CN109084992A - Method based on engine bench test unmanned vehicle intelligence - Google Patents
Method based on engine bench test unmanned vehicle intelligence Download PDFInfo
- Publication number
- CN109084992A CN109084992A CN201810842710.7A CN201810842710A CN109084992A CN 109084992 A CN109084992 A CN 109084992A CN 201810842710 A CN201810842710 A CN 201810842710A CN 109084992 A CN109084992 A CN 109084992A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- unmanned vehicle
- data
- main
- virtual
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M17/00—Testing of vehicles
- G01M17/007—Wheeled or endless-tracked vehicles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Traffic Control Systems (AREA)
Abstract
Method based on engine bench test unmanned vehicle intelligence, comprising the following steps: unmanned vehicle to be measured setting is subjected to drive simulating on testboard bay；Construct each data type in virtual scene and virtual scene；Main vehicle obtains the operating parameter of unmanned vehicle to be measured and the information of road surface of testboard bay simulation and simulates main vehicle into virtual scene；Main vehicle is interacted with each data type in virtual scene, generates virtual driving behavioral data；The driving behavior data that main vehicle is generated are as sample set, using random forests algorithm using the most classification of output times in decision tree as the classification of the test sample.Due to taking data-driven, the driving behavior data between main vehicle and each data type are acquired, so that data acquisition cost is lower, provided initial data authenticity with higher, the controllability of data realizes that detection scene is reappeared.
Description
Technical field
This application involves unmanned vehicle technology fields, and in particular to based on the intelligent method of engine bench test unmanned vehicle.
Background technique
As intelligence degree is continuously improved, more and more automatic driving vehicles are come out one after another.How to unmanned
The problem of behavior of vehicle carries out effectively verifying and assessment, is a worth further investigation.
The environmental model of the existing method simulation intelligent based on engine bench test unmanned vehicle is relatively simple, cannot sufficiently survey
Try the intelligent of unmanned vehicle.
Summary of the invention
The application provides a kind of method intelligent based on engine bench test unmanned vehicle, provides subject to more unmanned vehicle intelligence
True test.
According in a first aspect, provide a kind of method based on engine bench test unmanned vehicle intelligence in a kind of embodiment, including
Following steps: unmanned vehicle to be measured setting is subjected to drive simulating on testboard bay；It constructs in virtual scene and virtual scene
Each data type；Main vehicle obtains the operating parameter of unmanned vehicle to be measured and the information of road surface of testboard bay simulation and by main vehicle mould
Intend into virtual scene；Main vehicle is interacted with each data type in virtual scene, generates virtual driving behavioral data；It will
The driving behavior data that main vehicle generates are as sample set, using the random forests algorithm classification that output times in decision tree are most
Classification as the test sample.
Preferably, the data type in the virtual scene includes from vehicle module, environment module, road module, road surface mould
Block, main garage are control module, and each data type has an autonomous communication ability, and main garage is control module according to perceiving
Virtual scene event and own variable generate corresponding driving behavior data.
Preferably, the main garage is that control module is equipped with basic driving behavior library and rudimentary driving behavior library, Yi Jixiang
The triggering rule base of behavior is answered, main vehicle matches the virtual scene event perceived with the event in rule base, generates phase
The driving behavior data answered.
Preferably, described that unexpected incidents and uncertain event model are preset out of vehicle module, make main vehicle and occurs prominent
The slave vehicle interaction for sending out sexual behavior part or uncertain event generates driving behavior data.
Preferably, the main vehicle is equipped with virtual vision sensor, virtual radar sensor and virtual GPS biography with from vehicle module
Sensor.
Preferably, the environment module includes traffic sign, traffic marking, plant, building, bridge, tunnel model, row
People, every kind of model and main garage are the regular storehouse matching in control module.
Preferably, the road module includes highway, urban road, backroad model, every kind of model and main vehicle
Regular storehouse matching in behaviour control module.
Preferably, the main garage is that the rule base in control module is equipped with priority level, and the priority for the behavior that happens suddenly is high
It is higher than the priority of complex behavior in priority, the priority of simple behavior of medium-term and long-term behavior.
Preferably, sideways inclined simulation is realized by testboard bay, the testboard bay also acquires driving for unmanned vehicle to be measured
Sail behavioral data.
It preferably, further include being divided to unmanned vehicle intelligence grade, and set Vehicular intelligent evaluation index, intelligence
It can change after grade classification using the corresponding driving behavior data of evaluation index as training set, be exported using random forests algorithm intelligent
Classification after grade classification.
It is acquired according to the method based on engine bench test unmanned vehicle intelligence of above-described embodiment due to taking data-driven
Driving behavior data between main vehicle and each data type, so that data acquisition cost is lower, provided initial data tool
There is higher authenticity, the controllability of data can realize that detection scene is reappeared.Further, using with autonomous communication ability
It is interacted from vehicle module, environment module, road module with main vehicle, main vehicle is made to obtain more true driving behavior data, it is right
Unmanned vehicle intelligence provides more accurate test foundation.
Detailed description of the invention
Fig. 1 is an embodiment flow chart；
Fig. 2 is an embodiment from vehicle driving behavior control flow chart；
Fig. 3 is road module list in an embodiment；
Fig. 4 is intelligent grade separation table in an embodiment；
Fig. 5 is bagging algorithm flow chart；
Fig. 6 is the classifier based on SVM；
Fig. 7 is the schematic diagram calculated using the classifier of SVM unmanned vehicle intelligence.
Specific embodiment
Below by specific embodiment combination attached drawing, invention is further described in detail.Wherein different embodiments
Middle similar component uses associated similar element numbers.In the following embodiments, many datail descriptions be in order to
The application is better understood.However, those skilled in the art can recognize without lifting an eyebrow, part of feature
It is dispensed, or can be substituted by other elements, material, method in varied situations.In some cases, this Shen
Please it is relevant it is some operation there is no in the description show or describe, this is the core in order to avoid the application by mistake
More descriptions are flooded, and to those skilled in the art, these relevant operations, which are described in detail, not to be necessary, they
Relevant operation can be completely understood according to the general technology knowledge of description and this field in specification.
It is herein component institute serialization number itself, such as " first ", " second " etc., is only used for distinguishing described object,
Without any sequence or art-recognized meanings.And " connection ", " connection " described in the application, unless otherwise instructed, include directly and
It is indirectly connected with (connection).
Referring to FIG. 1, method of the application based on engine bench test unmanned vehicle intelligence, comprising the following steps:
101, unmanned vehicle to be measured setting is subjected to drive simulating on testboard bay.
It is the experiment vehicle for being placed in testboard bay by unmanned vehicle to be measured, unmanned vehicle to be measured is verified and assessment
Object.Testboard bay is test pavement simulating system, provides the simulation of rolling, three course, pitching angles for unmanned vehicle to be measured, simulation
State of motion of vehicle in reality scene.
102, each data type in server construction virtual scene and virtual scene.
Data type is that by autonomous, signal procedure function calculation procedure.Be arranged in virtual scene it is some from
It controls, the model of communication function, so that entire environment has intelligence, is created condition to verify with assessment.In one embodiment,
It is control module, each data type that data type, which includes from vehicle module, environment module, road module, road surface module, main garage,
With autonomous communication ability, main garage is that control module generates accordingly according to the virtual scene event and own variable perceived
Driving behavior data.
The intelligent behavior of unmanned vehicle to be measured is mainly reflected in the reaction made to sudden, uncertain event.Therefore,
When constructing virtual scene, it is necessary to set the model of different degrees of sudden or uncertain event.Out of vehicle module
Default unexpected incidents and uncertain event model, on the one hand can the truer main vehicle of simulation real traffic environment,
On the other hand it is that virtual scene introduces different sudden and uncertain event models, makes main vehicle and unexpected incidents occur
Or the slave vehicle interaction of uncertain event generates driving behavior data, can be used as verifying and assesses unmanned vehicle intelligent behavior to be measured
Part foundation.
From vehicle module can simulate the driving behavior under different drivers, same driver's different conditions, for main vehicle from
Main driving brings different difficulty, this also provides a variety of conditions with assessment for the intelligent behavior verifying of unmanned vehicle to be measured.
From vehicle module it is that each from vehicle module has preset itself speed, acceleration, steering wheel angle from vehicle, therefore main
Garage is the design of the triggering rule base of the behavior of control module it needs to be determined that the speed of main vehicle, position, range information with from vehicle
Speed, acceleration, the relationship between steering wheel angle.As shown in Fig. 2, first according to the initialization determined from vehicle module from vehicle
The parameters such as speed, acceleration, position, then constantly call the ambient enviroment sensing module of virtual scene, once detect master
Vehicle calls the unexpected incidents plot of setting in detection range, goes out together " examination paper " to main vehicle, verifies in this way for intelligent behavior
Condition is provided with assessment.It, will be according to normal detection method, to preceding, left and right inspection if not main vehicle occurs within the set range
Whether have barrier, execute corresponding controlling behavior if surveying.
103, main vehicle obtains the operating parameter of unmanned vehicle to be measured and the information of road surface of testboard bay simulation and simulates main vehicle
Into in virtual scene, main vehicle is interacted with each data type in virtual scene, generates virtual driving behavioral data.
Main garage is the triggering that control module is equipped with basic driving behavior library and rudimentary driving behavior library and respective behavior
Rule base, main vehicle match the virtual scene event perceived with the event in rule base, generate corresponding driving behavior
Data.Virtual unmanned vehicle basic act is contained in rudimentary driving behavior library, such as is turned to, advances, retreated.Rudimentary behavior is just
The automatic Pilot behavior of normal scene, is mainly used for timely reaction of the main vehicle to normal traffic scene.Fundamental reaction behavior is deposited in library
The motivation class behavior of virtual unmanned vehicle has been stored up, such as: it independently perceives, make decisions on one's own.Rudimentary behavior layer is adopted with basic act interlayer
It is connected with containment type structure, fundamental reaction behavior contains the function of rudimentary behavior, together using rudimentary driving behavior as basic unit
When may be constructed more complicated task level behavior again.All behaviors trigger rule base and virtual scene event matches by behavior
Carry out automatic Pilot control.The parallel of rudimentary behavior and fundamental reaction behavior may be implemented in vehicle main in this way, both ensure that burst
The real time reaction of traffic scene, and it can be given full play to and recognize planning ability, realize the automatic Pilot of vehicles in complex traffic scene.
Main garage is that the rule base in control module is equipped with priority level, and the priority for the behavior that happens suddenly is higher than medium-term and long-term behavior
Priority, the priority of simple behavior be higher than complex behavior priority.
Main garage is that control module autonomous can be communicated with main vehicle, and record in main vehicle and virtual scene each model it
Between the behavior that occurs.Find that interbehavior has occurred in main vehicle and certain module by detection trigger, the result of processing will be transmitted
To intelligent behavior verifying and evaluation module, foundation is provided with assessment for the intelligent behavior verifying of vehicle.
The motivation class fundamental reaction behavior of virtual unmanned vehicle (including main vehicle and from vehicle), which mainly has, at present turns on light automatically, is U-shaped
Turn around, signal lamp identification, variation lane, with speeding, overtaking other vehicles, being driven out to rotary island, parking stall identification, side coil park, vertically park, tightly
Anxious braking, lane departure warning, lane are kept etc., and encapsulate as opened/stopping, left/right turn, forward/backward, plus/minus is fast
Etc. rudimentary reflex action.These behaviors are indicated with parameterized form, are stored in respective behavior library.
The action selection mechanism of virtual unmanned vehicle determines its behaviour decision making process, is the pass for realizing its high degree of autonomy
Key.Virtual unmanned vehicle behavior controller is responsible for selection, activation and the termination of fundamental reaction behavior by action selection mechanism.
In simulation process, the perception event of task level complex behavior, virtual environment that virtual unmanned vehicle cognition module is planned
And real time control command of user etc. is used as external event (Event), the current internal state value with virtual unmanned vehicle
(InnerState) current demand of virtual unmanned vehicle is expressed as the input of behavioral module together.When certain demand is more than it
When action selection threshold values, corresponding rule of conduct will be triggered.If simple or urgent situation then triggers rudimentary behavior,
Timely reaction is generated to act；If medium-term and long-term complex behavior, then behavior controller selects one or one according to respective rule
Group fundamental reaction behavior, forms action command.Action command is realized through motion layer, to meet the current demand of visual human.One
Denier demand is met, and built-in attribute value will gradually resume normal level；If more than one demand needs to handle, by preferential
Grade sequence.The basic principle that priority determines is that most important demand will obtain highest priority, simple behavior priority grade high level cadre
The priority of complex behavior, the priority of long-term event in emergency event priority high level cadre.Action selection rule specifies behavior
Activation condition and its triggering result.
For the behavioural characteristic of true representation unmanned vehicle, the virtual unmanned vehicle of virtual scene can automatically perceive external dynamic
The variation of environment and therein attribute autonomously decides on behavior according to current goal or demand, while can also be with other
Virtual unmanned vehicle or user exchange information, change its own state.By effectively behaviour control, which can not only be to burst
Event carries out real time reaction, and has stronger cognition planning ability, can generate unmanned vehicle behavior true to nature.
Virtual unmanned vehicle obtains the stimulation of external Virtual environmental information and internal vehicle body state by sensing module in real time.It is right
In the perception of external Virtual environment, virtual unmanned vehicle is equipped with virtual vision sensor, virtual radar sensor and virtual GPS sensing
Device can get its current location and direction, can perceive the letter of stationary body in virtual environment, dynamic object and the traffic scene that happens suddenly
Breath wherein, completed by tested vehicle completely, transmits relevant data to server by thread by the cognition partial function of main vehicle
In.
Environment module includes traffic sign, traffic marking, plant, building, bridge, tunnel model, pedestrian, every kind of model with
Main garage is the regular storehouse matching in control module.Main vehicle is in autonomous driving, it has to be possible to accurately identify ambient enviroment mould
Type, could embody has certain intelligent behavior.
Road module includes highway, urban road, backroad model, and every kind of model and main garage are control module
In regular storehouse matching.Fig. 3 lists 3 kinds of ambient enviroment models, and model 1 is based on highway environment, has lane line without pedestrian,
It is required that unmanned vehicle to be measured has automatic Pilot, identifies lane line, function of overtaking other vehicles；Model 2 is avenue environment, has lane line simultaneously
And there is pedestrian, it is desirable that unmanned vehicle to be measured must identify in addition to having the function under the conditions of model 1 and avoid pedestrian automatically；Model 3
Highest is required to the intelligent behavior of vehicle, is had outside the function under the conditions of model 2, it is also necessary to identify unstructured road.In short,
Different ambient enviroment modules is established to meet the verifying and evaluation test of different intelligent degree vehicle behavior.
Main vehicle real-time perception information of road surface, the gradient, side tilt angle including road surface, is then transferred to test pavement simulating system.
Therefore main vehicle needs to carry out real-time perfoming detection to topographical surface, obtains the altitude data of four wheels, according to four wheels
Height can calculate the gradient, the side tilt angle on road surface.From vehicle behavior simulation be also required to carry out ambient enviroment model perception with
Detection, to realize automatic Pilot.
In one embodiment, by testboard bay realize sideways inclined simulation, the testboard bay also acquire it is to be measured nobody
The driving behavior data of vehicle.
104, the driving behavior data that intelligent behavior verifying and evaluation module generate main vehicle are as sample set, using random
Forest algorithm is using the most classification of output times in decision tree as the classification of the test sample.
It in one embodiment, further include being divided to unmanned vehicle intelligence grade, and setting Vehicular intelligent is evaluated
Index, it is defeated using random forests algorithm using the corresponding driving behavior data of evaluation index as training set after intelligent grade classification
Classification after intelligent grade classification out.
As shown in figure 4, the unmanned vehicle level of intelligence division based on engine bench test will be according to the intelligent attribute of unmanned vehicle
Grade classification is foundation, accomplishes to be entered tired by letter, realized the staged test assignment and testing requirement of engine bench test by easy entry hardly possible.
At present in terms of intelligent automobile intelligence, in the industry it is commonly accepted that U.S.'s SAE hierarchical definition.Based on this and combine existing rank
The complexity of Duan Zhongguo road traffic condition, it can be divided into driving auxiliary (DA), part automatic Pilot (PA), have ready conditions and drive automatically
Sail (CA), highly automated driving (HA), fully automated five grades (as shown in the table) of driving (FA).And it is with this five ranks
The output of unmanned vehicle intelligent quantization, each behavior are fed back to input, and establish unmanned vehicle intelligence analysis and assessment model to vehicle
Intelligence is evaluated.
Random forests algorithm is proposed by Leo Breiman and Adele Cutler.The algorithm combines Breimans's
" random subspac " method of " Bootstrap aggregating " thought and Ho.Its essence is one include multiple rhymed formula
The classifier of plan tree, these decision trees have been formed by random method, therefore also referred to as stochastic decision tree, in random forest
Tree between be not associated.It is in fact exactly that each decision tree is allowed to classify when test data enters random forest,
Finally taking that class that classification results are most in all decision trees is final result.Therefore it includes multiple determine that random forest, which is one,
The classifier of plan tree, and its output classification be by set the classification of output individually mode depending on.
Bootstrap method resampling
If containing n different sample { x1, x2, xn } in set S, if being taken out from set S with putting back to every time
Take a sample, extract n times altogether, form new set S, then in set S do not include some sample xi (i=1,2,
3, n) probability is
As n → ∞, have
Therefore, (all it is n), is cheated in new set and contain duplicate sample although the total sample number of new set s is equal
(putting back to extraction) only contains about 1-0.368*100%=in former set S if going out duplicate sample in new set S
63.2% sample.
Bagging algorithm is summarized
Bagging (abbreviation of Bootstrap aggregating) algorithm is earliest Ensemble Learning Algorithms, is thought substantially
Road is as shown in Figure 5.Specific steps can be described as:
Using Bootstrap method resampling, T training set S1, S2 ..., ST is randomly generated；
Using each training set, corresponding decision tree C1, C2 ..., Ct are generated；
It for test set sample X, is tested using each decision tree, obtains corresponding classification C1 (X), C2 (X) ...,
Cr(X)；
Using the method for ballot, most classifications will be exported in T decision tree as classification belonging to test set sample X.
The algorithm flow of random forest
Random forests algorithm is similar with Bagging algorithm, is all based on Bootstrap method resampling, generates multiple instructions
Practice collection.Unlike, random forests algorithm is when constructing decision tree, using the method for randomly selecting Split Attribute collection.
Detailed random forests algorithm process is as follows (might as well to set the attribute number of sample as M, m is greater than zero and whole less than M
Number):
Using Bootstrap method resampling, T training set S1, S2 ..., ST is randomly generated.
Using each training set, corresponding decision tree C1, C2 ..., CT are generated；At each non-leaf nodes (internal node)
Before upper selection attribute, Split Attribute collection of the m attribute as present node is randomly selected, and from M attribute with this m attribute
In best divisional mode divided that (in general, in the growth course of entire forest, the value of m is maintained not to the node
Become).
Each tree is all completely grown up, and without beta pruning.
It for test set sample X, is tested using each decision tree, obtains corresponding classification C1 (X), Cz (X) ...,
CT(X)。
Using the method for ballot, most classifications will be exported in T decision tree as classification belonging to test set sample X.
The architecture of classifier design based on support vector machines, support vector machines is as shown in Figure 6.Wherein, b is biasing
Parameter.
Unmanned vehicle test data will be from the real data that the gradational vehicle of tool measures on testboard.Data
Comprising specific amount of sample and specific characteristic component, level of intelligence will be provided as the class label of each sample.Just
Formula will be trained available disaggregated model to SVM (support vector machines) before being classified, the model pair recycled
Test set, which carries out class label prediction, can be obtained level of intelligence evaluation.
Intelligent evaluation rubric
It will be tested on rack by the intelligent vehicle for the different intelligent rank assert.Pass through a large amount of multiple repetition inspections
It surveys, completes the acquisition to the different scenes data generated in primary server.Using the data of acquisition as training set to SVM classifier
It is trained, determines optimal punishment parameter and function parameter.Next step collecting test data, test set sample are sent to decision
Tree selects mode according to ballot, determines the intelligence to tested vehicle.
Model overall flow
Vehicular intelligent evaluation index is divided into the advanced behavior j of vehicle basic act i, vehicle, basic traffic behavior k and advanced
Traffic behavior m sets up separately fixed with parameter: straight way lane keeps i1, stop line parking i2, U-shaped curved i3, speed limit i4, evacuation static-obstacle
i5；The advanced behavior j of vehicle is set separately are as follows: bend lane keeps j1, vehicle sound instruction j2, intersection passing j3, dynamic to advise
Draw j4, GPS navigation performance j5；Basic traffic behavior: forbid drive in the wrong direction k1, spacing holding k2；Advanced traffic behavior: traffic sign is known
Other m1, signal lamp identify m2, emergency braking m3.
Driving behavior data include specific amount of sample and specific characteristic component, and level of intelligence is as each sample
Class label.Available disaggregated model is trained to SVM (support vector machines) before formally being classified, recycled
Obtained model, which carries out class label prediction to test set, can be obtained level of intelligence evaluation.
As shown in fig. 7, level of intelligence evaluation Y=[y1, y2, y3, y4, y5], y1 drives auxiliary (DA), the part y2 is driven automatically
Sail (PA), y3 has ready conditions the highly automated driving (HA) of automatic Pilot (CA), y4, the fully automated driving (FA) of y5.
Wherein, X (i1) -- X (m3) is the sample set acquired under each intelligent evaluation index；K () is kernel function；X is spy
Vector is levied, X (i1) is respectively corresponded -- X (m3)；Xi1--Xm3For the parameter of corresponding vehicle behavior.
Use above specific case is illustrated the present invention, is merely used to help understand the present invention, not to limit
The system present invention.For those skilled in the art, according to the thought of the present invention, can also make several simple
It deduces, deform or replaces.
Claims (10)
1. the method based on engine bench test unmanned vehicle intelligence, it is characterised in that the following steps are included:
Unmanned vehicle to be measured setting is subjected to drive simulating on testboard bay；
Construct each data type in virtual scene and virtual scene；
Main vehicle obtains the operating parameter of unmanned vehicle to be measured and the information of road surface of testboard bay simulation and simulates main vehicle into virtual field
Jing Zhong；Main vehicle is interacted with each data type in virtual scene, generates virtual driving behavioral data；
The driving behavior data that main vehicle is generated are most by output times in decision tree using random forests algorithm as sample set
Classification of the classification as the test sample.
2. the method as described in claim 1 based on engine bench test unmanned vehicle intelligence, it is characterised in that: the virtual scene
In data type include from vehicle module, environment module, road module, road surface module, main garage be control module, each data
Type has autonomous communication ability, and main garage is that control module is generated according to the virtual scene event and own variable perceived
Corresponding driving behavior data.
3. the method as claimed in claim 2 based on engine bench test unmanned vehicle intelligence, it is characterised in that: the main garage is
Control module is equipped with the triggering rule base of basic driving behavior library and rudimentary driving behavior library and respective behavior, and main vehicle will be felt
The virtual scene event known is matched with the event in rule base, generates corresponding driving behavior data.
4. the method as claimed in claim 2 based on engine bench test unmanned vehicle intelligence, it is characterised in that: described from vehicle module
Interior default unexpected incidents and uncertain event model make main vehicle and the slave vehicle of unexpected incidents or uncertain event occur
Interaction generates driving behavior data.
5. the method as described in claim 1 based on engine bench test unmanned vehicle intelligence, it is characterised in that: the main vehicle with from
Vehicle module is equipped with virtual vision sensor, virtual radar sensor and virtual GPS sensor.
6. the method as claimed in claim 3 based on engine bench test unmanned vehicle intelligence, it is characterised in that: the environment module
Including traffic sign, traffic marking, plant, building, bridge, tunnel model, pedestrian, every kind of model and main garage are control module
In regular storehouse matching.
7. the method as claimed in claim 3 based on engine bench test unmanned vehicle intelligence, it is characterised in that: the road module
Including highway, urban road, backroad model, every kind of model and main garage are the regular storehouse matching in control module.
8. the method as claimed in claim 3 based on engine bench test unmanned vehicle intelligence, it is characterised in that: the main garage is
Rule base in control module is equipped with priority level, and the priority for the behavior that happens suddenly is higher than the priority of medium-term and long-term behavior, simple row
For priority be higher than complex behavior priority.
9. the method as claimed in claim 2 based on engine bench test unmanned vehicle intelligence, it is characterised in that: pass through testboard bay
Realize sideways inclined simulation, the testboard bay also acquires the driving behavior data of unmanned vehicle to be measured.
10. the method as described in claim 1 based on engine bench test unmanned vehicle intelligence, it is characterised in that: further include to nothing
People's vehicle intelligence grade divides, and setting Vehicular intelligent evaluation index, with evaluation index after intelligent grade classification
Corresponding driving behavior data are training set, the classification after intelligent grade classification is exported using random forests algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810842710.7A CN109084992B (en) | 2018-07-27 | 2018-07-27 | Method for testing intelligence of unmanned vehicle based on rack |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810842710.7A CN109084992B (en) | 2018-07-27 | 2018-07-27 | Method for testing intelligence of unmanned vehicle based on rack |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109084992A true CN109084992A (en) | 2018-12-25 |
| CN109084992B CN109084992B (en) | 2020-06-16 |
Family
ID=64831053
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810842710.7A Active CN109084992B (en) | 2018-07-27 | 2018-07-27 | Method for testing intelligence of unmanned vehicle based on rack |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109084992B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110864913A (en) * | 2019-11-28 | 2020-03-06 | 苏州智加科技有限公司 | Vehicle testing method and device, computer equipment and storage medium |
| CN110955972A (en) * | 2019-11-28 | 2020-04-03 | 苏州智加科技有限公司 | Virtual scene generation method and device, computer equipment and storage medium |
| CN111079800A (en) * | 2019-11-29 | 2020-04-28 | 上海汽车集团股份有限公司 | Acceleration method and acceleration system for intelligent driving virtual test |
| CN111178402A (en) * | 2019-12-13 | 2020-05-19 | 赛迪检测认证中心有限公司 | Scene classification method and device for road test of automatic driving vehicle |
| CN111649957A (en) * | 2020-06-08 | 2020-09-11 | 山东省交通规划设计院有限公司 | Tunnel environment automatic driving vehicle driving capability test system and test method |
| CN111881031A (en) * | 2020-07-23 | 2020-11-03 | 深圳慕智科技有限公司 | Intelligent transportation software and hardware precision disturbance method library and risk index construction method |
| CN112230228A (en) * | 2020-09-30 | 2021-01-15 | 中汽院智能网联科技有限公司 | Intelligent automobile vision sensor testing method based on field testing technology |
| CN112373483A (en) * | 2020-11-23 | 2021-02-19 | 浙江天行健智能科技有限公司 | Vehicle speed and steering prediction method based on forward neural network |
| CN112912883A (en) * | 2021-02-07 | 2021-06-04 | 华为技术有限公司 | Simulation method and related equipment |
| CN113219944A (en) * | 2021-04-27 | 2021-08-06 | 同济大学 | Intelligent vehicle control strategy test platform for mixed traffic flow working condition |
| CN113627740A (en) * | 2021-07-20 | 2021-11-09 | 东风汽车集团股份有限公司 | Driving load evaluation model construction system and construction method |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102393744A (en) * | 2011-11-22 | 2012-03-28 | 湖南大学 | Navigation method of pilotless automobile |
| CN103234763A (en) * | 2013-04-09 | 2013-08-07 | 北京理工大学 | System and method for quantitatively evaluating unmanned vehicles |
| CN103335853A (en) * | 2013-07-18 | 2013-10-02 | 中国科学院自动化研究所 | Unmanned driving vehicle cognitive competence testing system and method |
| CN104477167A (en) * | 2014-11-26 | 2015-04-01 | 浙江大学 | Intelligent driving system and control method thereof |
| CN107024927A (en) * | 2016-02-01 | 2017-08-08 | 上海无线通信研究中心 | A kind of automated driving system and method |
| CN107272687A (en) * | 2017-06-29 | 2017-10-20 | 深圳市海梁科技有限公司 | A kind of driving behavior decision system of automatic Pilot public transit vehicle |
| CN107272683A (en) * | 2017-06-19 | 2017-10-20 | 中国科学院自动化研究所 | Parallel intelligent vehicle control based on ACP methods |
| CN107526906A (en) * | 2017-10-11 | 2017-12-29 | 吉林大学 | A kind of driving style device for identifying and method based on data acquisition |
| CN107918392A (en) * | 2017-06-26 | 2018-04-17 | 怀效宁 | A kind of personalized driving of automatic driving vehicle and the method for obtaining driver's license |
| CN107957583A (en) * | 2017-11-29 | 2018-04-24 | 江苏若博机器人科技有限公司 | A kind of round-the-clock quick unmanned vehicle detection obstacle avoidance system of Multi-sensor Fusion |
| GB2557252A (en) * | 2016-12-02 | 2018-06-20 | Christian Schwazl | Physical environment simulator for vehicle testing |
| CN108267325A (en) * | 2018-01-29 | 2018-07-10 | 上海测迅汽车科技有限公司 | Unmanned vehicle material object is in ring test method |
| CN108267322A (en) * | 2017-01-03 | 2018-07-10 | 北京百度网讯科技有限公司 | The method and system tested automatic Pilot performance |
-
2018
- 2018-07-27 CN CN201810842710.7A patent/CN109084992B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102393744A (en) * | 2011-11-22 | 2012-03-28 | 湖南大学 | Navigation method of pilotless automobile |
| CN103234763A (en) * | 2013-04-09 | 2013-08-07 | 北京理工大学 | System and method for quantitatively evaluating unmanned vehicles |
| CN103335853A (en) * | 2013-07-18 | 2013-10-02 | 中国科学院自动化研究所 | Unmanned driving vehicle cognitive competence testing system and method |
| CN104477167A (en) * | 2014-11-26 | 2015-04-01 | 浙江大学 | Intelligent driving system and control method thereof |
| CN107024927A (en) * | 2016-02-01 | 2017-08-08 | 上海无线通信研究中心 | A kind of automated driving system and method |
| GB2557252A (en) * | 2016-12-02 | 2018-06-20 | Christian Schwazl | Physical environment simulator for vehicle testing |
| CN108267322A (en) * | 2017-01-03 | 2018-07-10 | 北京百度网讯科技有限公司 | The method and system tested automatic Pilot performance |
| CN107272683A (en) * | 2017-06-19 | 2017-10-20 | 中国科学院自动化研究所 | Parallel intelligent vehicle control based on ACP methods |
| CN107918392A (en) * | 2017-06-26 | 2018-04-17 | 怀效宁 | A kind of personalized driving of automatic driving vehicle and the method for obtaining driver's license |
| CN107272687A (en) * | 2017-06-29 | 2017-10-20 | 深圳市海梁科技有限公司 | A kind of driving behavior decision system of automatic Pilot public transit vehicle |
| CN107526906A (en) * | 2017-10-11 | 2017-12-29 | 吉林大学 | A kind of driving style device for identifying and method based on data acquisition |
| CN107957583A (en) * | 2017-11-29 | 2018-04-24 | 江苏若博机器人科技有限公司 | A kind of round-the-clock quick unmanned vehicle detection obstacle avoidance system of Multi-sensor Fusion |
| CN108267325A (en) * | 2018-01-29 | 2018-07-10 | 上海测迅汽车科技有限公司 | Unmanned vehicle material object is in ring test method |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110864913B (en) * | 2019-11-28 | 2021-09-03 | 苏州智加科技有限公司 | Vehicle testing method and device, computer equipment and storage medium |
| CN110955972A (en) * | 2019-11-28 | 2020-04-03 | 苏州智加科技有限公司 | Virtual scene generation method and device, computer equipment and storage medium |
| CN110864913A (en) * | 2019-11-28 | 2020-03-06 | 苏州智加科技有限公司 | Vehicle testing method and device, computer equipment and storage medium |
| CN111079800A (en) * | 2019-11-29 | 2020-04-28 | 上海汽车集团股份有限公司 | Acceleration method and acceleration system for intelligent driving virtual test |
| CN111079800B (en) * | 2019-11-29 | 2023-06-23 | 上海汽车集团股份有限公司 | Acceleration method and acceleration system for intelligent driving virtual test |
| CN111178402A (en) * | 2019-12-13 | 2020-05-19 | 赛迪检测认证中心有限公司 | Scene classification method and device for road test of automatic driving vehicle |
| CN111178402B (en) * | 2019-12-13 | 2023-04-07 | 赛迪检测认证中心有限公司 | Scene classification method and device for road test of automatic driving vehicle |
| CN111649957A (en) * | 2020-06-08 | 2020-09-11 | 山东省交通规划设计院有限公司 | Tunnel environment automatic driving vehicle driving capability test system and test method |
| CN111881031A (en) * | 2020-07-23 | 2020-11-03 | 深圳慕智科技有限公司 | Intelligent transportation software and hardware precision disturbance method library and risk index construction method |
| CN112230228A (en) * | 2020-09-30 | 2021-01-15 | 中汽院智能网联科技有限公司 | Intelligent automobile vision sensor testing method based on field testing technology |
| CN112230228B (en) * | 2020-09-30 | 2024-05-07 | 中汽院智能网联科技有限公司 | Intelligent automobile vision sensor testing method based on field testing technology |
| CN112373483A (en) * | 2020-11-23 | 2021-02-19 | 浙江天行健智能科技有限公司 | Vehicle speed and steering prediction method based on forward neural network |
| CN112373483B (en) * | 2020-11-23 | 2022-07-29 | 浙江天行健智能科技有限公司 | Vehicle speed and steering prediction method based on forward neural network |
| CN112912883A (en) * | 2021-02-07 | 2021-06-04 | 华为技术有限公司 | Simulation method and related equipment |
| CN112912883B (en) * | 2021-02-07 | 2022-06-28 | 华为技术有限公司 | Simulation method and related equipment |
| CN113219944A (en) * | 2021-04-27 | 2021-08-06 | 同济大学 | Intelligent vehicle control strategy test platform for mixed traffic flow working condition |
| CN113627740A (en) * | 2021-07-20 | 2021-11-09 | 东风汽车集团股份有限公司 | Driving load evaluation model construction system and construction method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109084992B (en) | 2020-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109084992A (en) | Method based on engine bench test unmanned vehicle intelligence | |
| US10346888B2 (en) | Systems and methods to obtain passenger feedback in response to autonomous vehicle driving events | |
| CN110562258B (en) | Method for vehicle automatic lane change decision, vehicle-mounted equipment and storage medium | |
| US20190129831A1 (en) | Autonomous Vehicle Simulation Testing Systems and Methods | |
| CN109466543A (en) | Plan autokinetic movement | |
| Dogan et al. | Autonomous driving: A comparison of machine learning techniques by means of the prediction of lane change behavior | |
| CN109421738A (en) | Method and apparatus for monitoring autonomous vehicle | |
| CN109242251A (en) | Vehicular behavior safety detecting method, device, equipment and storage medium | |
| CN107782564A (en) | A kind of automatic driving vehicle evaluation system and method | |
| CN109100155A (en) | A kind of unmanned vehicle is in ring high-speed simulation test macro and method | |
| CN109901574A (en) | Automatic Pilot method and device | |
| CN110531740A (en) | A kind of intelligent vehicle intelligence degree quantization assessment method | |
| CN107117179A (en) | The autonomous control noted based on occupant | |
| CN106548679A (en) | A kind of intelligent driving training system | |
| CN107117166A (en) | Autonomous dangerous item station | |
| CN106644503A (en) | Intelligent vehicle planning capacity testing platform | |
| CN107121979A (en) | Autonomous confidence control | |
| CN107117162A (en) | Autonomous vehicle control conversion | |
| US11657318B2 (en) | Assessing ride quality for autonomous vehicles | |
| CN108332977A (en) | A kind of classifying and analyzing method joining automotive test scene to intelligent network | |
| CN111275249A (en) | Driving behavior optimization method based on DQN neural network and high-precision positioning | |
| CN110371132A (en) | Driver's adapter tube appraisal procedure and device | |
| CN116134292A (en) | Tool for performance testing and/or training an autonomous vehicle planner | |
| CN109782751A (en) | Method and apparatus for autonomous system performance and benchmark test | |
| CN114787894A (en) | Perceptual error model |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |