fluctuation in steering angle while driving, the range of fluctuation or delay time of the trajectory until straight driving occurs after steering, and the speed of acceleration / deceleration response to depression of the accelerator pedal or brake pedal, may be detected, and a driving performance code may be determined corresponding to the ease of driving based on these detected index values.
[0018] When the learning data is collected as described above, a learning calculation process is executed using the electrical state values as input data and the motion performance code as correct answer data according to an algorithm of a supervised machine learning model of any format, so that a motion performance code is output for a certain electrical state value when that electrical state value is obtained, thereby forming a classifier that outputs a motion performance code estimated to be obtained when an arbitrary electrical state value is input. As the algorithm of the machine learning model, linear discrimination, quadratic discrimination, neural network, support vector machine, etc. may be used.
[0019] Selection of detection location for electrical status values In the driving performance determination device of this embodiment, the more sensitive the part for detecting the electric state value is to the change of the electric state (electric potential distribution, current distribution) in the vehicle, that is, the larger the change of the electric state value of the part, the better it can capture the change of the electric state, and this is advantageous. In fact, the inventors of this embodiment arranged the electric potential sensors s11 to s16 and the current sensor s17 in the vehicle 10 as shown in FIG. 2(A) and applied various changes of the electric state, and found that the manner of change of the detection value differed for each sensor. Then, the algorithm of the ReliefF method, which lowers the rank of the sensor signal whose average value of the sensor is not consistent in the same electric state and raises the rank of the sensor signal whose value is sufficiently different for different electric states, was applied to determine the importance score of each sensor, and the result was as shown in FIG. 2(B). In this figure, it is shown that the higher the importance score, the higher the sensitivity to the change of the electric state. In addition, as will be described later, it was also found that the driving performance code can be suitably determined by using the detection value of a sensor with a high importance score. Thus, in this embodiment, when various electrical state changes are applied to the vehicle, the parts where the detected electrical state values change significantly are searched for, and such parts where the electrical state values change significantly may be selected as the parts for detecting the electrical state values.
[0020] Example of determining the degree of athletic performance based on electrical state values In order to confirm that the degree of driving performance can be determined based on electrical state values detected in various parts of a vehicle, an experiment was conducted as follows.
[0021] In the experiment, first, as shown in Fig. 2(A), potential sensors s11 to s16 and a current sensor s17 were arranged at various locations of the vehicle 10 so that the potential value and current value of each location could be detected. Here, each potential sensor detects the potential based on the negative terminal of the auxiliary battery at each point on the body iron plate, and the current sensor 17 detects the current flowing through the earth wire to the earth point (common with s13) provided on the B pillar so that the current driving the ECU (not shown) mounted on the left B pillar returns to the negative electrode of the battery via the body. Then, the vehicle 10 was driven at a speed of 40 to 70 km / h on a circuit in different electrical states I, II, III, IV, and V (not shown), and the driver's sensory evaluation of the driving performance (good, bad) was recorded for each lap while detecting the electrical state value with each sensor. The data values of the electrical status were calculated as the average value every 10 seconds of the values detected by each sensor every 10 ms while driving at 10 km / h or faster (each successive average period was overlapped by 5 seconds).
[0022] FIG. 3(A) is a distribution diagram in which the voltage values detected by the sensor s13, which has the highest sensitivity among the seven sensors, and the sensor s16, which has the second highest sensitivity, are plotted as values on the horizontal axis and the vertical axis, respectively, under conditions in which each electrical state I, II, III, and IV is given. In addition, in the sensory evaluation by the driver, the driving performance was poor in the electrical states I and IV, and good in the electrical states II and III. Referring to the figure, it is possible to draw three discrimination lines 31 to 33 in the plot space, which divide the area according to the spread of the plots of each electrical state, and further, it is observed that the area of distribution of the plots (□, ○) in the state in which the driving performance was good and the area of distribution of the plots (+, ×) in the state in which the driving performance was poor can be roughly separated with the discrimination line 100 as a boundary. In addition, a classifier was constructed that classifies the driving performance as "good" or "poor" using the secondary discrimination method, using the electrical state values detected by the seven sensors while the vehicle was running under each of the electrical states I, II, III, and IV and the driver's sensory evaluation of the driving performance as learning data. The electrical state values obtained in each of the electrical states I, II, III, and IV were input to the classifier to classify the driving performance. The accuracy rates were as follows: I:96.8%,II: 90.3%,III:91.7%,IV:98.3% A high accuracy rate was obtained in both conditions.
[0023] Furthermore, under the condition of electrical state V, the motor performance was good, and when the voltage values detected by sensors s13 and s16 in that state were plotted in the same manner as in Fig. 3(A), the plot was generally distributed within the "good" region defined by the discrimination line 100 determined using the data of electrical states I, II, III, and IV, as shown in Fig. 3(B). The percentage of correct answers ("good") in the motor performance discrimination results obtained by inputting the electrical state value obtained in electrical state V to a discriminator constructed using the data of electrical states I, II, III, and IV was 83.3%.
[0024] According to the above results, it was shown that there is a correlation between the electrical state values detected at each part of the vehicle and the degree of driving performance, and that the degree of driving performance can be determined with a high accuracy rate based on the electrical state values.
[0025] The above description has been given in relation to the embodiment of the present invention, but it will be apparent to those skilled in the art that many modifications and changes can be easily made thereto, and that the present invention is not limited to the embodiment exemplified above, but can be applied to various devices without departing from the concept of the present invention.
Claims
[Claim 1] A vehicle dynamics performance determination device, an electrical condition detection means for detecting an electrical condition value in at least one portion of the vehicle; a driving performance level determination means for determining a level of driving performance of the vehicle based on the electric state value detected by the electric state detection means; Including, The driving performance degree determination means uses electrical state values detected in a vehicle traveling under various electrical states and the degree of driving performance obtained in each of the electrical states as learning data, and includes an identification means that learns according to a machine learning algorithm to output the degree of driving performance corresponding to an electrical state value when an electrical state value in the learning data is input, and is configured to determine that the degree of driving performance output by the identification means for the electrical state value obtained by the electrical state detection means is the degree of driving performance when that electrical state value was obtained.
Citation Information
Patent Citations
Method of determining earth cable connection points in vehicle
JP2015147487A
Vehicle current sensor
JP2018155625A
Aerodynamic characteristics controller of vehicle through electron donation to vehicle body
JP2022143555A
Device and method for securing an electric line with a controllable switching element
US20120134062A1
