TW201942911A - Data analysis server equipment and optimized combination method - Google Patents

Abstract

A data analysis server device and its optimal combination analysis method. This optimal combination analysis method first establishes several gene sequences during the initialization phase, and then uses these gene sequences to perform calculations such as mating and mutation, thereby generating the optimal gene sequence. And this best combination analysis method can be used for energy consumption estimation, traffic information estimation, and physiological information estimation.

Description

Data analysis server equipment and its best combination analysis method

The invention relates to a data analysis technology related to driving behavior, and in particular to a data analysis server device and an optimal combination analysis method thereof.

According to the research and analysis report of the Taiwan Economic Research Institute, in the cost structure ratio of the automobile freight industry and the automobile passenger industry, the cost of fuel oil accounts for 24% to 29% in each year. It can be seen that fuel cost is one of the main cost factors of vehicles, especially reflected in the automobile freight industry, whose fuel cost is higher than the cost of salaries and welfare benefits, and ranks first in the cost structure. In view of this, if a system and method capable of monitoring fuel consumption can be developed, this problem can be effectively dealt with.

In the prior art, although there are technologies that use historical data of vehicle types, fuel gauge voltage, and driving speed to obtain and correct fuel value values, there are also technologies that detect battery voltage and use them to calculate vehicle fuel consumption, or diagnose fuel tanks Technology for returning fuel quantity data, etc. However, these previous technologies each lack effective feedback methods, or cannot comprehensively estimate the fuel cost required by the freight industry through factors such as road traffic conditions, driver differences, etc., which have their own shortcomings and need to be addressed. Improvement.

The invention provides a data analysis server device and an optimal combination analysis method thereof, which comprehensively considers traffic information or physiological information of driving behavior, thereby providing the best evaluation information.

The optimal combination analysis method of the present invention is suitable for analyzing information of driving behavior response. This optimal combination analysis method includes the following steps. Gene sequences are generated, and each gene sequence contains several chromosomes, and these chromosomes are statistical quantities related to the evaluation information caused by driving behavior at different time points, and each statistical quantity is the number of evaluation information at different time points that meets the numerical interval. The traffic information or physiological information reflected by the driving behavior is input into the adaptation function to calculate the score of these gene sequences, and these gene sequences are used as the weight value of the adaptation function. These gene sequences are subjected to selection procedures, mating procedures, and mutation procedures, and the optimal gene sequence is generated when the scores of these gene sequences converge, and this optimal gene sequence is a collection of evaluation information of driving behavior.

In another aspect, the data analysis server device of the present invention includes a communication module, a memory, and a processor. The communication module receives traffic information or physiological information reflected in driving behavior. The memory records traffic information or physiological information, and several modules. The processor is coupled to the communication module and the memory, and accesses and executes those modules stored in the memory. And those modules include the best combination analysis module. The best combination analysis module performs the following steps. Gene sequences are generated, and each gene sequence contains several chromosomes, and these chromosomes are statistical quantities related to the evaluation information caused by driving behavior at different time points, and each statistical quantity is the number of evaluation information at different time points that meets the numerical interval. The traffic information or physiological information reflected by the driving behavior is input into the adaptation function to calculate the score of these gene sequences, and these gene sequences are used as the weight value of the adaptation function. These gene sequences are subjected to selection procedures, mating procedures, and mutation procedures, and the optimal gene sequence is generated when the scores of these gene sequences converge, and this optimal gene sequence is a collection of evaluation information of driving behavior.

Based on the above, the embodiment of the present invention can improve the genetic algorithm. First, a plurality of excellent gene sequences are established at the initialization stage, and then the plurality of gene sequences are used to calculate the optimal gene sequence by performing mating and mutation calculations. And this improved genetic algorithm can be combined with the adaptation function of neural network. The embodiment of the invention can be applied to the estimation of traffic information and physiological information, and the gene sequence is used as the weight value of the adaptation function to obtain the optimal gene sequence.

In order to make the above features and advantages of the present invention more comprehensible, embodiments are hereinafter described in detail with reference to the accompanying drawings.

FIG. 1 is a system architecture diagram according to an embodiment of the present invention. Referring to FIG. 1, this system includes at least a plurality of vehicle devices 1, a plurality of user devices 2, a data analysis server device 3, and a database device 4.

The vehicle device 1 may be a car, a locomotive, a bus, or a train. In this embodiment, the vehicle device 1 includes at least a positioning module 14, an intermediary software module 12, and a communication module 10.

The positioning module 14 may support a global positioning system (for example, GPS, Big Dipper, Galileo positioning system, etc.) or a wireless network signal positioning (for example, base station positioning, Wi-Fi positioning, etc.) method, and obtain position information and vehicle speed information. In this embodiment, the positioning module 14 can support a global positioning system, and can obtain the latitude and longitude coordinates and vehicle speed information of the vehicle equipment through a satellite signal.

The communication module 10 can support wireless network transmission, and can establish communication between the user equipment 2 and the data analysis server equipment 3. In this embodiment, the communication module 10 can support 4G (Long Term Evolution, LTE) communication to connect to the 4G network and establish a communication connection with the data analysis server device 3.

The intermediary software module 12 is stored in a memory (for example, a hard disk, a memory, a temporary memory, etc.) of the vehicle device 1 and is loaded and executed by a processor (for example, a CPU, a chip, a microprocessor, etc.), and Supports transmission protocols such as HyperText Transfer Protocol (HTTP), Message Queuing Telemetry Transport (MQTT), or restricted application protocols. The vehicle device 1 may be connected to the data analysis server device 3 via the communication module 10 through the intermediary software module 12 to transmit vehicle device information, traffic information, and / or physiological information to the data analysis server device 3. Vehicle equipment information may include vehicle number, vehicle model, driver number, time information, location information, speed information, and so on. Traffic information can be travel time, traffic volume, speed, etc. The physiological information may be a heart rate value, a heart rate change value, and the like. In this embodiment, the intermediary software module 12 can support Hypertext Transfer Protocol and Representational State Transfer (REST), and the intermediary software module 12 can call the application program interface of the data analysis server device 3 (Application Program Interfaces, APIs), and transmits vehicle equipment information, traffic information, and / or physiological information to the data analysis server device 3 in a periodic or aperiodic manner through the communication module 10.

In this embodiment, the vehicle device 1 includes a vehicle number, a vehicle model, and a driver number. Assume that there are _CN vehicle equipment, _TN vehicle models, and D _N drivers in the system. Vehicle equipment 1 can send vehicle equipment information, traffic information, and / or physiological information to the data analysis server equipment 3 every 30 seconds. Moreover, each vehicle device 1 may include an identification device, and the driver of each vehicle device 1 may insert his identification document into the identification device to obtain driver identity information.

As shown in Table 1. For example: driver 1 drives vehicle equipment 1 with vehicle number 1 on 2015/01/01, the vehicle model of vehicle equipment 1 is vehicle model 1, and vehicle equipment 1 can obtain the vehicle equipment at 06:00:00 via the positioning module 14. 1 location information (that is, longitude 102.5423383 degrees and latitude 24.09490167 degrees) and vehicle speed information (real-time speed 44 km / h), and can call the REST APIs of the data analysis server 3 through the intermediary software module 12, and the vehicle equipment information Send to data analysis server 3. Table (1), vehicle equipment information

The user equipment 2 may be a smart phone, a tablet computer, a computer host, a notebook computer, or other devices. The user equipment 2 includes at least a user interface 24, an intermediary software module 22, and a communication module 20.

The user interface 24 can be presented through a display (for example, LCD, LED, OLED display, etc.) to provide the user with the operation of the user equipment 2 and obtain the vehicle number, time information, energy information and other evaluation information input by the user. The energy information can be oil quantity information or electricity information, and the analysis result can be queried from the data analysis server device 3, and the analysis result can be displayed on the user interface 24.

The communication module 20 can support any type of wireless network transmission or wired network transmission, and can establish communication between the user equipment 2 and the data analysis server equipment 3. In this embodiment, the communication module can support 4G communication, and the user equipment 2 can connect to the 4G network through the communication module 20 and establish a communication connection with the data analysis server device 3.

The intermediary software module 22 is stored in a memory (for example, a hard disk, a memory, a register, etc.) and is executed after being loaded by a processor (for example, a CPU, a chip, a microprocessor, etc.), and supports hypertext transmission. Protocols, or telemetry transmission of message sequences, or restricted application protocols, and can be connected to the data analysis server device 3 via the communication module 20 to transmit vehicle number, time information, and energy information to the data analysis server device 3. The energy information may include fuel quantity information or power information. The intermediary software module 22 may receive the analysis result of the data analysis server device 3. In this embodiment, the intermediary software module 22 can support hypertext transfer protocol and representational status transmission. The user device 2 can call the REST APIs of the data analysis server device 3 through the intermediary software module 22, and use the user in using The vehicle number, time information, and energy information input by the user interface 24 are transmitted to the data analysis server device 3 through the communication module 20. It can be known from this that the user equipment 2 can obtain the energy information and other evaluation information input by the user and send it to the data analysis server equipment 3.

In this embodiment, the user equipment 2 can be operated by the user aperiodically, and the vehicle number, time information, and fuel quantity information input by the user are obtained through the user interface 24, and the vehicle number is transmitted through the communication module 20 , Time information, and energy information (ie oil quantity information) to the data analysis server device 3 and its energy consumption estimation module, as shown in Table (2): Table (2), Fuel consumption information

In this embodiment, the vehicle equipment 1 of the vehicle number 1 is refueled at 43.04 liters at 2015/01/05 18:51:00. The user can operate the user equipment 2 according to the refueling invoice information, and enter the vehicle number via the user interface 24 ( Ie vehicle number 1), time information (ie 2015/01/05 18:51:00), and fuel quantity information (ie 43.04 liters), and the intermediary software module 22 can call the REST APIs of the data analysis server device 3, The entered vehicle number, time information, and energy information (ie, fuel quantity information) are transmitted to the data analysis server device 3.

In this embodiment, the user equipment 2 can be used by the user for non-periodic operation. The user vehicle 24 receives the input vehicle number, time information, and power information through the user interface 24, and transmits 22 the vehicle number and time information through the intermediary software module. , And energy information (ie, power information) to the data analysis server device 3, as shown in table (3): Table (3), power consumption information

In this embodiment, the vehicle equipment 1 of vehicle number 2 is charged at 17.22 degrees (kWh (1kWh)) at 2015/01/05 12:50:00, and the user operates the user equipment 2 according to the charging information, via the user interface 24 Enter the vehicle number (that is, vehicle number 2), time information (that is, 2015/01/05 12:50:00), and power information (that is, 17.22 degrees), and the intermediary software module 22 can call the data analysis server device 3 REST APIs to transmit the entered vehicle number, time information, and energy information (that is, power information) to the data analysis server device 3.

The data analysis server device 3 includes at least an intermediary software module 32, a communication module 30, and an optimal combination analysis module 34. In this embodiment, the data analysis server device 3 may support a Linux operating system, a Microsoft Windows operating system, and the like, and various types of servers such as a network service server may be built on the operating system.

The intermediary software module 32 is stored in a memory (for example, a hard disk, a memory, a temporary register, etc.) and executed by a processor (for example, a CPU, a chip, a microprocessor, etc.), and supports hypertext transmission. Transmission protocols such as protocols, message sequence telemetry transmissions, or restricted application protocols. The data analysis server device 3 may be connected to the vehicle device 1 and the user device 2 through the communication module 30 through the intermediary software module 32 to receive vehicle device information, traffic information, and / or physiological information transmitted by the vehicle device 1 and The vehicle number, time information, and energy information transmitted by the user equipment 2 are received, and a message may be transmitted to the vehicle equipment 1 or the user equipment 2, and the received vehicle equipment information and energy information may be stored in the database equipment 4. In this embodiment, the intermediary software module 32 may be implemented using a Tomcat network service server, and several REST APIs may be built for the vehicle device 1 and the user device 2 to connect, and the vehicle may be received via a hypertext transfer protocol The vehicle equipment information, traffic information, and / or physiological information transmitted by the device 1 and the vehicle number, time information, and energy information transmitted by the user device 2 can be received, and the message can be transmitted to the vehicle device 1 or the user device 2 to better receive the information. The vehicle equipment information and energy information are stored in the database equipment 4.

The communication module 30 can support any type of wired network transmission (for example, Ethernet, fiber optic network, etc.), establish vehicle equipment 1 and data analysis server equipment 3, user equipment 2 and data analysis server equipment 3 And communication transmission between the database device 4 and the data analysis server device 3.

The optimal combination analysis module 34 is stored in a memory (for example, a hard disk, a memory, a register, etc.) and is executed after being loaded by a processor (for example, a CPU, a chip, a microprocessor, etc.), and is executable. The optimal combination analysis method of the embodiment of the present invention (to be detailed in the subsequent embodiments) collects vehicle equipment information, traffic information and / or physiological information transmitted by vehicle equipment 1, vehicle number, time information transmitted by user equipment 2, And energy information, and analyze the amount of energy consumed by each driving behavior and other assessment information. The driving behavior can be vehicle speed information. The optimal combination analysis module 34 can store the amount of energy consumed by each driving behavior to the database device 4 and generate a set of estimated driving behavior energy consumption information. Alternatively, the optimal combination analysis module 34 may estimate the corresponding evaluation information (for example, the degree of traffic jam, driver fatigue, etc.) based on the traffic information or physiological information reflected by the driving behavior, which will be described in detail in the subsequent embodiments.

The data analysis server device 3 may be connected to the external geographic information server through the communication module 30 through the intermediary software module 32, and query the external geographic information server with REST APIs to obtain the road type corresponding to the vehicle equipment information and its location information. The road type and vehicle equipment information are merged into the modified vehicle equipment information, and the modified vehicle equipment information is stored in the database equipment 4. The external geographic information server can be Google Map server or Chunghwa Telecom GeoWeb map server, as shown in Table (4): Table (4), modified vehicle equipment information stored in database device 4

The database device 4 includes at least a storage module 44, a computing module 42, and a communication module 40. In this embodiment, the database device 4 may be implemented using Microsoft Structured Query Language (SQL) server, MySQL, PostgreSQL, Oracle database server, MongoDB server, HBase server, etc., and The data of the data analysis server device 3 is received and stored through the communication module 40.

The communication module 40 can support any type of wired network transmission, and establish a communication connection between the database device 4 and the data analysis server device 3.

The computing module 42 (for example, various types of processors or chips) can receive the request transmitted by the data analysis server device 3 through the communication module 40 and access the storage module 44 according to the obtained request.

The storage module 44 (for example, a hard disk, a memory, or a memory card) may be coupled to the computing module 42 and provide external operations such as adding, modifying, deleting, and querying stored data. In this embodiment, the storage module 44 stores a vehicle number and vehicle model comparison table (as shown in Table (5)), modified vehicle equipment information (as shown in Table (4)), and fuel consumption information (such as Table (2)), power consumption information, etc. (as shown in Table (3)), traffic information, and physiological information.

When a new vehicle device 1 is to be added, the system manager can log in the corresponding vehicle number and vehicle model of the new vehicle device 1 to the vehicle number and vehicle model comparison table, and the vehicle number and vehicle model comparison table can provide a data analysis server. Device 3 creates the modified vehicle equipment information after query. Table (5), vehicle number and vehicle model comparison table

In an embodiment, the vehicle equipment 1 may further include an energy detection device (not shown). The energy detection device can detect the energy information of the vehicle equipment 1. The energy information may be fuel quantity information or electricity information. Energy information It may be included in the vehicle equipment information, and the vehicle equipment information may be transmitted to the data analysis server equipment 3 through the intermediary software module and the communication module.

In addition, the energy detection device may periodically or aperiodically detect the fuel quantity information of the vehicle equipment 1, and record the vehicle number, time information, and fuel quantity information of the vehicle equipment 1, and then transmit the vehicle number through the intermediary software module 12. , Time information, and energy information (that is, oil quantity information) to the data analysis server device 3, as shown in Table (6): Table (6), Fuel consumption information

In this embodiment, the vehicle equipment 1 of vehicle number 3 is before 2015/01/01 06:00:00, and its energy detection device detects that the remaining fuel quantity of the fuel tank of the vehicle equipment 1 is 4 liters. At 01/01 06:00:00, it was detected that the remaining fuel in the fuel tank was 3.973 liters. The fuel consumption information calculated by the energy detection device is 0.027 liters, as shown in table (6), and the vehicle number can be assigned. , Time information, and fuel quantity information are transmitted to the data analysis server device 3 through the intermediary software module 12.

As shown in Table (7), when vehicle equipment 1 of vehicle number 4 is at 2015/01/01 06:00:00, its energy detection device detects vehicle equipment 1 at 2015/01/01 05:59. A total of 0.013 kWh (kWh (1kWh)) was consumed from 30 to 2015/01/01 06:00:00. The recordable power consumption information is 0.013 kWh, as shown in Table (7). The number, time information, and power information are transmitted to the data analysis server device 3 through the intermediary software module 12. Table (7), power consumption information

In this embodiment, the user does not need to input the vehicle number, time information, and fuel quantity information through the user interface 24, and then transmits the vehicle number, time information, and energy information to the data analysis server device through the intermediary software module 12. 3.

In addition, the data analysis server device 3 intermediary software module 32 may receive the vehicle device information, traffic information and / or physiological information of the vehicle device 1 and the vehicle number, time information, and energy information of the vehicle device 1 through the communication module 30. And save those received information to database device 4. The best combination analysis module 34 of the data analysis server device 3 can execute the best combination analysis method, collect vehicle equipment information, traffic information, physiological information, vehicle number, time information, and energy information transmitted by the vehicle equipment 1, and Analyze the amount of energy consumed by each driving behavior. This driving behavior may be vehicle speed information, the data analysis server device 3 may store the amount of energy consumed by each driving behavior to the database device 4. The data analysis server device 3 can store data to the database device 4, and then the data analysis server device 3 executes an energy consumption estimation method to calculate the amount of energy consumed by each driving behavior.

Based on the foregoing system architecture, several embodiments of the present invention applying the best combination analysis method will be further described below to help the reader understand the spirit of the present invention.

An embodiment of the present invention is applied to an energy consumption estimation method. The steps performed by the optimal combination analysis module 34 of the data analysis server device 3 include at least a driving behavior collection procedure, an energy information collection procedure, and an optimal combination analysis procedure.

In the process of collecting driving behavior, the vehicle equipment 1 reports vehicle equipment information, traffic information, and / or physiological information to the data analysis server equipment 3, and the data analysis server equipment 3 analyzes the vehicle equipment information, traffic information, and / or physiology Information, and store vehicle equipment information, traffic information and / or physiological information to the database device 4. The data analysis server device 3 can calculate the number of various driving behaviors of each vehicle device, each vehicle model, each road type, and each driver in a period of time.

In the energy information collection process, the user equipment 2 reports energy information or other evaluation information to the data analysis server device 3 or the vehicle equipment energy detection device detects the energy information and returns the energy information to the data analysis server device 3, and then The data analysis server device 3 analyzes the energy information, and stores the energy information and / or evaluation information in the database device 4, and the data analysis server device 3 can calculate each vehicle device 1 and each vehicle model in a period of time. , Each road type, the amount of energy consumed by each driver, and / or the statistical amount of evaluation information (for example, the number of vehicles on a certain road section, the cumulative number of sober states, etc.).

In the optimal combination analysis program, the optimal combination analysis module 34 obtains the number of driving behaviors and the energy consumption amount, and the optimal combination analysis module 34 executes a genetic algorithm to analyze the energy consumption amount of each driving behavior and outputs A collection of driving behavior energy consumption estimation information or other driving behavior assessment information collection.

It is worth noting that the vehicle equipment 1 can also execute the collection of driving behavior procedures to obtain the modified vehicle equipment information (as shown in Table (4)), and can be based on the vehicle equipment, vehicle model, driver, road type and other information. Count the number of driving behaviors. This driving behavior can be vehicle speed information, which is defined as v. In this embodiment, the driving behavior is based on a 10 km segment per hour, but not limited to 10 km.

The revised vehicle equipment information statistics show the number of driving behaviors of each vehicle equipment in 2015, as shown in Table (8). The vehicle speed information reported by vehicle number 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the vehicle speed information reported by vehicle equipment _CN in 2015 is 0 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (8), statistics on the number of driving behaviors according to vehicle number 1

The revised vehicle equipment information counts the number of driving behaviors of various vehicle models in 2015, as shown in Table (9). The vehicle speed information reported by vehicle model 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the vehicle speed information reported by the vehicle model T _N in 2015 is 0 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (9), statistics on the number of driving behaviors by vehicle model

The revised vehicle equipment information statistics show the number of driving behaviors of each driver in 2015, as shown in Table (10). The speed information reported by driver 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the driving speed information reported by driver D _N in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (10), statistics on the number of driving behaviors by driver

The revised vehicle equipment information statistics show the number of driving behaviors of various road types in 2015, as shown in Table (11). The speed information reported by road type 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the road speed R _N reported the speed information of 0 km / h in 2015. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (11), statistics on the number of driving behaviors by road type

The revised vehicle equipment information statistics show the driving behaviors of each vehicle equipment 1 and each driver in 2015, as shown in Table (12). The speed information reported by driver 1 driving vehicle equipment 1 in 2015 was 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Driver 2 reported vehicle speed information of vehicle number 1 in 2015 as 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. The speed information reported by driver 1 driving vehicle number 2 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the driver ’s D _{N ’s} vehicle number C _{N ’s} reported speed information in 2015 is 0 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (12), Statistics of driving behaviors by vehicle equipment and driver

The revised vehicle equipment information statistics of its various vehicle models and drivers' driving behaviors in 2015 are shown in Table (13). The speed information reported by driver 1 driving vehicle model 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. The speed information reported by driver 2 driving vehicle model 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Driver 1 driving vehicle model 2 reported speed information of 0 km / h in 2015. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the speed information reported by the driver D _N driving the vehicle model T _N in 2015 is 0 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (13), statistics of driving behaviors by vehicle type and driver

The revised vehicle equipment information statistics show the number of driving behaviors of each vehicle equipment 1 and various road types in 2015, as shown in Table (14). The vehicle speed information reported by vehicle equipment 1 on road type 1 in 2015 was 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. The vehicle speed information reported by vehicle number 1 on road type 2 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. The vehicle speed information reported by road number 1 for road type 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the vehicle speed information reported by vehicle equipment C _N on road type R _N in 2015 is 0 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (14), statistics of driving behaviors by vehicle equipment and road type

The revised vehicle equipment information statistics show the number of driving behaviors of each driver and road type in 2015, as shown in Table (15). Driver 1 drives those vehicles and equipment 1 The road speed 1 reported the speed information of 0 km / h in 2015. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Driver 1 drives those vehicle equipment 1 on road type 2 and the speed information reported in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Driver 2 drives those vehicles and equipment 1 The road speed information reported by road type 1 in 2015 is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. By analogy, the driver D _N drives those vehicle equipment 1 on the road type R _N. The speed information reported by the vehicle in 2015 is 0 km / h. The total number of records, such as: ..., speed information greater than 120 km / h pen. Table (15), statistics on the number of driving behaviors by driver and road type

It should be noted that the number of driving behaviors generated by the statistically modified vehicle equipment information is not limited to the year. It can be calculated using a time interval. The time interval must include year, quarter, month, week, day, hour, minute, second, etc. .

Take the month as an example: The vehicle speed information reported by vehicle number 1 in January 2015 is 0 km / h. Pen, vehicle number 1The data of vehicle speed information reported in March 2015 is 0 km / h. And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the vehicle speed information reported by vehicle model 1 in January 2015 is 0 km / h. Pen, vehicle model 1The data of the speed information reported in March 2015 is 0 km / h. And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the speed information reported by driver 1 in January 2015 was 0 km / h. Pen, driver 1 reported the speed information of 0 km / h in March 2015. And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the speed information reported by road type 1 in January 2015 is 0 km / h. The number of data for the pen and road type 1 reported in March 2015 is 0 km / h. And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the number of data reported by driver 1 in driving vehicle number 1 in January 2015 was 0 km / h. Pen, driver 1 driving vehicle number 1 in 2015 M. The speed information reported by the vehicle is 0 km / h. And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the speed information reported by driver 1 driving vehicle model 1 in January 2015 is 0 km / h. Pen, driver 1, driving vehicle model 1, the vehicle speed information reported in March 2015 is 0 km / h, the total number of data is And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the number of data for vehicle number 1 on road type 1 in January 2015 was 0 km / h. The number of pens and vehicle numbers 1 on road type 1 in 2015 M is 0 km / h. The total number of data is And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; driver 1 drives those vehicle equipment 1 on road type 1 and the speed information reported in January 2015 is 0 km / h. Pen, driver 1 drives those vehicle equipment 1 on road type 1 and the speed information reported in March 2015 is 0 km / h. The total number of records is And the sum of the months of the year is equal to the sum of the whole year (i.e. ),So on and so forth.

The vehicle equipment 1 executes the energy information collection program to obtain the fuel consumption information (as shown in Table (2)) or the power consumption information (as shown in Table (3)) from the user equipment 2 and combine the vehicle number and Vehicle model comparison table (as shown in Table (5)) and modified vehicle equipment information (as shown in Table (4)). According to vehicle number, vehicle model, driver, etc., a period of time can be used to count the energy consumption. The time interval must include year, quarter, month, week, day, hour, minute, second, and so on.

Or, when the data analysis server device 3 executes the energy information collection program, it may obtain fuel consumption information (as shown in Table (6)) or power consumption information (as shown in Table (7) from the energy detection device of the vehicle equipment 1). )), And must combine the vehicle number and vehicle model comparison table (as shown in table (5)) and the modified vehicle equipment information (as shown in table (4)). The data analysis server device 3 can be based on the vehicle number, vehicle model, driver and other information, and can use a period of time to count the amount of energy consumption, and this time interval can include year, quarter, month, week, day, hour, minute , Seconds, etc.

This available year is a time interval. Energy consumption information can be fuel consumption information or power consumption information. The data analysis server equipment 3 can be calculated according to the vehicle number, vehicle model, and driver respectively. The total energy consumption of vehicle equipment 1 in 2015 is The total energy consumption of vehicle equipment N in 2015 is , And so on; The total energy consumption of vehicle model 1 in 2015 is The total energy consumption of vehicle model N in 2015 is And so on; in calculating the energy consumption of the vehicle model, the data analysis server device 3 can take out the vehicle number (ie vehicle equipment) of the same vehicle model according to the vehicle number and vehicle model comparison table, and those vehicle equipment The energy consumption corresponding to the time interval is added up to the energy consumption of the vehicle model.

Driver 1's total energy consumption in 2015 is The total energy consumption of driver N in 2015 is , And so on; The total energy consumption of driver 1 driving vehicle equipment 1 in 2015 is The total energy consumption of driver N ₁ driving vehicle equipment N ₂ in 2015 is , And so on; The total energy consumption of driver 1 driving vehicle model 1 in 2015 is The total energy consumption of driver N ₁ driving vehicle model N ₂ in 2015 is ,So on and so forth.

The data analysis server equipment 3 available month is the time interval, statistics of energy consumption information, energy consumption information can be fuel consumption information or power consumption information, and can be calculated according to the vehicle number, vehicle model, and driver respectively: Vehicle equipment 1Total energy consumption in January 2015 is The total energy consumption of vehicle equipment N in March 2015 is And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the total energy consumption of vehicle model 1 in January 2015 is The total energy consumption of vehicle model N in M in 2015 is And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the total energy consumption of driver 1 in January 2015 is The total energy consumption of Driver N in M in 2015 is And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the total energy consumption of driver 1 driving vehicle equipment 1 in January 2015 is The total energy consumption of driver N ₁ driving vehicle equipment N ₂ in 2015 M is And the sum of the months of the year is equal to the sum of the whole year (i.e. ), And so on; the total energy consumption of driver 1 driving vehicle model 1 in January 2015 is The total energy consumption of driver N ₁ driving vehicle model N ₂ in 2015 M is And the sum of the months of the year is equal to the sum of the whole year (i.e. ),So on and so forth.

After obtaining the number of driving behaviors and the amount of energy consumption, the optimal combination analysis module 34 can execute the optimal combination analysis method. Please refer to FIG. 2 for a flowchart of the optimal combination analysis method.

The best combination analysis module 34 establishes initial data (step S210). This initial data includes the number of driving behaviors, the amount of energy consumption, and the number of maternal gene sequences. Evolution times Overlap times Mating rate Mutation rate . Number of evolutions The initial value is 0, and each time a genetic algorithm (including selection, mating, and mutation procedures) is executed, the number of evolutions is increased by one until the number of evolutions Equal to the number of overlapping .

The optimal combination analysis module 34 then executes an adaptation function generation algorithm (step S211) to generate an adaptation function, which is used to calculate a score of a gene sequence, and each gene sequence includes several chromosomes. These coloring systems are related to the statistical quantities of evaluation information (for example, energy consumption, number of vehicles, exhaustion, etc.) caused by driving behavior at different time points, and each statistical quantity is the number of evaluation information that meets the value interval at different time points.

The optimal combination analysis module 34 then executes a gene sequence generation algorithm (step S212) to generate a gene sequence according to the number of chromosomes required for the adaptation function, and can also depend on the number of gene sequences of the mother group. Gene sequences are generated for the mother population.

The optimal combination analysis module 34 then executes a genetic sequence score calculation algorithm (step S213) to input the genetic sequence and driving speed information, traffic information, or physiological information reflected in the driving behavior into the adaptation function, and calculates the genetic sequence score , And those gene sequences are used as weight values for the fitness function.

The optimal combination analysis module 34 then determines whether the function converges (step S214). Equal to the number of overlapping At this time, the best combination analysis module 34 outputs the best gene sequence, and this best gene sequence is a set of estimated driving energy consumption information or a set of evaluation information of other driving behaviors. On the other hand, when the number of evolutions Less than the number of lapping , The optimal combination analysis module 34 will evolve Plus one.

The optimal combination analysis module 34 then uses a gene sequence selection algorithm (ie, a selection program) to copy two sets of gene sequences in those gene sequences (step S216), thereby forming two parent gene sequences.

The optimal combination analysis module 34 then executes a genetic sequence mating algorithm (step S217, that is, a mating program), and the mating rate is determined based on the mating rate. , Mating the two parent gene sequences to generate two first-generation daughter gene sequences.

The optimal combination analysis module 34 then executes a genetic sequence mutation algorithm (step S218, that is, a mutation program) to determine the mutation rate based on the mutation rate. The two first-generation daughter gene sequences are mutated to form two second-generation daughter gene sequences.

The best combination analysis module 34 replaces the newly generated two second-generation daughter gene sequences with the two gene sequences in those gene sequences in the mother group, and the substituted gene sequences correspond to the two worst scores ( Step S219).

The best combination analysis module 34 can further obtain two sets of new gene sequences (step S220), and use the gene sequence score calculation algorithm to calculate the scores of those new gene sequences, and execute the gene algorithm again.

For example, suppose the number of maternal gene sequences in the initial data Set to 14, its evolution times The initial value is 0, and the number of times it is overlapped 1000, and its mating rate 100%, its mutation rate Is 7%.

The number of driving behaviors can be calculated by vehicle equipment (as shown in Table (8)), the number of driving behaviors by vehicle model (as shown in Table (9)), and the number of driving behaviors by driver ( (As shown in Table (10)), counting the number of driving behaviors by road type (as shown in Table (11)), counting the number of driving behaviors by vehicle equipment and drivers (as shown in Table (12)), by vehicle Model and driver count the number of driving behaviors (as shown in Table (13)), count the number of driving behaviors according to vehicle equipment and road types (as shown in Table (14)), or count each driving behavior by driver and road type Number of actions (as shown in Table (15)).

The energy consumption amount may be the vehicle equipment energy consumption amount, the vehicle model energy consumption amount, the vehicle model energy consumption amount, the driver energy consumption amount, the vehicle equipment and driver energy consumption amount, or the vehicle model and driver energy consumption amount.

In this example, the number of driving behaviors is calculated based on the vehicle equipment and the driver (as shown in Table (12)), and the amount of energy consumption is the amount of vehicle equipment and driver energy consumption.

Take driver i driving vehicle equipment j as an example: driver i driving vehicle equipment j reported in 2015 that the speed information is 0 km / h. There are a total of data of pen and vehicle speed information between 0 and 10 km / h. The total number of records, such as: ..., speed information greater than 120 km / h The number of data reported by driver i driving vehicle equipment in January 2015 was 0 km / h. The total number of records of the vehicle speed information reported in February 2015 is 0 km / h. The total number of records for which the speed information reported in December 2015 is 0 km / h is: And the sum of the months of the year is equal to the sum of the whole year (i.e. ); The total energy consumption of driver i driving vehicle equipment j in 2015 is The total energy consumption of driver i driving vehicle equipment j in January 2015 is The total energy consumption of driver i driving vehicle equipment j in 2015 M is And the sum of the months of the year is equal to the sum of the whole year (i.e. ).

In this embodiment, the number of driving behaviors of driver 1 driving vehicle equipment 1 throughout the year is a set , And driver 1 drives vehicle equipment 1 and its annual gasoline consumption is 10921.364 liters.

In the adaptation function generation algorithm, the optimal combination analysis module 34 can generate a multivariate linear function as the adaptation function, and the adaptation function can be used to calculate the gene sequence score s. In this embodiment, taking driver i to drive vehicle equipment j as an example, referring to FIG. 3, the adaptation function is: Where the gene sequence score s is lower in this embodiment, the better, that is, the best solution is .

A set of gene sequences The gene sequence contains 14 chromosomes (the cardinality of the set ), Of which the first dyeing system The chromosome can be a floating-point number encoding and can be regarded as the energy consumption amount corresponding to the idling speed (vehicle speed information is 0 km / h) of driver i driving vehicle equipment j.

In addition, in the gene sequence generation algorithm, the optimal combination analysis module 34 can generate a gene sequence according to the number of chromosomes required for the adaptation function, and can also depend on the number of maternal group gene sequences. A plurality of gene sequences are generated from the mother population.

Take driver i driving vehicle equipment j as an example, the number of maternal gene sequences in this example Line 14 and the number of chromosomes are 14, the gene sequence generation algorithm will randomly generate 14 gene sequences. These gene sequences each contain 14 chromosomes, and the optimal combination analysis module 34 uses those gene sequences as maternal gene sequences.

In addition, taking driver i driving vehicle equipment j as an example, those gene sequences are expressed in this embodiment as: gene sequence 1 line Gene sequence 2 , And so on; gene sequence 14 . Table (16), Maternal gene sequence

In this embodiment, the maternal gene sequences are randomly generated, and those of the staining system are encoded by floating point numbers. Taking driver 1 driving vehicle equipment 1 as an example, as shown in Table (17), those gene sequences are expressed in this embodiment as follows: Gene sequence 1 line Gene sequence 2 , And so on; gene sequence 14 . Table 17.Genetic gene sequence of driver 1 driving vehicle equipment 1

In the gene sequence score calculation algorithm, the optimal combination analysis module 34 can input each gene sequence in the maternal gene sequence into an adaptation function and calculate the gene sequence score. . Taking driver i driving vehicle equipment j as an example, the gene sequence score corresponding to gene sequence 1 is The gene sequence score corresponding to the gene sequence h .

Take driver 1 driving vehicle equipment 1 as an example, the gene sequence score corresponding to each gene sequence of the maternal gene sequence: ; ; So on and so forth, .

After the optimal combination analysis module 34 executes the gene sequence score calculation algorithm, it will judge the number of evolutions. Is it equal to the number of overlaps . If the number of evolutions Equal to the number of overlapping Then, the best combination analysis module 34 outputs the best gene sequence, which is a group of gene sequences of the maternal gene sequence, and this gene sequence corresponds to (has) the best gene sequence score, and the best gene A sequence is a collection of information on driving energy consumption estimates. On the other hand, when the number of evolutions Less than the number of lapping , The optimal combination analysis module 34 will evolve Plus one.

The process of the genetic algorithm includes a gene sequence selection algorithm (or selection program), a gene sequence mating algorithm (or mating program), and a gene sequence mutation algorithm (or mutation program).

In this embodiment, the genetic sequence selection algorithm is a roulette wheel selection. The optimal combination analysis module 34 can use the roulette method to copy two sets of the maternal gene sequences to form two maternal gene sequences. . Taking driver 1 driving vehicle equipment 1 as an example, the optimal combination analysis module 34 selects the gene sequence 1 ( ) And gene sequence 2 ( ), And copy it into the first-generation maternal gene sequence, as shown in Table (18): Table (18), the first-generation maternal gene sequence

In this embodiment, in the process of performing the genetic sequence mating algorithm, the optimal combination analysis module 34 may And perform a 1-point crossover, assuming a crossover point Randomly generated as 2, after the mating process of the two genes, the first-generation mother gene sequence was changed to the first-generation daughter gene sequence (as shown in Table (19)): The first-generation daughter gene sequence 1 First generation daughter gene sequence 2 . Table (19), first generation daughter gene sequence

In this embodiment, the optimal combination analysis module 34 may execute the mutation sequence of the gene sequence according to the mutation rate. And randomly generate a binary vector To perform the two-gene sequence mutation process. Suppose , The value of the nth chromosome in the gene sequence will become another value that is not the original value, and the other value must be randomly generated. For example, the first-generation daughter gene sequence transformed from the first-generation mother gene sequence is used as the second-generation mother gene sequence (as shown in Table (20)), and it is assumed that Then, the second-generation mother gene sequence is transformed into the second-generation daughter gene sequence after the mutation process, as shown in Table (21). Table (20), second generation maternal gene sequence Table (21), second-generation daughter gene sequence

The optimal combination analysis module 34 then replaces the newly generated two second-generation daughter gene sequences with the two gene sequences of those in the mother group, and those gene sequences that correspond to the worst score Both. In this embodiment, take the mother gene sequence of driver 1 driving vehicle device 1 as an example, the gene sequence score corresponding to gene sequence 2 is 1062.54674, the gene sequence score corresponding to gene sequence 14 is 1009.53678, and the two gene sequences are mother The worst-scoring gene sequence in the population. The optimal combination analysis module 34 is about to replace two of those gene sequences in the mother group with two second-generation daughter gene sequences. The results after substitution are shown in Table (22). Table (22). Mother group gene sequence of driver 1 driving vehicle device 1 after one round of evolution

The best combination analysis module 34 will also use the gene sequence score calculation algorithm to calculate the scores of other gene sequences, and judge the number of evolutions. Is it equal to the number of overlaps . If the number of evolutions Equal to the number of overlapping Then, the optimal combination analysis module 34 outputs an optimal gene sequence. If the number of evolutions Less than the number of lapping , The optimal combination analysis module 34 will evolve the number of times Add one and execute the genetic algorithm again. In this embodiment, the optimal combination analysis module 34 will perform a genetic sequence score calculation algorithm to calculate two new gene sequences (i.e., gene sequence 2) of the mother group gene sequence of driver 1 driving vehicle equipment 1 after one round of evolution. And gene sequence 14), the gene sequence scores corresponding to the two new gene sequences can be obtained: ;

When the number of evolutions Equal to the number of overlapping When the optimal combination analysis module 34 outputs the best gene sequence, the best gene sequence is one of the maternal gene sequences and the gene sequence corresponds to the best gene sequence score, and the best gene sequence is the driving behavior energy Consumption estimation information collection or other estimation information of driving behavior.

In this embodiment, taking the driver 1 driving the vehicle device 1 as an example, the gene sequence 14 will be output Optimal gene sequence That is, the amount of gasoline consumed by driver 1 driving vehicle equipment 1 at idle speed (vehicle speed information is 0 km / h) for 30 seconds is 0.012500034 liters, and driver 1 driving vehicle equipment 1 vehicle speed information is 0-10 km / h for 30 seconds Gasoline consumption is 0.018487159 liters, and so on:

In this embodiment, the optimal combination analysis module 34 can perform a dynamic method to correct the chromosome during the execution of the genetic sequence mutation algorithm. This dynamic method refers to the gene sequence to be substituted into the score calculated by the adaptation function for correction. Taking driver 1 driving vehicle equipment 1 as an example, the optimal combination analysis module 34 can perform And randomly generate a binary vector To perform the two-gene sequence mutation process. Suppose , The value of the nth chromosome in the gene sequence is corrected by substituting the reference gene sequence into the score calculated by the adaptation function.

For example, its second-generation maternal gene sequence (as shown in Table 19) can be mutated by the gene sequence mutation method, and it is assumed that , Its second-generation maternal gene sequence 1 Chromosome 1 2nd generation maternal gene sequence 2 Chromosome 1 The following calculations can be used to mutate, and the second-generation mother gene sequence is transformed into the second-generation daughter gene sequence after the mutation process. .

In the above-mentioned genetic sequence mutation algorithm, the optimal combination analysis module 34 may also set the upper limit upper_bound and the lower limit lower, and then refer to the score calculated by substituting the gene sequence into the adaptation function for correction. For example, based on the mutation example described above, the optimal combination analysis module 34 can use the following formula to mutate, and then the second-generation mother gene sequence will be transformed into the second-generation daughter gene sequence after the mutation process. .

Please refer to FIG. 4 for a method for generating an adaptation function according to another embodiment. A neural network can be used as an adaptation function, and the adaptation function can be used to calculate a gene sequence score s. In this embodiment, taking driver i driving vehicle equipment j as an example, the neural network has a hidden layer, and the hidden layer has Neurons, the adaptation function is: Where the gene sequence score is s. In this embodiment, the lower the score, the better, that is, the best solution is , A set of gene sequences , This gene sequence contains Chromosomes (that is, the cardinality of the set ), Of which the first dyeing system The chromosome can be a floating-point number. Optimal combination analysis module 34 can use genetic algorithms to obtain a set of optimal gene sequences , And this best gene sequence corresponds to the best gene sequence score s (the lowest score), and the best gene sequence can be combined with the number of driving behaviors Estimate the amount of energy consumed.

In this embodiment, it is described that a neural network has a hidden layer, but it is not limited to this. The neural network may also have a plurality of hidden layers, and the weight value between neurons can be used as a chromosome. Based on the foregoing assumptions, a plurality of gene sequences are generated, and a set of optimal gene sequences can be obtained by using a genetic algorithm.

Please refer to FIG. 4 next is a method for generating a gene sequence according to an embodiment of the present invention. The optimal combination analysis module 34 is based on statistics of driving behaviors (for example, statistics of specific speed intervals, statistics of specific traffic-related information, statistics of specific physiological information, etc.) and statistics of evaluation information (for example, (The amount of energy consumption, etc.) to establish a plurality of target functions (step S510); the optimal combination analysis module 34 randomly generates a plurality of parameter values of each target function, and calculates those target functions to generate error values for each target function ( Step S511); the optimal combination analysis module 34 corrects the optimal solution of the parameter values of each target function according to the error value of each target function (step S512); the optimal combination analysis module 34 outputs the optimal solution of the parameter values of each target function Go to those other target functions (step S513), and recalculate the error values of each target function; the optimal combination analysis module 34 determines whether the error value of each target function is lower than the convergence threshold value (step S514, that is, determines whether to converge) ; When the error value of each target function is lower than the convergence threshold, the optimal combination analysis module 34 combines the parameter values with the smallest output error (step S515); otherwise, if each target function If the error value is higher than the convergence threshold value, the optimal combination analysis module 34 corrects the optimal solutions of the parameter values of each objective function according to the error values, and outputs the optimal solutions of the parameter values of each objective function to those other objective functions, and Calculate the error value of each objective function, and continue to calculate until convergence.

It should be noted that the above target functions can be established based on the number of driving behaviors and the amount of energy consumption in each month. In this embodiment, taking the driver i driving the vehicle equipment j as an example, the optimal combination analysis module 34 can generate a plurality of target functions in the following manner: The first target function: The second objective function: … The twelfth objective function: Thirteenth target function: Fourteenth objective function: .

In this method, the best combination analysis module 34 can also set an upper limit value upper_bound and a lower limit value lower_bound, and those parameters of the first objective function A value between the upper limit upper_bound and the lower limit lower_bound may be randomly generated. In addition, the optimal combination analysis module 34 calculates the error value and the correction parameter according to the target function after randomly generating those values. Initial value: .

According to the above calculation method, those parameters of the second objective function It is necessary to randomly generate values between the upper limit upper_bound and the lower limit lower_bound. The best combination analysis module 34 generates those values randomly, and then calculates the error value and correction parameters according to the target function Initial value: .

According to the above calculation method, and so on, those parameters of the fourteenth objective function It is necessary to randomly generate values between the upper limit upper_bound and the lower limit lower_bound. The best combination analysis module 34 generates those values randomly, and then calculates the error value and the correction parameter according to the target function. Initial value: .

After the initial value calculation is completed, the optimal combination analysis module 34 can output the optimal solution of the parameter values of each target function to those other target functions, and recalculate the error value of each target function. Taking the first objective function as an example, those parameters can be reset using the following methods, and the error value and the correction parameters are calculated according to the objective function. : .

According to the above calculation method, taking the second objective function as an example, those parameters can be reset using the following methods, and then the error value and the correction parameters are calculated according to the objective function. : .

According to the above calculation method, and so on. Taking the fourteenth objective function as an example, those parameters can be reset using the following methods, and the error value and the correction parameter are calculated according to the objective function. : .

According to the above calculation method, determine whether the error value of each objective function is lower than the convergence threshold. When the error value of each target function is lower than this convergence threshold, the optimal combination analysis module 34 outputs the parameter value combination with the smallest error. If the error value of each objective function is higher than the convergence threshold, the optimal combination analysis module 34 repeatedly performs parameter correction, corrects the optimal solution of each objective function parameter value according to the error value, and outputs the objective function parameter value. The best solution is to those other objective functions, and the error value of each objective function is recalculated, and the calculation is continued until the error value is lower than the convergence threshold.

Based on the aforementioned genetic algorithm, the present invention further proposes a method for estimating traffic information, as shown in FIG. 6. The data analysis server device 3 can collect the traffic information of the n road segments at the t-th time point (step S610). The traffic information may be travel time, traffic volume, or speed. For example, the traffic information of section 1 at the t time point is The traffic information of section 2 at the t time point is , ..., the traffic information of section n at time t is .

The optimal combination analysis module 34 then inputs the traffic information of the n road segments at the t time point into the optimal combination analysis method described in FIG. 2 (step S611). It is worth noting that at this time, the formula of the adaptation function generation method is , As shown in Figure 7. The best combination analysis module 34 can then use the best combination analysis method to obtain the set of weights And the gene sequence is used as the weight value of the fitness function (that is, the weight value of the traffic information association between each link) ), And the best combination is the link factor influence weight (step S612).

In addition, the present invention also provides a method for estimating physiological information. As shown in FIG. 8, the data analysis server device 3 can collect physiological information for driving m time points (step S810), and the physiological information can be a heart rate value or a heart rate variability. The value is shown in Figure 9.

The optimal combination analysis module 34 then inputs the physiological information reflected by the driving behavior at m time points into the optimal combination analysis method described in FIG. 2 (step S811), and the adaptation function is a one-variable n-th order equation , As shown in Figure 10.

Best combination analysis module 34 uses the best combination analysis method to obtain the weight set The best combination is the time factor (t) affecting the weight (that is, the weight value associated with the time series and the heart rate variation) (step S812).

It should be noted that in the evaluation methods shown in FIGS. 6 and 8, the optimal combination analysis module 34 may then perform a selection procedure, a mating procedure, and a mutation procedure on those gene sequences, and generate them when the scores of those gene sequences converge. Optimal gene sequence, and this optimal gene sequence is a collection of evaluation information for driving behavior.

To sum up, the optimal combination analysis method of the embodiment of the present invention can improve the genetic algorithm, and firstly establish several excellent gene sequences during the initialization phase, and then use these gene sequences to perform mating, mutation and other calculations to generate the best genes. Sequence, and this improved genetic algorithm can be combined with the adaptation function of the neural network. The embodiments of the present invention can be used for energy consumption estimation, traffic information estimation, and physiological information estimation according to the needs of users, so as to obtain the optimal gene sequence.

Although the present invention has been disclosed as above with the examples, it is not intended to limit the present invention. Any person with ordinary knowledge in the technical field can make some modifications and retouching without departing from the spirit and scope of the present invention. The protection scope of the present invention shall be determined by the scope of the attached patent application.

1‧‧‧Vehicle equipment

10, 20, 30, 40‧‧‧ communication module

12, 22, 32‧‧‧ Intermediary software modules

14‧‧‧ Positioning Module

20‧‧‧User Equipment

24‧‧‧user interface

34‧‧‧Best combination analysis module

40‧‧‧Database equipment

42‧‧‧ Computing Module

44‧‧‧Storage Module

S210 ~ S220, S510 ~ S515, S610 ~ S612, S810 ~ S812‧‧‧steps

~ Number of driving behaviors

~ , W ₁ ~ w _n ‧‧‧ gene sequence

‧‧‧Total energy consumption

~ , ~ , ~ , ~ ‧‧‧chromosome

~ ‧‧‧Traffic Information

t ⁰ ~ t ⁿ ‧‧‧ time factor

FIG. 1 is a system architecture diagram according to an embodiment of the present invention. FIG. 2 is a flowchart of an optimal combination analysis method according to an embodiment of the present invention. FIG. 3 is a schematic diagram of an adaptation function according to an embodiment of the present invention. FIG. 4 is a schematic diagram of an adaptation function according to another embodiment of the present invention. FIG. 5 is a flowchart of a gene sequence generation algorithm according to an embodiment of the present invention. FIG. 6 is a flowchart of an evaluation algorithm of traffic information related evaluation information according to an embodiment of the present invention. FIG. 7 is a schematic diagram of an adaptation function related to traffic information according to an embodiment of the present invention. FIG. 8 is a flowchart of an evaluation algorithm for evaluation information related to physiological information according to an embodiment of the present invention. FIG. 9 is a timing statistics chart of a heart rhythm value according to an embodiment of the present invention. FIG. 10 is a schematic diagram of an adaptation function related to physiological information according to an embodiment of the present invention.

Claims (10)

An optimal combination analysis method suitable for analyzing information about driving behavior response. The optimal combination analysis method includes: generating multiple gene sequences, each of which contains multiple chromosomes, and the chromosomes are related to driving behavior in The statistical quantity of evaluation information caused by different time points, and each of the statistical quantities is the quantity that the evaluation information meets a numerical interval at different time points; the traffic information or physiological information reflected by the driving behavior is input into an adaptation function Formula to calculate the scores of the gene sequences, wherein the gene sequences are used as weight values of the fitness function; and the gene sequences are subjected to selection procedures, mating procedures, and mutation procedures, and the scores of the gene sequences are used When converged, an optimal gene sequence is generated, wherein the optimal gene sequence is a collection of evaluation information of the driving behavior. The optimal combination analysis method according to item 1 of the scope of patent application, wherein the step of generating the optimal gene sequence when the scores of the gene sequences converge includes: outputting the optimal number of times when the number of evolutions is equal to the number of overlapping bands. A good gene sequence, in which the number of evolutions plus one is performed each time the selection procedure, the mating procedure, and the mutation program are performed; and when the number of evolutions is less than the number of overlaps, the number of evolutions is increased by one. The optimal combination analysis method according to item 1 of the scope of patent application, wherein generating the gene sequences includes: establishing a plurality of target functions according to the traffic information or physiological information of the driving behavior and the evaluation information; randomly generating each of the Multiple parameter values of the target function, and calculate each target function to generate an error value for each of the target functions; modify each of the target functions according to the error value of each of the target functions The optimal solution of the parameter value; outputting the optimal solution of the parameter value of each of the target functions to each of the other target functions, and recalculating the error value of each of the target functions; and determining each of the target functions Whether the error value of is lower than a convergence threshold value, where if it is lower than the convergence threshold value, the parameter value combination with the smallest error is output, and if it is higher than the convergence threshold value, each error is corrected according to the error value of the objective function. An optimal solution of each of the parameter values of the objective function, and output the optimal solution of each of the parameter values of each of the objective functions to each of the other objective functions and recalculate each of the objective functions Formula error To continuously calculate converges to below the threshold. According to the optimal combination analysis method described in item 1 of the patent application scope, generating the gene sequences includes: obtaining traffic information of different road sections at different points in time, wherein the traffic information is travel time, traffic volume, or speed. According to the optimal combination analysis method described in item 1 of the patent application scope, generating the gene sequences includes: obtaining physiological information at different time points, wherein the physiological information is a heart rhythm value or a heart rhythm variation value. A data analysis server device includes: a communication module that receives traffic information or physiological information reflected by driving behavior; a memory that records the traffic information or the physiological information and a plurality of modules; and a processor, Coupling the communication module and the memory, and accessing and executing the modules stored in the memory, the modules include: an optimal combination analysis module, executing: generating a plurality of gene sequences, wherein Each gene sequence contains multiple chromosomes, and the chromosomes are related to the statistical quantity of evaluation information caused by driving behavior at different time points, and each of the statistical quantities is a value interval where the evaluation information meets a numerical interval Quantity; inputting the traffic information or physiological information of the driving behavior into an adaptation function to calculate a score of the gene sequences, wherein the gene sequences are used as weight values of the adaptation function; and selecting the gene sequences Procedures, mating procedures, and mutation procedures, and generate an optimal basis when the scores of the gene sequences converge Sequence, wherein the optimal gene sequence is a collection of evaluation information of the driving behavior. The data analysis server device according to item 6 of the scope of patent application, wherein when an evolution number is equal to a stack number, the optimal combination analysis module outputs the optimal gene sequence, wherein each time the selection process is performed, The mating program and the mutation program, the optimal combination analysis module adds one to the number of evolutions; and when the number of evolutions is less than the number of overlaps, the best combination analysis module adds one to the number of evolutions. The data analysis server device described in item 6 of the scope of patent application, wherein the best combination analysis module creates a plurality of objective functions based on the traffic or physiological information of the driving behavior and the evaluation information, and randomly generates each of the objectives Complex parameter values of the function, and calculate each target function to generate an error value for each of the target functions, and correct each parameter value of each of the target functions according to the error value of each of the target functions The best solution, output the best solution of each parameter value of the target function to each other of the target function, and recalculate the error value of each target function to determine the error value of each target function Whether it is lower than a convergence threshold value, wherein if it is lower than the convergence threshold value, the parameter value combination with the smallest error is output; if it is higher than the convergence threshold value, each goal is corrected according to the error value of each goal function The optimal solution of each of the parameter values of the function, and output the optimal solution of each of the parameter values of each of the target functions to each of the other target functions and recalculate the error of each of the target functions Value on a continuous basis Calculate until it is below the convergence threshold. The data analysis server device as described in item 6 of the scope of patent application, wherein the best combination analysis module obtains traffic information of different road sections at different points in time through the communication module, and the traffic information is travel time and traffic volume , Or speed. The data analysis server device described in item 6 of the scope of the patent application, wherein the optimal combination analysis module obtains physiological information at different time points through the communication module, and the physiological information is a heart rhythm value or a heart rhythm variation value.