CN114564849A - Data-driven vehicle economy simulation test scene generation method - Google Patents
Data-driven vehicle economy simulation test scene generation method Download PDFInfo
- Publication number
- CN114564849A CN114564849A CN202210268449.0A CN202210268449A CN114564849A CN 114564849 A CN114564849 A CN 114564849A CN 202210268449 A CN202210268449 A CN 202210268449A CN 114564849 A CN114564849 A CN 114564849A
- Authority
- CN
- China
- Prior art keywords
- speed
- state
- scene
- vehicle
- sub
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/80—Technologies aiming to reduce greenhouse gasses emissions common to all road transportation technologies
- Y02T10/84—Data processing systems or methods, management, administration
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Computational Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Aviation & Aerospace Engineering (AREA)
- Automation & Control Theory (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a data-driven vehicle economy simulation test scene generation method, which comprises the following steps: based on the historical driving data of the vehicle, carrying out speed segment division on the driving speed of the vehicle, carrying out speed state division on the speed segments, defining a state space, and calculating the state transition probability among the speed states to obtain a Markov state transition probability matrix of the driving speed of the vehicle; meanwhile, semantic description of keywords is carried out on the speed fragment, scene elements are selected and reasonably combined with the scene elements, and sub-scenes are reproduced; and outputting a speed state chain based on the Markov state transition probability matrix, randomly selecting a sub-scene corresponding to the speed state, converting the speed state chain into a sub-scene chain, and outputting a complete simulation test scene. The invention randomly generates a simulation test scene by taking the dynamic transfer characteristics of the vehicle driving data as constraints, and improves the test level of the vehicle economy simulation test.
Description
Technical Field
The invention relates to the technical field of vehicle engineering, in particular to a data-driven vehicle economy simulation test scene generation method.
Background
The vehicle economy test method generally comprises a laboratory chassis dynamometer test, a road test, a simulation test and the like. Compared with road test, the test of the laboratory chassis dynamometer is carried out under a controlled condition, and the accuracy and the repeatability are relatively high; the simulation method can greatly save research time and cost because field experiments can not be carried out. In a vehicle economy simulation test, energy consumption or emission results of a vehicle are usually obtained based on driving conditions, the driving conditions of the vehicle reflect the most representative vehicle driving characteristics of a certain vehicle type, a certain road section or a certain area, but the energy consumption or emission results of the vehicle obtained based on the driving conditions are often greatly different from actual conditions.
Virtual simulation testing has recently received wide attention from the automobile industry as one of basic testing techniques for forming tests of automatically driven automobiles, and the construction of a simulation scene is an important component forming the simulation testing. In the driving field, a scene is considered as a comprehensive reflection of the driving environment and the driving behavior of the automobile in a certain time and space range, describing external roads, weather and traffic participants as well as the driving tasks and status information of the vehicle itself. In the research of an automatic driving test scene, a learner divides the scene generation research into two main methods of mechanism modeling and data-driven modeling, wherein the mechanism modeling method focuses on creating the scene based on theory, and the data-driven method focuses on realizing scene reproduction and key scene derivation based on data. Compared with the running condition, the simulation test scene is abstract and summary of the real world, especially scene reproduction to a certain degree can be realized based on data-driven scene generation, and the approximation degree of the vehicle economy simulation test and the real condition can be improved through reasonable scene construction and vehicle dynamics modeling.
How to extract the real-world vehicle driving rule and reproduce and synthesize the scene is the problem which needs to be solved based on the data-driven vehicle economy simulation test scene generation.
Disclosure of Invention
The invention aims to design and develop a vehicle economy simulation test scene generation method based on data driving, which extracts the dynamic transfer characteristics of the vehicle running speed by using a state transfer probability matrix and carries out sub-scene reproduction of speed segments to realize scene synthesis.
The technical scheme provided by the invention is as follows:
a data-driven vehicle economy simulation test scene generation method comprises the following steps:
step one, acquiring the running speed of a vehicle, and dividing the running speed of the vehicle into speed segments:
when v is less than or equal to 1.5km/h, the idle speed is a segment;
when v is more than 1.5km/h and a is more than 0.15m/s2To accelerate the segment;
when v is more than 1.5km/h and a is less than or equal to 0.15m/s2Is a deceleration segment;
when v > 1.5km/h and-0.15 m/s2＜a≤0.15m/s2The time is a uniform speed segment;
wherein v is the vehicle speed and a is the acceleration of the vehicle;
step two, calculating a state transition probability matrix of the speed segment, and performing sub-scene reproduction on the speed segment;
wherein the state transition probability matrix is:
where P is the state transition probability matrix, PijFor transition probability from state i to state j, state i and state j refer to any of n speed states, and i is 1,2, …, n, j is 1,2, …, n;
and thirdly, outputting a speed state chain based on the state transition probability matrix, randomly selecting a sub-scene corresponding to the speed state, and converting the speed state chain into a sub-scene chain to obtain a complete simulation test scene.
Preferably, the first step further comprises:
and keeping the type of the speed segment consistent with the two sides when the speed segment is less than 5s and the types of the speed segments before and after the speed segment are the same.
Preferably, the calculating the state transition probability matrix of the speed segment in the second step specifically includes:
in the formula, pijFor transition probability from state i to state j, NijThe frequency of the segment for the transition of state i to state j, state i and state j referring to any of the n speed states, ΣjNijThe frequency of the segments for all states to which state i can transition;
step 4, calculating a state transition probability matrix:
preferably, the sub-scene reproduction of the speed segment comprises:
semantic description of keywords and selection of scene elements.
Preferably, the semantic description of the keyword includes a sub-scene distance, a sub-scene termination condition and a scene element;
wherein the sub-scene distance is a driving distance of the speed segment;
the sub-scene termination condition is that termination is carried out according to a sub-scene termination time.
Preferably, the sub-scene end time satisfies:
in the formula, LsenarioFor a sub-scene distance,
and k is the number of the speed segment set of the speed segment in the speed state i.
Preferably, the scene elements include a static scene element, a dynamic scene element, a meteorological environment element, and a vehicle element to be measured.
Preferably, the static scene elements include:
road type, traffic facilities, geographic information, and static obstacles;
the dynamic scene elements include:
dynamic traffic identification and traffic participants;
the meteorological environment elements comprise:
precipitation, light, temperature, humidity and climate;
the vehicle element under test includes:
an initial state of the vehicle, a travel target of the vehicle, and a behavior element.
Preferably, the outputting the speed state chain based on the state transition probability matrix specifically includes:
selecting an initial speed state, generating random numbers uniformly distributed in (0,1) by a Monte Carlo method, and determining a next speed state to obtain a speed state chain.
Preferably, the randomly selecting the sub-scenes corresponding to the speed states specifically includes:
one speed segment from state 1 and speed range [0,5) is selected as the start of the speed state chain, and the speed difference between adjacent speed segments does not exceed 0.5 km/h.
The invention has the following beneficial effects:
(1) the method for generating the vehicle economy simulation test scene based on data driving, which is designed and developed by the invention, utilizes the Markov state transition probability matrix to extract the dynamic transition characteristic of the actual vehicle speed, thereby improving the similarity degree of the simulation test scene and the real world;
(2) compared with the vehicle economy simulation test based on the working condition, the randomness and the selectivity of the method for generating the vehicle economy simulation test scene based on the data drive are stronger, so that the simulation test result is closer to the actual result;
(3) the data-driven vehicle economy simulation test scene generation method is designed and developed, aiming at the sub-scene reproduction mode of the speed segment, the combination mode of scene elements is greatly reduced, and the scene generation efficiency is improved;
(4) the invention discloses a data-driven vehicle economy simulation test scene generation method which is designed and developed based on the invention, takes a scene as an information basis and a vehicle model as a material basis, provides more complete information environment for hardware-in-the-loop tests, driver-in-the-loop tests and other in-the-loop tests, and improves the vehicle economy test level.
Drawings
Fig. 1 is a schematic flow chart of a method for generating a simulation test scenario of vehicle economy based on data driving according to the present invention.
FIG. 2 is a schematic diagram of the vehicle speed correlation at a time scale of 1s according to the present invention.
FIG. 3 is a schematic diagram of the vehicle speed correlation at a time scale of 2s according to the present invention.
FIG. 4 is a schematic diagram of the vehicle speed correlation at a time scale of 5s according to the present invention.
FIG. 5 is a schematic diagram of the vehicle speed correlation at a time scale of 10s according to the present invention.
FIG. 6 is a schematic diagram of the velocity segmentation according to the present invention.
FIG. 7 is a diagram illustrating a Markov state transition probability distribution according to the present invention.
Fig. 8 is a schematic diagram of accelerating reproduction of a sub-scene according to the embodiment of the present invention.
Fig. 9 is a schematic diagram of the recurrence of the deceleration sub-scene in the embodiment of the present invention.
Fig. 10 is a schematic diagram of uniform-velocity sub-scene reproduction in the embodiment of the present invention.
Fig. 11 is a schematic view of a complete scene synthesis process according to the present invention.
Detailed Description
The present invention is described in further detail below in order to enable those skilled in the art to practice the invention with reference to the description.
As shown in fig. 1, the method for generating a simulation test scene based on data-driven vehicle economy provided by the invention is characterized in that based on historical driving data of a vehicle, speed segment division is performed on the driving speed of the vehicle, speed state division is performed on the speed segment, a state space is defined, and a state transition probability is calculated to obtain a markov state transition probability matrix of the driving speed of the vehicle; performing semantic description of keywords, selecting scene elements and reasonably combining the scene elements based on the speed segments, and reproducing the sub-scenes; and outputting a speed state chain based on the Markov state transition probability matrix, randomly selecting a sub-scene corresponding to the speed state, converting the speed state chain into a sub-scene chain, and outputting a complete simulation test scene.
Before describing the operation of the present invention in detail, first, the markov property of the vehicle running speed is verified:
let X (T) be a random process that varies with time T (T ∈ T), known at time T0In the state X (t)0) Then the process is at time t1In the state X (t)1) And time t0The previous state has nothing to do with the time t0In the state X (t)0) The characteristic is called non-aftereffect or Markov property, and has non-aftereffect or Markov propertyThe stochastic process of mutability is called markov process.
Describing Markov property by probability distribution function, and setting S as state space of random process X (t) at X (t)i)＝xi,xiE S, i is 1,2, …, n, for an arbitrary time ti(ti∈T,t1＜t2＜…tn) Having X (t)n) Is equal to X (t)n-1)＝xn-1When X (t)n) The conditional probability distribution function of (1), as follows:
P{X(tn)≤xn|X(t1)≤x1,…,X(tn-1)≤xn-1}＝P{X(tn)≤xn|X(tn-1)≤xn-1},xn∈S；
during the running of the vehicle, the vehicle speed of the current state of the vehicle is related only to the vehicle speed of the previous state, and is not related to the previous historical state. For the vehicle running speed VnHaving the formula:
P{V(tn)≤vn|V(t1)≤v1,…,V(tn-1)≤vn-1}＝P{V(tn)≤vn|V(tn-1)≤vn-1},vn∈S；
i.e. the vehicle driving process may be considered a markov process.
Calculating the correlation between two adjacent states at a certain time interval by using the Pearson correlation coefficient:
wherein X is a speed variable, and Y is a speed variable which is adjacent to X for a certain time scale.
TABLE 1 vehicle speed correlation coefficient for different time scales
As shown in fig. 2-5 and table 1, it can be demonstrated that on a small time scale (i.e., within 10 s), the vehicle driving process can be considered a markov process.
The invention specifically comprises the following steps:
step one, as shown in fig. 6, acquiring the running speed of the vehicle, and dividing the running speed of the vehicle into speed segments:
dividing constant speed and idle speed segments according to the acceleration as zero, dividing acceleration segments according to the acceleration as more than zero, dividing deceleration segments according to the acceleration as less than zero, but setting an acceleration threshold value when the acceleration is-0.15 m/s because the actually acquired vehicle running condition data is influenced by interference and equipment noise, the acceleration as zero almost does not exist and the vehicle speed fluctuates2To 0.15m/s2When the acceleration of the point is marked to be zero, the point is classified into a zero acceleration segment, in order to distinguish an idle segment from a non-idle segment, a speed threshold value is set to be 1.5km/h, and the speed segments are divided as follows:
(a) when v is less than or equal to 1.5km/h, the idle speed is a segment;
(b) when v is more than 1.5km/h and a is more than 0.15m/s2To accelerate the segment;
(c) when v is more than 1.5km/h and a is less than or equal to 0.15m/s2Is a deceleration segment;
(d) when v > 1.5km/h and-0.15 m/s2＜a≤0.15m/s2The time is a uniform speed segment;
where v is the vehicle speed and a is the acceleration of the vehicle.
In order to maintain the continuity of the vehicle state, the division result is corrected for a speed segment whose time length is too small: keeping the type of the speed segment consistent with the two sides when the speed segment is less than 5s and the front and rear speed segments are the same in type;
step two, as shown in fig. 7, calculating a state transition probability matrix of the speed segment, and performing sub-scene reproduction on the speed segment;
the calculating the state transition probability matrix of the speed segment specifically includes:
(1) and dividing speed state:
calculating the average speed of each speed segment, defining a state of the average speed of each speed segment according to the interval of 5km/h, and dividing the speed state into a state 1, a state 2, a state 3, a state … … and a state n;
(2) defining a state space:
all velocity states (state 1, state 2, state 3, … …, state n) constitute a state space;
(3) calculating the state transition probability:
the state transition frequency is a maximum likelihood estimation of the state transition probability, the number of speed segments in the speed state and the state relation of two adjacent segments are counted based on the n divided speed states, the probability is estimated by the frequency, the transition probability between different states is calculated, and the calculation formula is as follows:
in the formula, pijFor transition probability from state i to state j, NijThe frequency of the segment for the transition of state i to state j, state i and state j referring to any of the n speed states, ΣjNijFrequency count for segments of all states to which state i can transition;
(4) and forming a state transition probability matrix P by the one-step transition probabilities among all the speed states, and obtaining the state transition probability matrix as follows:
the sub-scene is a scene segment reproduced at a certain speed segment, and is the minimum unit for scene composition.
The sub-scene reproduction of the speed segment specifically comprises:
and performing semantic description on the keywords and selecting scene elements based on the speed segments, and reproducing the sub-scenes.
Wherein the semantic description of the keyword comprises:
1. sub-scene distance:
setting the driving distance of the speed segment as a sub-scene distance, wherein the sub-scene distance refers to the length of a road in a sub-scene;
2. sub-scene termination conditions:
setting the sub-scene termination condition to be time-based termination, wherein the sub-scene termination time is calculated according to the following formula:
in the formula, LsenarioFor a sub-scene distance,
the index i is the average speed of the speed segment, and the index j is the number of the speed segment set of the speed segment in the speed state i.
3. Scene elements:
the scene elements comprise static scene elements in a certain space range, dynamic scene elements in a certain space-time range, traffic participant elements, meteorological environment elements and detected vehicle elements.
The static scene elements comprise road types, traffic facilities, geographic information and static obstacles, the static scene elements are selected and attributes are set, and the attribute setting comprises the size, the position, the number and the like of the elements.
The dynamic scene elements comprise dynamic traffic indicating facilities and traffic participant elements, the dynamic traffic indicating facilities take traffic lights as examples, and the attributes are set as the size, the position and the conversion frequency of the traffic lights; traffic participant elements include motor vehicles, non-motor vehicles, pedestrians, and animals; the traffic participant elements are selected and set with attributes, wherein the attributes comprise size, position, distance/direction from the vehicle, speed (size/direction), acceleration, motion trail and the like.
Meteorological elements include light, temperature, humidity, climate, etc.
The tested vehicle elements comprise the initial state of the tested vehicle, the running target of the vehicle and the behavior elements.
Specifically, as shown in table 2:
TABLE 2 scene element Table
The selected scene elements are static scene elements, dynamic scene elements, traffic participant elements, meteorological environment elements and detected vehicle elements, and the scene elements are combined.
Step three: outputting a speed state chain based on a Markov state transition probability matrix, randomly selecting a sub-scene corresponding to the speed state, converting the speed state chain into a sub-scene chain, and outputting a complete simulation test scene:
(1) output speed state chain:
selecting an initial speed state, generating uniformly distributed random numbers according with a state transition probability matrix by adopting a Monte Carlo simulation mode, determining the next speed state, and knowing that the state transition probability of the state i to the states 1,2, 3, … and n is (p) through the calculated state transition probability matrixi1 pi2 … pin) And the probability accumulated value is 1, the interval (0,1) is divided into n intervals according to the probability value, the random numbers uniformly distributed in the interval (0,1) are generated by adopting a Monte Carlo method, the interval in which the random numbers fall corresponds to the speed state at the next moment, all subsequent speed states are generated according to the speed state, and the speed state chain with the required length is output.
(2) And randomly extracting a speed segment in each speed state:
selecting a speed segment from a low-speed state as the beginning of a speed state chain, wherein the speed difference between the front and rear adjacent segments is required to be not more than 0.5 km/h;
preferably, a speed segment from state 1, speed range [0,5) is selected as the beginning of the speed state chain.
(3) And randomly extracting the sub-scenes of each speed segment.
(4) And converting the speed state chain into a sub-scene chain, meeting the continuity requirement among the sub-scenes and outputting a synthesized complete scene.
Examples
The method comprises the steps of collecting the running speed of a vehicle, dividing the running speed of the vehicle into speed segments, defining the average speed of each speed segment as 5km/h, dividing the speed states into a state 1, a state 2, a state 3, … … and a state 12, wherein the range of the average speed of the speed segment in the state 1 is [0,5 ], the range of the average speed of the speed segment in the state 2 is [5,10 ], and the like, and the range of the average speed of the speed segment in the state 12 is [55, 60).
All velocity states (state 1, state 2, state 3, … …, state 12) constitute a state space.
Counting the number of speed segments in the speed state and the state relation of two adjacent segments based on the divided 12 speed states, calculating the transition probability among different states according to the frequency estimation probability, and forming a state transition probability matrix P by the one-step transition probability among all the speed states to obtain a state transition probability matrix:
semantic description of keywords is carried out based on the speed segments, and the keyword semantic description of one acceleration sub-scene can be that the sub-scene distance is 80m and the scene termination time is 10 s; on a double-lane straight road, the right side of a lane has a speed limit sign of 50km/h, and a small passenger car is positioned 20m right in front of a detected vehicle and runs at a constant speed of 30 km/h; the tested vehicle is positioned on a left lane, the initial speed is 30km/h, the vehicle accelerates to move to the right lane, and the speed is guaranteed to be below 50 km/h. "
As shown in fig. 8, selecting static scene elements according to semantic description of keywords: a three-lane straight lane (80 m); speed limit signs (50 km/h). Selecting dynamic scene elements according to the semantic description of the keywords: passenger car (<3.5 m). Auxiliary scene elements: green belts; a curb; a building is provided. And fourthly, combining scene elements by taking the initial position of the vehicle to be detected as a reference position: the speed limit sign is arranged on the right side of the lane and is 30m away from the vehicle to be detected; the passenger car is arranged 20m ahead of the tested vehicle and runs at a constant speed of 30 km/h.
The keyword semantic description of one speed-reducing sub-scene can be that the sub-scene distance is 100m, and the scene ending time is 15 s; on the two-lane straight road, a pedestrian crossing is arranged at a position 100m away from the detected vehicle, and an adult male with the height of 1.75m passes through the pedestrian crossing at the speed of 1.5m/s when the detected vehicle is 20m away from the pedestrian crossing; the tested vehicle is positioned on the right lane, and when the vehicle runs to 10m away from the crosswalk at a constant speed of 36km/h from the initial position, the vehicle decelerates at 18km/h until the vehicle decelerates to 0 in front of the crosswalk. "
As shown in fig. 9, firstly, static scene elements are selected according to semantic descriptions of keywords: a two-lane straight lane (100 m); pedestrian crossing traffic markings. Selecting dynamic scene elements according to the semantic description of the keywords: pedestrians (1.2 m or more). Auxiliary scene elements: green belts; a curb; a building is provided. And fourthly, combining scene elements by taking the initial position of the vehicle to be detected as a reference position: the pedestrian crosswalk traffic marking is arranged at a distance of 100m from the vehicle to be tested; the pedestrian is arranged at the right side of the running direction of the detected vehicle, the behavior is triggered based on an event, and the event is that the detected vehicle runs to a distance of 10m from the pedestrian crossing.
The keyword semantic description of a constant-speed sub-scene can be that the sub-scene distance is 100m and the scene termination time is 5 s; on a straight road of a single lane, the detected vehicle runs at a constant speed of 45km/h from the initial position, and no interference is caused to traffic participants in all directions. "
As shown in fig. 10, picking static scene elements according to semantic description of keywords: one-way lane straight (100 m). Selecting dynamic scene elements according to the semantic description of the keywords: none. Auxiliary scene elements: a green belt; a curb; a building is provided. And fourthly, combining scene elements by taking the initial position of the tested vehicle as a reference position.
And outputting a speed state chain based on the Markov state transition probability matrix, randomly selecting a sub-scene corresponding to the speed state, and converting the speed state chain into a sub-scene chain.
As shown in fig. 11, the synthesized complete scene is output.
The invention discloses a method for generating a vehicle economy simulation test scene based on data driving, which is designed and developed, utilizes a Markov state transition probability matrix to extract the dynamic transition characteristic of the actual vehicle speed, and improves the similarity degree of the simulation test scene and the real world; compared with the vehicle economy simulation test based on the working condition, the randomness and the selectivity are stronger, so that the simulation test result is closer to the actual result; aiming at the sub-scene reproduction mode of the speed segment, the combination mode of scene elements is greatly reduced, and the scene generation efficiency is improved; the method takes a scene as an information basis and a vehicle model as a material basis, provides more complete information environment for hardware-in-the-loop and driver-in-the-loop tests and the like, and improves the level of vehicle economy tests.
While embodiments of the invention have been described above, it is not intended to be limited to the details shown, particular embodiments, but rather to those skilled in the art, and it is to be understood that the invention is capable of numerous modifications and that various changes may be made therein without departing from the spirit and scope of the invention as defined by the appended claims and their equivalents.
Claims (10)
1. A data-driven vehicle economy simulation test scene generation method is characterized by comprising the following steps:
step one, acquiring the running speed of a vehicle, and dividing the running speed of the vehicle into speed segments:
when v is less than or equal to 1.5km/h, the idle speed is a segment;
when v is more than 1.5km/h and a is more than 0.15m/s2To accelerate the segment;
when v is more than 1.5km/hAnd a is less than or equal to 0.15m/s2Is a deceleration segment;
when v > 1.5km/h and-0.15 m/s2＜a≤0.15m/s2The time is a uniform speed segment;
wherein v is the vehicle speed and a is the acceleration of the vehicle;
step two, calculating a state transition probability matrix of the speed segment, and performing sub-scene reproduction on the speed segment;
wherein the state transition probability matrix is:
where P is the state transition probability matrix, PijFor transition probabilities from state i to state j, state i and state j refer to any of n speed states, and i is 1,2, …, n, j is 1,2, …, n;
and thirdly, outputting a speed state chain based on the state transition probability matrix, randomly selecting a sub-scene corresponding to the speed state, and converting the speed state chain into a sub-scene chain to obtain a complete simulation test scene.
2. The data-driven-based vehicle economy simulation test scenario generation method of claim 1, wherein the step one further comprises:
and keeping the type of the speed segment consistent with the two sides when the speed segment is less than 5s and the types of the speed segments before and after the speed segment are the same.
3. The method for generating the vehicle economy simulation test scenario based on data driving according to claim 2, wherein the calculating the state transition probability matrix of the speed segment in the second step specifically comprises:
step 1, calculating the average speed of each speed segment, defining a state of the average speed of each speed segment according to the interval of 5km/h, and dividing the speed state into a state 1, a state 2, a state 3, … … and a state n;
step 2, all speed states form a state space;
step 3, calculating the transition probability among different speed states:
in the formula, pijTransition probability, N, for transition from state i to state jijThe frequency of the segment for the transition of state i to state j, state i and state j referring to any of the n speed states, ΣjNijThe frequency of the segments for all states to which state i can transition;
step 4, calculating a state transition probability matrix:
4. the data-driven-based vehicle economy simulation test scenario generation method of claim 3, wherein the sub-scenario rendering of the speed segments comprises:
semantic description of keywords and selection of scene elements.
5. The data-driven-based vehicle economy simulation test scenario generation method of claim 4, wherein the semantic description of the keyword comprises sub-scenario distance, sub-scenario termination conditions, and scenario elements;
wherein the sub-scene distance is a driving distance of a speed segment;
the sub-scene termination condition is that termination is carried out according to a sub-scene termination time.
6. The data-driven-based vehicle economy simulation test scenario generation method of claim 5, wherein the sub-scenario termination time satisfies:
in the formula, LsenarioFor a sub-scene distance,
and k is the number of the speed segment set of the speed segment in the speed state i.
7. The data-driven-based vehicle economy simulation test scenario generation method of claim 6, wherein the scenario elements comprise static scenario elements, dynamic scenario elements, meteorological environment elements and vehicle under test elements.
8. The data-driven-based vehicle economy simulation test scenario generation method of claim 7, wherein the static scenario elements comprise:
road type, traffic facilities, geographic information, and static obstacles;
the dynamic scene elements include:
dynamic traffic identification and traffic participants;
the meteorological environment elements comprise:
precipitation, light, temperature, humidity and climate;
the vehicle element under test includes:
the initial state of the vehicle, the travel target of the vehicle, and the behavior elements.
9. The method for generating the vehicle economy simulation test scenario based on data driving according to claim 8, wherein the outputting the speed state chain based on the state transition probability matrix specifically comprises:
selecting an initial speed state, generating random numbers uniformly distributed in (0,1) by a Monte Carlo method, and determining a next speed state to obtain a speed state chain.
10. The data-driven-based vehicle economy simulation test scenario generation method of claim 9, wherein the randomly selecting the sub-scenario corresponding to the speed state specifically comprises:
one speed segment from state 1 and speed range [0,5) is selected as the start of the speed state chain, and the speed difference between adjacent speed segments does not exceed 0.5 km/h.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210268449.0A CN114564849A (en) | 2022-03-18 | 2022-03-18 | Data-driven vehicle economy simulation test scene generation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210268449.0A CN114564849A (en) | 2022-03-18 | 2022-03-18 | Data-driven vehicle economy simulation test scene generation method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114564849A true CN114564849A (en) | 2022-05-31 |
Family
ID=81720300
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210268449.0A Pending CN114564849A (en) | 2022-03-18 | 2022-03-18 | Data-driven vehicle economy simulation test scene generation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114564849A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115509909A (en) * | 2022-09-26 | 2022-12-23 | 北京百度网讯科技有限公司 | Test method, test device, electronic equipment and storage medium |
-
2022
- 2022-03-18 CN CN202210268449.0A patent/CN114564849A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115509909A (en) * | 2022-09-26 | 2022-12-23 | 北京百度网讯科技有限公司 | Test method, test device, electronic equipment and storage medium |
| CN115509909B (en) * | 2022-09-26 | 2023-11-07 | 北京百度网讯科技有限公司 | Test method, test device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106355907B (en) | Signalized crossing queue length real-time estimation method based on track of vehicle | |
| CN112133089B (en) | Vehicle track prediction method, system and device based on surrounding environment and behavior intention | |
| CN111178402B (en) | Scene classification method and device for road test of automatic driving vehicle | |
| CN110716529A (en) | Automatic generation method and device for automatic driving test case | |
| CN115081508B (en) | Traffic running risk parallel simulation system based on traffic digital twin | |
| CN111243338A (en) | Vehicle acceleration-based collision risk evaluation method | |
| CN113511204B (en) | Vehicle lane changing behavior identification method and related equipment | |
| CN106564503B (en) | The behavioural information for generating abnormal driving behavior determines method and device | |
| CN114564849A (en) | Data-driven vehicle economy simulation test scene generation method | |
| CN111824169A (en) | Method for reducing exhaust gas emissions of a drive system of a vehicle having an internal combustion engine | |
| CN115657002A (en) | Vehicle motion state estimation method based on traffic millimeter wave radar | |
| CN115795808A (en) | Automatic driving decision dangerous scene generation method, system, equipment and medium | |
| JP3520331B2 (en) | Traffic flow simulation device | |
| CN112849144B (en) | Vehicle control method, device and storage medium | |
| Scora et al. | Real-time roadway emissions estimation using visual traffic measurements | |
| CN112559968A (en) | Driving style representation learning method based on multi-situation data | |
| CN115114786B (en) | Assessment method, system and storage medium for traffic flow simulation model | |
| WO2023197409A1 (en) | Speed control model generation method and apparatus, and vehicle control method and apparatus | |
| CN115092138A (en) | Vehicle expressway lane change track planning method based on natural driver characteristics | |
| CN113361613B (en) | Method and device for classifying on-ramp vehicle lane change simulation model based on track data | |
| CN113268857B (en) | Urban expressway interweaving area microscopic traffic simulation method and device based on multiple intelligent agents | |
| CN109712425B (en) | Method and device for determining bus position based on sparse positioning point | |
| CN113284338A (en) | Method for calculating influence of motor vehicle emergency avoidance no-lamp control pedestrian crossing on traffic flow | |
| CN113673146A (en) | NSGA-II-based vehicle safety multi-objective optimization method | |
| Zhang et al. | A method for evaluating the complexity of test scenarios for autonomous vehicles |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |