CN113065685A - High-precision map traffic rule model based on automatic driving application scene and implementation method thereof - Google Patents
High-precision map traffic rule model based on automatic driving application scene and implementation method thereof Download PDFInfo
- Publication number
- CN113065685A CN113065685A CN202110225611.6A CN202110225611A CN113065685A CN 113065685 A CN113065685 A CN 113065685A CN 202110225611 A CN202110225611 A CN 202110225611A CN 113065685 A CN113065685 A CN 113065685A
- Authority
- CN
- China
- Prior art keywords
- road
- model
- vehicle
- traffic
- element layer
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000011218 segmentation Effects 0.000 claims abstract description 5
- 230000006399 behavior Effects 0.000 claims description 54
- 238000012549 training Methods 0.000 claims description 16
- 238000002474 experimental method Methods 0.000 claims description 9
- 238000012544 monitoring process Methods 0.000 claims description 7
- 238000005259 measurement Methods 0.000 claims description 6
- 230000008447 perception Effects 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 4
- 230000001105 regulatory effect Effects 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 238000003672 processing method Methods 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 230000003068 static effect Effects 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 230000003993 interaction Effects 0.000 abstract 1
- 238000007726 management method Methods 0.000 description 12
- 230000008859 change Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000009467 reduction Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- General Business, Economics & Management (AREA)
- Software Systems (AREA)
- Marketing (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Traffic Control Systems (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
The invention discloses a high-precision map traffic regulation model based on an automatic driving application scene and an implementation method thereof, wherein the high-precision map traffic regulation model based on the automatic driving application scene comprises a road model, a full-element traffic regulation model, a network model and an intelligent driving behavior model, the road model comprises a road grade element layer and a road type element layer, the full-element traffic regulation model comprises a road segmentation element layer, a road time interval element layer, a preceding lane element layer, a traffic signal element layer, a dangerous road section rule element layer and a basic traffic regulation element layer, the network model comprises a public network model and an ad hoc network management model, the intelligent driving behavior model is provided with different behavior control systems according to different vehicle types, and the vehicle judges own behavior mode and driving behavior according to different vehicle types and the high-precision map model. The invention realizes complete digitalization of interaction among vehicles, between vehicles and pedestrians and between vehicles and roads.
Description
The technical field is as follows:
the invention relates to an automatic driving high-precision map model, in particular to a high-precision map traffic rule model based on an automatic driving application scene and an implementation method thereof.
Background art:
with the increasing progress of the automatic driving technology, the requirement on high-precision maps is higher and higher. The method is only limited to the basic auxiliary function of the common digital map, and cannot meet the requirements of automatic driving on complex traffic environment driving and operation. The traffic model is added to the high-precision map based on automatic driving, which becomes an important step for establishing a digital map, guiding automatic driving behaviors by using the map, standardizing traffic safety by using the map and realizing real unmanned driving. The high-precision map integrated with the traffic rule is used for replacing the real-life road traffic rule needing manual intervention and identification, and is the necessary requirement of unmanned driving on urban traffic safety and traffic driving regulations.
The invention content is as follows:
the invention provides a high-precision map traffic rule model based on an automatic driving application scene and an implementation method thereof to solve the problems in the prior art.
The invention adopts the following technical scheme: a high-precision map traffic rule model based on an automatic driving application scene is characterized in that: the intelligent driving system comprises a road model, a full-factor traffic rule model, a network model and an intelligent driving behavior model;
the road model comprises a road grade element layer and a road type element layer, wherein the road grade element layer is divided into a first-level express way and a second-level express way; a primary main road, a secondary main road and a tertiary main road; a secondary trunk road; a primary branch and a secondary branch; the road type element layer is divided into structured roads: the method comprises the steps of (1) guiding an automatic driving vehicle according to point cloud measurement and traffic rules on a road with a fixed traffic rule and a single driving mode; unstructured road: the driving environment is complex, and vehicles and roads with various vehicle types guide automatic driving of the vehicles through perception, traffic rules and a road obstacle monitoring system;
the full-element traffic rule model comprises a road segmentation element layer, a road time interval element layer, a preceding lane element layer, a traffic signal element layer, a dangerous road section rule element layer and a basic traffic rule element layer;
the road segmentation element layer divides the road into geocodes according to the traffic models on the basis of the road types and the road levels, each section of different road models comprise different traffic rule models, and the behaviors and the driving rules of the automatic driving vehicle are expressed and guided through the geocodes;
the road time period element layer respectively uses an early peak time period collection method, a late peak time period collection method and collection of a middle common time period, the holiday trip peak time period collection method and the collection are divided into fixed time intervals to form two time period models, namely a working day traffic time period model and a holiday traffic time period model, the time period models are set in basic element data of a high-precision map, and corresponding time period models are set according to road types and road grades;
marking the vehicles in advance in a high-grade map in the element layer of the preceding lane, and searching intersections of the preceding lane by using high-precision sensing equipment and measuring equipment through a formulated rule when the automatic driving vehicles are set at the marked positions for driving so as to determine a passing standard and a passing instruction;
the high-precision map in the traffic signal element layer prompts the intersection position of an automatic driving vehicle by utilizing signals in advance, trains the time required for starting deceleration at the intersection required to stop and the safe distance required to be kept with a front vehicle before reaching the intersection, the safe distance is determined by automatic driving vehicle sensing equipment, the speed of the automatic driving vehicle and the deceleration time, and a mature traffic signal model is obtained by multiple times of training and sampling;
the dangerous road section rule element layer adopts a method of sampling for multiple times and training a sample model, the driving environment of the unmanned vehicle in severe weather is simulated, and the unmanned vehicle is informed to the intersection and the traffic road section with frequent accidents according to different schemes according to the simulation condition;
the basic traffic rule element layer is merged into a high-precision map based on the running of the automatic driving vehicle according to the classification and element specification modes of all models, and the automatic driving vehicle obtains traffic rule information in the traffic rule element layer according to the real-time position of the automatic driving vehicle to automatically run;
the network model comprises a public network model and an ad hoc network management model;
the intelligent driving behavior model is divided into a passenger vehicle model, a freight vehicle model, a traction vehicle model and a special purpose vehicle model, different behavior control systems are set according to different vehicle types, and the vehicle judges the behavior mode and the driving behavior according to different vehicle types and high-precision map models.
The invention also adopts the following technical scheme: a method for realizing a high-precision map traffic rule model based on an automatic driving application scene comprises the following steps:
the method comprises the following steps: extracting road elements according to the collected point cloud data and multiple collected vehicle measurements, and determining the road shape: at intersections, turning roads and straight lanes, carrying out feature analysis on lane lines by using a traditional computer vision method, carrying out feature classification by using a convolutional neural network model, judging lane line features, and carrying out training experiments on the categories of the lane lines;
step two: dynamic and static geocoding is carried out on different types of vehicles, pedestrians and different types of obstacles, multiple experimental tests are carried out, unstructured segmented roads are classified, a road obstacle monitoring system is formed, and corresponding instructions are completed according to different codes;
step three: the road gradient is used as an important image factor of a running mode and a vehicle behavior of a designated vehicle on a map, the congestion degree of a segmented slope road is measured through experiments, the congestion degree of the segmented slope road and an automatic driving perception model are measured based on the experiments, the vehicle behavior at the next moment is judged in a geographic coding mode, the running traffic of the slope road is regulated into a geographic code, and the automatic driving vehicle is guided to run and the automatic driving vehicle behavior is output;
step four: setting a geocode and a direction aiming at an intersection curve marked with a guide lane, and strictly driving according to the lane direction; the left turning lane takes the turning center point of the intersection as the origin of coordinates, and appoints and pre-judges the vehicle behavior and the vehicle track; the bend in passing signal classifies and marks the surrounding ground objects according to the vehicle sensing system, and one side passes through according to the marking priority; when a curve meets a stop signal, marking a stop line in a map by geocoding, and performing highest priority processing on the stop line, the stop signal in the map and other marks;
step five: the method is combined with an image processing method, a map is projected under a Frenet coordinate system by a map projection mechanism provided by GWS84 through the conversion of the coordinate system, intersection geocoding and a driving direction are marked under the coordinate system, an initial direction is defined on a sectional lane based on a high-precision map according to traffic rules, and whether the direction of a vehicle at the next moment is changed or not and a direction changing mode are determined according to the traffic rules of the high-precision map intersection and a lane changing traffic rule starting point as the basis of whether the direction is changed or not according to the codes at the intersection and the lane changing position.
The invention has the following beneficial effects: the invention aims to realize an automatic driving system of level L4 or above, a traffic rule high-precision map is used, the function of the high-precision map is enriched, the high-precision map can guide the behavior of an automatic driving vehicle, the path planning and the traffic control planning are realized, the traffic management mechanism in a real scene is thoroughly liberated, no traffic signals such as intersection traffic and the like exist on a real road, signal lamps, speed-limiting boards, cameras and other physical burdens are realized, the automatic driving vehicle is guided to drive and traffic control constraint is completely carried out by using the high-precision map, and the map realizes the prejudgment and the automatic driving behavior limitation such as speed reduction, turning stop and the like on a fixed road level and road type. All automatic driving behaviors and traffic rule systems are completely embodied in the map, so that full digitalization of communication among unmanned vehicles, vehicles and pedestrians and vehicles and roads is realized, and a foundation is laid for realizing a real digital city.
The high-precision map of the traffic rules can realize traffic management facilities such as no signals and the like, signal boards, cameras and the like, can greatly reduce resources consumed by rectifying, maintaining and setting the traffic management facilities, can also greatly reduce time cost, increase traffic efficiency and bring quality promotion and change for life production.
The specific implementation mode is as follows:
the invention relates to a high-precision map traffic regulation model based on an automatic driving application scene, which comprises a road model, a full-element traffic regulation model, a network model and an intelligent driving behavior model.
1. And (3) road model:
road grade element layer: the road grade distribution mode not only considers the problems of time consumption and resources, but also considers the requirements of different places, different policies and different local productivity levels and the relation of a supply network, and needs to reach balance through data and a training model, and the road grade division is determined by applying the road grade division to the supply network to reach a balance state for multiple times according to the training model.
The road grade is divided into a first-level express way and a second-level express way; a primary main road, a secondary main road and a tertiary main road; a secondary trunk road; a primary branch and a secondary branch.
Road type element layer: the classification according to the technical standard of roads can be divided into: the linear standard, the load standard and the clearance standard need to be combined with different environments and different traffic demands to carry out road survey design for classification. In the high-precision map, the classification and definition standards are combined, and in order to reflect the influence of different road types on urban traffic and guide automatic driving behaviors more accurately, the road types of the basic data city are divided into: structuring the road: and (4) guiding automatic driving of the vehicle according to the point cloud measurement and the traffic rule on the road with a fixed traffic rule and a single driving mode. Unstructured road: the driving environment is complex, and vehicles and roads with various vehicle types guide the automatic driving of the vehicles through perception, traffic rules and a road obstacle monitoring system.
Setting a map passing through an intersection or an electronic traffic sign intersection to communicate with an intelligent driving vehicle, namely, a critical distance for deceleration or turning, if the map is an electronic traffic signal or the like, namely, the minimum distance of the basic element layer parking area of the standard automatic driving vehicle and the high-precision map, when the electronic traffic signal timing is started, the vehicle with the distance larger than the minimum distance can safely run, the driving time and the braking time of the automatic driving vehicle at the intersection passing through the traffic sign are estimated according to the traffic rules and the automatic driving vehicle behavior model, model training is carried out for multiple times according to intersections of traffic signs in different map models, and according to historical records, different traffic signs including signs such as red and green, speed reduction signs and the like and intersections corresponding to different road types and different road grades in the high-precision map model are confirmed, and time models of different driving sections such as stopping, speed reduction, constant speed and the like of the automatic driving vehicle are formulated.
2. The full-element traffic rule model is as follows:
road segment element layer: on the basis of road types and road levels, the roads are divided into geographical codes according to traffic models, each section of different road models comprises different traffic rule models, and the behaviors and the driving rules of the automatic driving vehicles are expressed and guided through the geographical codes. The geocoding is generated in the full-element traffic rule model, different and unique codes in each road model are generated on the basis of the road model, and the geocoding is established through a road model feature library and the full-element traffic rule, contains topological relation and intersection complexity information, and guides the behavior and the driving mode of an automatic driving vehicle.
Road time period element layer: the training and customizing method of the road time period element layer respectively uses an early peak time period collection method, a late peak time period collection method and a middle common time period collection method; the method comprises the steps of collecting travel peak periods in holidays, dividing the collection into fixed time intervals, forming two period models which are respectively a working day traffic period model and a holiday traffic period model, setting the period models into basic element data of a high-precision map, setting the corresponding period models according to road types and road grades, setting the period models to be key points for guiding the high-precision map of automatic driving vehicle behaviors to replace traffic management facilities in real life, enabling the high-precision map to be communication between vehicles and vehicles, enabling the vehicles and the roads to correspond, and enabling the communication between the vehicles and people to be completely embodied on the high-precision map, and achieving digitization of the high-precision map. The time period model guides and standardizes the behavior of the automatic driving vehicle on the basis of high-precision map basic data element layers such as road grades and road types, the unmanned driving vehicle runs in the range specified by the time period model, the running data and the running state of the automatic driving vehicle are obtained, and the automatic driving vehicle runs more safely and efficiently under the control of a traffic management range.
Element layer of the leading lane: when the lane is changed and the vehicle passes through the crossroad, the high-precision map electronic mark is used for leading the automatic driving vehicle of the leading lane, the vehicle which is led ahead by the real traffic management rule is marked in the high-grade map, and when the automatic driving vehicle is set at the mark for driving, the high-precision sensing equipment and the measuring equipment are used for searching the crossroad of the leading lane by establishing the rule so as to determine the traffic standard and the traffic instruction.
Traffic signal element layer: according to traffic safety management rules and methods, vehicles and pedestrians are permitted to pass when the green light is on, vehicles which cross a stop line when the yellow light is on can continue to pass, vehicles and pedestrians are prohibited to pass when the red light is on, the whole set of traffic rules are beyond high-precision map data layer guidance and specification of automatic driving vehicles, and in order to enable unmanned vehicles to run more safely, real traffic safety rules are insufficient. The traffic signal model not only covers a vehicle running basis specified by traffic management, the traffic signal model meets the passing time and the stopping time of a traffic signal, but also meets the safety distance required to be kept when the vehicle runs and stops, the safety distance is subjected to a plurality of times of experimental training through different vehicle speed sections, the high-precision map prompts the crossing position of the automatic driving vehicle by utilizing a signal in advance, the time required for starting to decelerate at the crossing required to be stopped and the safety distance required to be kept by a front vehicle before reaching the crossing, the safety distance is determined by the automatic driving vehicle sensing equipment, the speed of the automatic driving vehicle and the deceleration time, and a mature traffic signal model is obtained through a plurality of times of training and sampling.
Dangerous road section rule element layer: according to the traffic control rule, the speed of the vehicle should be reduced and the vehicle should be slowly driven when the vehicle is driven at night or at intersections where danger is likely to occur, or when the vehicle is driven under the weather conditions of sand dust, hail, rain, snow, fog, ice and the like. However, the rule model of the dangerous road section not only needs the unmanned vehicle to travel slowly at a low speed when meeting the driving environments, but also needs to consider the situation that the sensing system of the unmanned vehicle sensing equipment is blocked when meeting severe weather or having an unclear visual field according to different roads, and when the sensing environment is severe, the unmanned vehicle depends on the specification and guidance of a high-precision map. The dangerous road end rule model adopts a method of sampling for multiple times and training a sample model, severe weather is simulated to be the driving environment of the unmanned vehicle, the unmanned vehicle is informed to intersections and traffic sections with frequent accidents according to different schemes according to simulation conditions, and different speed limits and stopping ranges are set according to different weather conditions and driving time conditions in each scheme, so that the high-precision map can comprehensively guide the automatic driving vehicle to run, and safety accidents are greatly reduced.
And the basic traffic regulation element layer comprises but is not limited to allowing straight running, forbidding straight running, allowing left turning, forbidding left turning, allowing right turning, forbidding right turning, allowing parking, forbidding parking, limiting speed, the highest limiting speed, the lowest limiting speed, allowing overtaking, forbidding overtaking, allowing turning around, forbidding turning around, temporary parking spots, parking lots and forbidding whistling. The basic traffic element data layer is merged into a high-precision map based on the running of the automatic driving vehicle according to the classification and element specification modes of all models, the automatic driving vehicle obtains the traffic rule information in the traffic rule element layer according to the real-time position of the automatic driving vehicle and automatically runs, and the automatic driving vehicle does not only use a sensing system carried by the automatic driving vehicle to identify physical traffic sign lines to obtain the traffic rule information, and is an important element for realizing the standardized safety and improving the efficiency of the automatic driving vehicle.
3. And (3) network model:
the intelligent driving behavior is completed within the specified signal coverage range through the signal system, the signal strength and the signal coverage range, and a stopping method or a braking method outside the signal coverage range is defined.
Public network model: and establishing a network model base according to different road types and different road grades, and regulating the driving range of the automatic driving vehicle to ensure effective signal communication between vehicles, vehicles and people, vehicles and roads and vehicles and networks in the network signal coverage range.
Ad hoc network management model: according to the difference of road grade and road type division, an emergency network management model is established, and when the signal intensity is weakened or the network signal is interfered or the intensity is reduced, an alternative model is started, so that the automatic driving vehicle can effectively receive and transmit the network signal, and the traffic safety is ensured.
After the automatic driving vehicle inputs a target place, the network model plans in a network coverage range according to a path and an automatic driving behavior and downloads basic map data and required business layer data before the automatic driving behavior occurs; in the running process of the vehicle, if the situation that the training model predicted or sensed by the vehicle sensing system and the vehicle behavior system is not covered is met, the network model and the intelligent driving behavior model carry out network signal emergency scheme communication, and when the vehicle driven by the automatic driving vehicle runs to an area outside the coverage of the network model, an emergency braking system scheme is adopted.
4. Intelligent driving is a model:
and setting a geocode according to the road section characteristics for the driving behavior data of the automatic driving vehicle and reporting the geocode to the cloud. And calculating according to the road model and the traffic model, combining with the traffic rules corresponding to the geocoding to obtain an optimal automatic driving behavior strategy, and outputting the optimal automatic driving behavior strategy to the intelligent driving system to control the vehicle behavior.
The automatic driving behavior models are classified according to vehicle types and are divided into models of passenger cars, trucks, traction cars and special-purpose cars, different behavior control systems are set according to different vehicle types, and the vehicles judge behavior modes and driving behaviors according to different vehicle types and high-precision map models.
The intelligent driving behavior model needs to be matched with a high-precision map basic data layer and a service data element layer, a driving scheme of the whole automatic driving vehicle is standardized according to a network model and a high-precision map rule instruction, and when the automatic driving vehicle drives in an area or a scene outside the coverage range of a training model, an emergency or preparation scheme is started by being matched with a high-precision element layer and the network model.
The invention relates to a method for realizing a high-precision map traffic rule model based on an automatic driving application scene, which comprises the following steps:
extracting road elements according to the collected point cloud data and multiple collected vehicle measurements, and determining the road shape: intersections, turning roads and straight lanes. The method comprises the steps of utilizing a traditional computer vision method to carry out feature analysis on a lane line, carrying out feature classification by using a convolutional neural network model, judging lane line features, carrying out training experiments on the categories of the lane line features, and carrying out multiple times of judging experiments so as to avoid single misjudgment under the condition that the lane line is incomplete or the lane line is shielded.
The road obstacle monitoring system includes: different types of vehicles, pedestrians and different types of obstacles are subjected to dynamic and static geocoding, multiple experimental tests are carried out, unstructured segmented roads are classified, a road obstacle monitoring system is formed, and corresponding instructions are completed according to different codes.
The road gradient is used as an important image factor of a map for specifying a vehicle running mode and vehicle behavior, and is also used as a factor for predicting the vehicle running mode at the next moment in an image in a traffic rule model and is divided into a straight ramp, a curved ramp and a straight-curved mixed ramp. According to different traffic rules on the slope with different crowdedness degrees in the implementation regulations of road traffic safety laws, the crowdedness degree of the segmented slope is measured through experiments, the next-moment vehicle behavior is judged in a geographic coding mode on the basis of the experimentally measured crowdedness degree of the segmented slope and an automatic driving perception model, and the slope driving traffic is regulated into the geographic coding to guide the automatic driving vehicle to drive and output the automatic driving vehicle behavior.
The curve curvature, the curve curvature of the curve at different element levels and the traffic rules determine the vehicle behavior of the autonomous vehicle at the curve. The method comprises the following steps: an intersection bend marked with a guide lane; a left turn lane; a right turn lane; a bend is marked when a release signal is met; and a curve is encountered by the stop signal. Setting a geocode and a direction at an intersection curve marked with a guide lane, and strictly driving according to the lane direction; the left turning lane takes the turning center point of the intersection as the origin of coordinates, and appoints and pre-judges the vehicle behavior and the vehicle track; the bend in passing signal classifies and marks the surrounding ground objects according to the vehicle sensing system, and one side passes through according to the marking priority; when a curve meets a stop signal, a stop line is marked in a map by geocoding, and the highest priority is processed on the stop line, the stop signal in the map and other marks.
And the driving direction is combined with an image processing method, a map is projected under a Frenet coordinate system by using a map projection mechanism provided by GWS84 through the conversion of the coordinate system, and intersection geocoding and the driving direction are marked under the coordinate system. On the sectional lane, an initial direction is defined based on a high-precision map according to traffic rules, the starting point of a high-precision map intersection traffic rule and a lane change traffic rule is used as a basis for judging whether to change the direction or not, and whether to change the direction of the vehicle at the next moment or not and a direction change mode are determined by combining codes at the intersection and the lane change position.
And the lane type library is divided according to the road type and the road grade, and is provided with a lane type element layer, wherein the lane type model needs to consider specific traffic management rules and comprises a truck automatic driving behavior model (empty vehicle and heavy vehicle), a passenger car automatic driving behavior model, a car automatic driving behavior model and a special vehicle automatic driving behavior model. Different traffic driving strategies are established according to different vehicle types.
The foregoing is only a preferred embodiment of this invention and it should be noted that modifications can be made by those skilled in the art without departing from the principle of the invention and these modifications should also be considered as the protection scope of the invention.
Claims (2)
1. A high-precision map traffic rule model based on an automatic driving application scene is characterized in that: the intelligent driving system comprises a road model, a full-factor traffic rule model, a network model and an intelligent driving behavior model;
the road model comprises a road grade element layer and a road type element layer, wherein the road grade element layer is divided into a first-level express way and a second-level express way; a primary main road, a secondary main road and a tertiary main road; a secondary trunk road; a primary branch and a secondary branch; the road type element layer is divided into structured roads: the method comprises the steps of (1) guiding an automatic driving vehicle according to point cloud measurement and traffic rules on a road with a fixed traffic rule and a single driving mode; unstructured road: the driving environment is complex, and vehicles and roads with various vehicle types guide automatic driving of the vehicles through perception, traffic rules and a road obstacle monitoring system;
the full-element traffic rule model comprises a road segmentation element layer, a road time interval element layer, a preceding lane element layer, a traffic signal element layer, a dangerous road section rule element layer and a basic traffic rule element layer;
the road segmentation element layer divides the road into geocodes according to the traffic models on the basis of the road types and the road levels, each section of different road models comprise different traffic rule models, and the behaviors and the driving rules of the automatic driving vehicle are expressed and guided through the geocodes;
the road time period element layer respectively uses an early peak time period collection method, a late peak time period collection method and collection of a middle common time period, the holiday trip peak time period collection method and the collection are divided into fixed time intervals to form two time period models, namely a working day traffic time period model and a holiday traffic time period model, the time period models are set in basic element data of a high-precision map, and corresponding time period models are set according to road types and road grades;
marking the vehicles in advance in a high-grade map in the element layer of the preceding lane, and searching intersections of the preceding lane by using high-precision sensing equipment and measuring equipment through a formulated rule when the automatic driving vehicles are set at the marked positions for driving so as to determine a passing standard and a passing instruction;
the high-precision map in the traffic signal element layer prompts the intersection position of an automatic driving vehicle by utilizing signals in advance, trains the time required for starting deceleration at the intersection required to stop and the safe distance required to be kept with a front vehicle before reaching the intersection, the safe distance is determined by automatic driving vehicle sensing equipment, the speed of the automatic driving vehicle and the deceleration time, and a mature traffic signal model is obtained by multiple times of training and sampling;
the dangerous road section rule element layer adopts a method of sampling for multiple times and training a sample model, the driving environment of the unmanned vehicle in severe weather is simulated, and the unmanned vehicle is informed to the intersection and the traffic road section with frequent accidents according to different schemes according to the simulation condition;
the basic traffic rule element layer is merged into a high-precision map based on the running of the automatic driving vehicle according to the classification and element specification modes of all models, and the automatic driving vehicle obtains traffic rule information in the traffic rule element layer according to the real-time position of the automatic driving vehicle to automatically run;
the network model comprises a public network model and an ad hoc network management model;
the intelligent driving behavior model is divided into a passenger vehicle model, a freight vehicle model, a traction vehicle model and a special purpose vehicle model, different behavior control systems are set according to different vehicle types, and the vehicle judges the behavior mode and the driving behavior according to different vehicle types and high-precision map models.
2. A realization method of a high-precision map traffic rule model based on an automatic driving application scene is characterized in that: the method comprises the following steps:
the method comprises the following steps: extracting road elements according to the collected point cloud data and multiple collected vehicle measurements, and determining the road shape: at intersections, turning roads and straight lanes, carrying out feature analysis on lane lines by using a traditional computer vision method, carrying out feature classification by using a convolutional neural network model, judging lane line features, and carrying out training experiments on the categories of the lane lines;
step two: dynamic and static geocoding is carried out on different types of vehicles, pedestrians and different types of obstacles, multiple experimental tests are carried out, unstructured segmented roads are classified, a road obstacle monitoring system is formed, and corresponding instructions are completed according to different codes;
step three: the road gradient is used as an important image factor of a running mode and a vehicle behavior of a designated vehicle on a map, the congestion degree of a segmented slope road is measured through experiments, the congestion degree of the segmented slope road and an automatic driving perception model are measured based on the experiments, the vehicle behavior at the next moment is judged in a geographic coding mode, the running traffic of the slope road is regulated into a geographic code, and the automatic driving vehicle is guided to run and the automatic driving vehicle behavior is output;
step four: setting a geocode and a direction aiming at an intersection curve marked with a guide lane, and strictly driving according to the lane direction; the left turning lane takes the turning center point of the intersection as the origin of coordinates, and appoints and pre-judges the vehicle behavior and the vehicle track; the bend in passing signal classifies and marks the surrounding ground objects according to the vehicle sensing system, and one side passes through according to the marking priority; when a curve meets a stop signal, marking a stop line in a map by geocoding, and performing highest priority processing on the stop line, the stop signal in the map and other marks;
step five: the method is combined with an image processing method, a map is projected under a Frenet coordinate system by a map projection mechanism provided by GWS84 through the conversion of the coordinate system, intersection geocoding and a driving direction are marked under the coordinate system, an initial direction is defined on a sectional lane based on a high-precision map according to traffic rules, and whether the direction of a vehicle at the next moment is changed or not and a direction changing mode are determined according to the traffic rules of the high-precision map intersection and a lane changing traffic rule starting point as the basis of whether the direction is changed or not according to the codes at the intersection and the lane changing position.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110225611.6A CN113065685B (en) | 2021-03-01 | 2021-03-01 | High-precision map traffic rule model based on automatic driving application scene and implementation method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110225611.6A CN113065685B (en) | 2021-03-01 | 2021-03-01 | High-precision map traffic rule model based on automatic driving application scene and implementation method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113065685A true CN113065685A (en) | 2021-07-02 |
| CN113065685B CN113065685B (en) | 2023-10-17 |
Family
ID=76559399
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110225611.6A Active CN113065685B (en) | 2021-03-01 | 2021-03-01 | High-precision map traffic rule model based on automatic driving application scene and implementation method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113065685B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113408654A (en) * | 2021-07-13 | 2021-09-17 | 航天海鹰机电技术研究院有限公司 | Urban road network fusion method, system and storage medium based on hierarchical merging |
| CN114648877A (en) * | 2022-05-13 | 2022-06-21 | 北京理工大学前沿技术研究院 | Method, system, equipment and storage medium for vehicle passing at signalless intersection |
| CN115440028A (en) * | 2022-07-22 | 2022-12-06 | 中智行(苏州)科技有限公司 | Traffic road scene classification method based on labeling |
| CN115713866A (en) * | 2022-10-11 | 2023-02-24 | 悉地（苏州）勘察设计顾问有限公司 | Road static information active service method based on vehicle running characteristics |
| CN116343469A (en) * | 2022-12-23 | 2023-06-27 | 昆易电子科技(上海)有限公司 | Data processing method and data processing device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609633A (en) * | 2017-05-03 | 2018-01-19 | 同济大学 | The position prediction model construction method of vehicle traveling influence factor based on deep learning in car networking complex network |
| CN108549366A (en) * | 2018-05-04 | 2018-09-18 | 同济大学 | Intelligent automobile road driving mapping experiment method parallel with virtual test |
| CN112101120A (en) * | 2020-08-18 | 2020-12-18 | 沃行科技(南京)有限公司 | Map model based on automatic driving application scene and application method thereof |
-
2021
- 2021-03-01 CN CN202110225611.6A patent/CN113065685B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609633A (en) * | 2017-05-03 | 2018-01-19 | 同济大学 | The position prediction model construction method of vehicle traveling influence factor based on deep learning in car networking complex network |
| CN108549366A (en) * | 2018-05-04 | 2018-09-18 | 同济大学 | Intelligent automobile road driving mapping experiment method parallel with virtual test |
| CN112101120A (en) * | 2020-08-18 | 2020-12-18 | 沃行科技(南京)有限公司 | Map model based on automatic driving application scene and application method thereof |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113408654A (en) * | 2021-07-13 | 2021-09-17 | 航天海鹰机电技术研究院有限公司 | Urban road network fusion method, system and storage medium based on hierarchical merging |
| CN113408654B (en) * | 2021-07-13 | 2024-02-20 | 航天海鹰机电技术研究院有限公司 | Urban road network fusion method, system and storage medium based on hierarchical merging |
| CN114648877A (en) * | 2022-05-13 | 2022-06-21 | 北京理工大学前沿技术研究院 | Method, system, equipment and storage medium for vehicle passing at signalless intersection |
| CN115440028A (en) * | 2022-07-22 | 2022-12-06 | 中智行(苏州)科技有限公司 | Traffic road scene classification method based on labeling |
| CN115440028B (en) * | 2022-07-22 | 2024-01-30 | 中智行(苏州)科技有限公司 | Traffic road scene classification method based on labeling |
| CN115713866A (en) * | 2022-10-11 | 2023-02-24 | 悉地（苏州）勘察设计顾问有限公司 | Road static information active service method based on vehicle running characteristics |
| CN115713866B (en) * | 2022-10-11 | 2023-08-22 | 悉地（苏州）勘察设计顾问有限公司 | Road static information active service method based on vehicle operation characteristics |
| CN116343469A (en) * | 2022-12-23 | 2023-06-27 | 昆易电子科技(上海)有限公司 | Data processing method and data processing device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113065685B (en) | 2023-10-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113065685B (en) | High-precision map traffic rule model based on automatic driving application scene and implementation method thereof | |
| CN108364494B (en) | Intelligent road traffic management method, system and platform | |
| CN110853357B (en) | Early warning control system and method for dangerous highway sections and construction areas | |
| CN107843440B (en) | System and method for testing performance of automatic driving vehicle | |
| Tengilimoglu et al. | Implications of automated vehicles for physical road environment: A comprehensive review | |
| CN109641589B (en) | Route planning for autonomous vehicles | |
| CN111367292A (en) | Intelligent road system for automatically driving automobile | |
| JP6785939B2 (en) | Systems and methods for generating surface map information in an emergency | |
| CN110907966A (en) | Emergency vehicle navigation system and method based on real-time traffic flow in time of Internet of things | |
| CN107924617A (en) | System and method for determining navigation information for autonomous vehicle | |
| CN107490384B (en) | Optimal static path selection method based on urban road network | |
| WO2021147748A1 (en) | Self-driving method and related device | |
| RU2766038C1 (en) | Method and apparatus for generating information on the traffic environment for a vehicle, and method for controlling movement for a vehicle | |
| CN114385661A (en) | High-precision map updating system based on V2X technology | |
| US20200320875A1 (en) | Coordinated Traffic Feature and Map Generation for Convoy Operations | |
| CN114139329A (en) | Virtual test scene construction method and device | |
| US20230140569A1 (en) | System and method for maps for an autonomous vehicle | |
| WO2022213542A1 (en) | Method and system for clearing information-controlled intersection on basis of lidar and trajectory prediction | |
| CN112906209B (en) | Closed-field-oriented vehicle-road cooperative test scene credibility assessment method | |
| CN116842688A (en) | Online compliance verification system oriented to automatic driving decision algorithm | |
| RU2725569C1 (en) | Method for controlling movement of unmanned vehicles (uv) in column and/or separate uv and monitoring by intelligent transport infrastructure (iti) of automobile network | |
| EP4080164A1 (en) | Identifying parkable areas for autonomous vehicles | |
| CN114004080A (en) | Method for evaluating road driving adaptability of automatic driving vehicle based on virtual simulation | |
| RU2777857C1 (en) | Device for formation of modified energy-efficient track of operated vehicle | |
| RU2777854C1 (en) | Device for formation of energy-efficient navigation route of operated vehicle |
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 |