CN117416342B - Intelligent parking method for unmanned vehicle - Google Patents

Intelligent parking method for unmanned vehicle Download PDF

Info

Publication number
CN117416342B
CN117416342B CN202311737049.0A CN202311737049A CN117416342B CN 117416342 B CN117416342 B CN 117416342B CN 202311737049 A CN202311737049 A CN 202311737049A CN 117416342 B CN117416342 B CN 117416342B
Authority
CN
China
Prior art keywords
node
parking
path
nodes
position node
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.)
Active
Application number
CN202311737049.0A
Other languages
Chinese (zh)
Other versions
CN117416342A (en
Inventor
杨扬
胡心怡
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Boonray Intelligent Technology Co Ltd
Original Assignee
Shanghai Boonray Intelligent Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Boonray Intelligent Technology Co Ltd filed Critical Shanghai Boonray Intelligent Technology Co Ltd
Priority to CN202311737049.0A priority Critical patent/CN117416342B/en
Publication of CN117416342A publication Critical patent/CN117416342A/en
Application granted granted Critical
Publication of CN117416342B publication Critical patent/CN117416342B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units, or advanced driver assistance systems for ensuring comfort, stability and safety or drive control systems for propelling or retarding the vehicle
    • B60W30/06—Automatic manoeuvring for parking
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • B60W60/0015—Planning or execution of driving tasks specially adapted for safety
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • B60W60/0025—Planning or execution of driving tasks specially adapted for specific operations
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2552/00—Input parameters relating to infrastructure
    • B60W2552/50—Barriers

Abstract

The invention relates to the technical field of automatic parking, and provides an intelligent parking method of an unmanned vehicle, which comprises the following steps: acquiring collision-free paths from the unmanned vehicle to parking points of all idle parking spaces; acquiring a position node on a collision-free path, and acquiring a node direction mutation rate; obtaining the neighborhood direction change difference of the position node, and further obtaining the direction continuous mutation consistency of the position node; setting a direction control node, and acquiring self-adaptive weights of the position nodes on the collision-free path according to the continuous abrupt change consistency of the direction of the position nodes and the distance between the position nodes and the direction control node; and acquiring a smooth path from the unmanned vehicle to a parking point of the idle parking space, acquiring the path goodness of the smooth path according to the smooth path, and realizing intelligent parking of the unmanned vehicle according to the path goodness of the smooth path. The invention aims to solve the problem of insufficient adaptability of the optimal path and the parking space to the intelligent parking scene in the intelligent parking process.

Description

Intelligent parking method for unmanned vehicle
Technical Field
The invention relates to the technical field of automatic parking, in particular to an intelligent parking method of an unmanned vehicle.
Background
With the development of sensor technology, artificial intelligence and other technologies, unmanned vehicles show great potential and application prospect in the traffic field. Compared with the traditional driving mode, the unmanned vehicle can realize intelligent driving decision and autonomously control the automobile, so that the unmanned vehicle can safely run in various complex traffic environments. The intelligent parking of the unmanned vehicle is a system for realizing automatic parking and taking of the vehicle by utilizing an intelligent technology, and the vehicle can autonomously sense, plan a path and finish a parking process by using technologies such as a sensor, computer vision and the like, so that more convenient and efficient parking experience is provided for an automobile driver.
Currently, an intelligent parking method of an unmanned vehicle mainly relies on a path planning algorithm to automatically generate an optimal path, and then the path is smoothed, so that the position and the path length of an obstacle are considered. However, in the intelligent parking scene, the influence of curvature change of the path and the direction of the vehicle reaching the parking point on the parking efficiency is not considered in path planning, so that the obtained optimal path and parking space have low adaptability to the intelligent parking scene, and the intelligent parking efficiency of the unmanned vehicle is low.
Disclosure of Invention
The invention provides an intelligent parking method of an unmanned vehicle, which aims to solve the problem of insufficient adaptability of an optimal path and a parking space to an intelligent parking scene in the intelligent parking process, and adopts the following technical scheme:
one embodiment of the invention provides an intelligent parking method of an unmanned vehicle, which comprises the following steps:
collecting parking points of idle parking spaces of the unmanned vehicle, and obtaining collision-free paths from the unmanned vehicle to the parking points of all the idle parking spaces;
acquiring position nodes on a collision-free path, acquiring a position node sequence, acquiring direction angles of the position nodes according to the position node sequence, acquiring a direction angle sequence, acquiring the distribution probability of the direction angles according to the direction angle sequence, and acquiring the node direction mutation rate according to the distribution probability of the direction angles and the adjacent position nodes of the direction angles in the position node sequence;
acquiring the angle difference of the position node according to the adjacent position node of the position node, acquiring the neighborhood node of the position node in the position node sequence, acquiring the neighborhood direction change difference of the position node according to the angle difference of the neighborhood node of the position node, and acquiring the continuous mutation consistency of the direction of the position node according to the neighborhood direction change difference of the position node and the node direction mutation rate;
acquiring position nodes of parking point positions on a collision-free path, setting direction control nodes according to the position nodes of the parking point positions, setting weights of the position nodes of the parking point positions and the direction control nodes, and acquiring self-adaptive weights of the position nodes on the collision-free path according to the continuous abrupt change consistency of the directions of the position nodes and the distances between the position nodes and the direction control nodes;
according to the weights and positions of all position nodes on all collision-free paths of the unmanned vehicle, obtaining a smooth path from the unmanned vehicle to a parking point of an idle parking space, obtaining the degree of confusion and the length of curvature change of the smooth path according to the smooth path, obtaining the path goodness of the smooth path according to the degree of confusion and the length of curvature change of the smooth path, and realizing intelligent parking of the unmanned vehicle according to the path goodness of the smooth path.
Further, the method for obtaining the position node sequence comprises the following specific steps:
and arranging the position nodes on the collision-free path according to the direction from the unmanned vehicle to the parking point of the idle parking space, and acquiring a position node sequence.
Further, the method for obtaining the direction angle of the position node according to the position node sequence includes the following specific steps:
each position node in the position node sequence is respectively used as a position node to be analyzed;
the last position node in the position node sequence where the position node to be analyzed is positioned is marked as a first adjacent position node of the position node to be analyzed;
the clockwise included angle between the connecting line between the position node to be analyzed and the first adjacent position node of the position node to be analyzed and the initial direction is recorded as the direction angle of the position node to be analyzed;
and arranging the direction angles of all the position nodes in order from small to large to obtain a direction angle sequence.
Further, the method for obtaining the distribution probability of the direction angles according to the direction angle sequence comprises the following specific steps:
dividing the direction angle sequence average into a first division threshold value number of direction angle sequence intervals;
the ratio of the total number of the repeated occurrence of direction angles contained in the direction angle sequence section to the number of the position nodes is recorded as the distribution probability of all direction angles contained in the direction angle sequence section.
Further, the specific method for obtaining the angle difference of the position node according to the adjacent position node of the position node includes:
and recording the difference between the direction angle of the position node to be analyzed and the direction angle of the first adjacent position node of the position node to be analyzed as the angle difference of the position node to be analyzed.
Further, the method for obtaining the neighborhood node of the position node in the position node sequence, and obtaining the neighborhood direction change difference of the position node according to the angle difference of the neighborhood node of the position node, includes the following specific steps:
marking a second division threshold position node adjacent to the position node to be analyzed in the position node sequence as a neighborhood node of the position node to be analyzed;
and carrying out anomaly detection on the angle differences of all the position nodes on the same collision-free path, obtaining outlier factors of each position node, and recording the average value of the outlier factors of the angle differences of the neighborhood nodes of the position node to be analyzed as the neighborhood direction change difference of the position node to be analyzed.
Further, the setting of the direction control node according to the position node of the parking point position includes the specific steps:
setting a direction control node in the parking space direction of a position node corresponding to the parking point of the idle parking space, wherein the distance between the position node corresponding to the parking point of the idle parking space and the direction control node is the average value of the distances between all adjacent position nodes on a collision-free path where the position node corresponding to the parking point of the idle parking space is located.
Further, the method for setting the weights of the position node and the direction control node of the parking point position comprises the following specific steps:
and setting the weights of the position nodes and the direction control nodes corresponding to the parking points of the idle parking spaces as a second preset threshold value.
Further, the specific method for obtaining the degree and length of confusion of curvature change of the smooth path according to the smooth path includes:
uniformly setting a third division threshold value on the smooth path, acquiring the curvature radius of each sample point, and taking the sum of the information entropy of the curvature radius of all sample points as the curvature change chaotic degree of the smooth path;
the length of the smooth path is obtained using an arc length integral formula.
Further, the intelligent parking of the unmanned vehicle is realized according to the path goodness of the smooth path, and the method comprises the following specific steps:
and taking the path with the maximum path goodness as the optimal path, taking the idle parking space corresponding to the optimal path as the optimal parking space, controlling the vehicle to reach the parking point of the optimal parking space along the optimal path by using the vehicle control system, and parking, thereby realizing intelligent parking of the unmanned vehicle.
The beneficial effects of the invention are as follows:
according to the method, when the collision-free path is subjected to smooth processing, the influence of weight setting on the obstacle avoidance effect is analyzed, larger self-adaptive weight is given to position nodes near the obstacle, smaller self-adaptive weight is given to position nodes close to the parking point, so that the smooth path from the unmanned vehicle to the parking point of the idle parking space is smoother, the influence of the obstacle on the possible damage of the vehicle is reduced, specifically, the continuous direction mutation consistency is obtained according to the change of the direction angle between the position nodes and the adjacent position nodes, the local direction mutation degree of the position nodes can be reflected by the continuous direction mutation consistency, and the smaller the continuous direction mutation consistency is, the greater the possibility that the position nodes are positioned around the obstacle is; then, setting a direction control node by analyzing the influence of the vehicle orientation on the parking efficiency when the vehicle reaches the parking point, setting the weight of each position node according to the distance between the position node and the direction control node and the local direction mutation degree, and obtaining a smooth path by the influence of the weight on the shape of the Bezier curve, thereby improving the adaptability of path planning to intelligent parking scenes; finally, the optimal path and the corresponding parking spaces are selected according to the curvature and the length of the smooth path from the vehicle to all the idle parking spaces, so that the intelligent parking efficiency of the unmanned vehicle is improved, and the problem of insufficient adaptability of the optimal path and the parking spaces to intelligent parking scenes in the intelligent parking process is solved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
Fig. 1 is a schematic flow chart of an intelligent parking method of an unmanned vehicle according to an embodiment of the present invention;
FIG. 2 is a schematic illustration of a parking spot;
fig. 3 is a schematic diagram of a directional control node arrangement.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, a flowchart of an intelligent parking method for an unmanned vehicle according to an embodiment of the invention is shown, the method includes the following steps:
and S001, collecting parking points of idle parking spaces of the unmanned vehicle, and obtaining collision-free paths from the unmanned vehicle to the parking points of all the idle parking spaces.
When the unmanned vehicle reaches the target area, a navigation plane map is obtained through a GPS navigation positioning technology, and the position of the unmanned vehicle is marked in the navigation plane map through a GPS real-time positioning technology. And detecting metal objects through the magnetic induction sensors on each parking space, acquiring an idle parking space, and marking parking points of the idle parking space in the navigation plane map. The parking point is the lateral distance from the rear of the parking space to the central point of the parking space, and the lateral distance isLongitudinal distance is->Is (are) located>Is 5 @, is given by>Is 2; will be->The parking point of each free parking space is marked as +.>. The orientation of the unmanned vehicle is obtained by the vehicle sensor.
The schematic diagram of the parking spot is shown in fig. 2.
In the intelligent parking process, when a traditional path planning algorithm plans the path of an unmanned vehicle reaching a parking spot of a parking space, the main consideration is the path length after avoiding obstacles, and the influence of the path smoothness degree, the vehicle orientation and the parking direction on the vehicle stability and the parking efficiency is ignored. Because the parking space is limited, after the vehicle reaches the parking point, the gesture of the vehicle has a great influence on parking efficiency, and when the included angle between the orientation of the vehicle and the parking orientation of the parking space is overlarge, the vehicle can be stably parked in the parking space through multiple adjustments.
Based on the analysis, generating a collision-free path from the unmanned vehicle to the parking points of all the idle parking spaces in the destination area by using an A-algorithm, and leading the unmanned vehicle to the first positionParking spot of individual free parking spaces>Is marked as +.>. The collision-free path is obtained by using an a-algorithm, which is a known technique and will not be described in detail.
So far, the collision-free path from the unmanned vehicle to the parking points of all the idle parking spaces is obtained.
Step S002, acquiring position nodes on a collision-free path, acquiring a position node sequence, acquiring a direction angle sequence according to the position node sequence, acquiring a distribution probability of the direction angles according to the direction angle sequence, and acquiring a node direction mutation rate according to the distribution probability of the direction angles and adjacent position nodes of the direction angles in the position node sequence.
To ensure the stability of the vehicle running, a Bezier curve is adopted for the collision-free pathSmoothing is performed. The Bezier curve is a parameterized curve, the shape of the Bezier curve is determined by the position and weight of nodes on the path, and the greater the weight of the nodes, the greater the influence on the Bezier curve, and the closer the curve is to the nodes. Therefore, the weights of the nodes on the path of the bezier curve need to be determined.
In order to uniformly express angles in the calculation process of vehicles, parking spaces and paths, taking the north direction as the initial direction, acquiring position nodes on the collision-free paths, and counting the position nodes on the collision-free pathsAnd arranging the position nodes on the collision-free path according to the direction from the unmanned vehicle to the parking point of the idle parking space, and acquiring a position node sequence.
To the first unmanned vehicleParking spot of individual free parking spaces>Is free of collisionBump path->The number of position nodes on the table is marked as +.>The (th) in the position node sequence>The individual location nodes are marked->First->The previous position node of the position node is marked +.>First->The next position node of the position nodes is marked +.>Wherein->。
The effect of the degree of path smoothness on the smoothness of the vehicle is mainly dependent on the angular variation between adjacent position nodes.
First, the location nodeThe last position node in the sequence of position nodes in which it is located +.>The clockwise angle between the connecting line and the initial direction is recorded as the position node +.>Is>。
And arranging the direction angles of all the position nodes in order from small to large to obtain a direction angle sequence.
Average division of a sequence of direction angles intoA sequence of individual direction angles, wherein ∈>Is 10. Counting the number of direction angles in each direction angle sequence interval, and recording the ratio of the total number of the repeated direction angles contained in the direction angle sequence interval to the number of the position nodes as the distribution probability of all the direction angles contained in the direction angle sequence interval.
For example, if the direction angle included in the direction angle sequence section is 15 °,15 °,20 °,30 °,30 °,30 °,45 °, the direction angles of the repeated occurrence included in the direction angle sequence section are 15 ° and 30 °, and the total number of direction angles of the repeated occurrence is 6. The distribution probability of the direction angles reflects the probability of the position nodes generating direction mutation, and when the distribution probability of the direction angles is larger, the more the repeated occurrence times of the direction angles in the direction angle sequence interval are, the smaller the probability of the position nodes generating the direction mutation is, and the more stable the vehicle running direction is at the position of the direction angles.
According to position nodeAnd position node->Adjacent location node->、/>Direction angle difference and position node +.>The distribution probability of the direction angle of (a) to obtain the position node +.>Node mutation Rate of->The calculation formula is as follows:
wherein,for the position node->Is a node direction mutation rate; />For the position node->Is a direction angle of (2); />For the position node->In the sequence of the position nodes, the position node +.>Later location node of (a)Is a direction angle of (2); />For the position node->In the sequence of the position nodes, the position node +.>Preceding position node +.>Is a direction angle of (2); />For the position node->Is a distribution probability of the direction angle of (a).
In the calculation process, the missing numerical values are filled by a mean filling method, so that each position node is ensured to have the corresponding direction angle of the front position node and the rear position node.
When the node direction mutation rate of the position node is smaller than 0, the change between the direction angle of the position node and the direction angle of the position node adjacent to the position node is not monotonic, and the path direction is more tortuous.
When the absolute value of the node direction mutation rate is closer to 1, it is explained that the smaller the amount of change between the direction angles of the position nodes and the direction angles of the position nodes adjacent to the position node is and the larger the distribution probability of the direction angles is, the smaller the local transition of the direction in which the vehicle runs is, and the more stable the vehicle running is.
So far, the node direction mutation rate of all the position nodes on the collision-free path is obtained.
Step S003, acquiring an angle difference of a position node according to a position node adjacent to the position node, acquiring a neighborhood node of the position node in the position node sequence, acquiring a neighborhood direction change difference of the position node according to the angle difference of the neighborhood node of the position node, and acquiring a continuous mutation consistency degree of the position node according to the neighborhood direction change difference of the position node and the node direction mutation rate.
The collision-free path generated by the A-algorithm is a path planning result of avoiding the obstacle, and in the determination process of the weight of the Bezier curve, the position nodes around the obstacle are given a larger weight, so that the curve is as close to the position nodes as possible, and the situation that the unmanned vehicle is scratched and damaged due to the fact that the unmanned vehicle is too close to the obstacle when driving according to the collision-free path is prevented.
In order to avoid an obstacle, a series of position nodes around the obstacle typically require a continuous uniform change of direction to successfully avoid the obstacle. The absolute value of the node direction mutation rate of the position nodes around the obstacle is greatly different from the number 1 and the direction angle change between the position nodes is similar. Therefore, whether the position node is located around the obstacle can be judged according to the node direction mutation rate and the direction angle of the neighborhood nodes of the position node.
Positioning nodeIs>And position node->Is>The difference between them is noted as position node +.>Is provided.
And arranging the angle differences of all the position nodes according to the sequence of the position nodes in the position node sequence to obtain a direction conversion sequence. And performing anomaly detection on all angle differences contained in the direction conversion sequence by using an LOF anomaly detection algorithm to obtain LOF outliers of each angle difference. The method for obtaining the LOF outlier factor by using the LOF anomaly detection algorithm is a known technique and will not be described in detail. When the LOF outlier factor is larger, the numerical distribution of the angle difference corresponding to the LOF outlier factor is more discrete.
Associating a position node in a sequence of position nodes with a position nodeAdjacent->The individual position nodes are designated as position nodes->Is a neighborhood node of (1), wherein->Is 10.
Positioning nodeThe average value of LOF outliers of angle differences of neighboring nodes of (2) is recorded as position node +.>Neighborhood direction change difference->。
According to position nodeNeighborhood direction change difference and node direction mutation rate +.>Acquiring location nodesOrientation continuous mutation identity +.>The calculation formula is as follows:
wherein,for the position node->Is consistent with the continuous mutation in the direction of the (a); />Is->Is->Node direction mutation rates of the neighborhood nodes; />For the position node->Is 10, the empirical value is 10;for the position node->Is a neighborhood direction change variance; />The empirical value is 1 for the first comparison factor.
The smaller the difference between the absolute value of the node direction mutation rate of the neighborhood node of the position node and the number 1 is, the smaller the change amount of the direction angle of the neighborhood node of the position node is, the smaller the direction transition of the vehicle driving is, and the more likely the position node is the position node which is used for avoiding the obstacle and continuously changes in direction around the obstacle, and at the moment, the smaller the continuous direction mutation consistency of the position node is.
So far, the continuous abrupt change consistency of the directions of all the position nodes on the collision-free path is obtained.
Step S004, position nodes of parking point positions on the collision-free path are obtained, direction control nodes are arranged according to the position nodes of the parking point positions, weights of the position nodes of the parking point positions and the direction control nodes are set, and self-adaptive weights of the position nodes on the collision-free path are obtained according to the continuous abrupt change consistency of the directions of the position nodes and the distances between the position nodes and the direction control nodes.
It should be noted that, the collision-free path obtained by the algorithm a is an optimal path based on a starting point and a destination point, and the angle between the direction of the vehicle and the direction of the parking space when the parking point of the idle parking space is reached is not considered, and when the angle between the direction of the vehicle and the direction of the parking space is too large, the vehicle needs to be adjusted for smooth parking for many times.
Will be the firstParking spot of individual free parking spaces>The location node at which is marked +.>Position node->A direction control node is arranged in the parking space direction of the car>Wherein the location node->And a direction control node->The distance between them is->The average value of the distances between all adjacent position nodes on the collision-free path. The schematic diagram of the directional control node arrangement is shown in fig. 3.
Positioning nodeAnd a direction control node->The weight of the vehicle is set to a second preset threshold to ensure that the vehicle is optimally parked when the unmanned vehicle reaches the parking point. Wherein the empirical value of the second preset threshold is 1. Meanwhile, the closer to the position node +.>The position node of (2) should be given a small weight to ensure that the vehicle moves smoothly when reaching the parking spot without large angular rotation.
Based on the analysis, according to the position nodeOrientation continuous mutation identity +.>Location nodePosition node corresponding to parking spot of idle parking space +.>The self-adaptive weight of the position node on the collision-free path is obtained by the distance acquisition, and the calculation formula is as follows:
wherein,for the position node->Is determined by the adaptive weights of (a); />For the position node->Is consistent with the continuous mutation in the direction of the (a); />For the position node->Is->Location node +.>A Euclidean distance between them;is an exponential function based on natural constants; />The empirical value is 1 for the first comparison factor.
When the continuous abrupt change consistency of the direction of the position node is larger, the possibility that the position node is positioned near the obstacle is larger, so that the position node is endowed with larger self-adaptive weight, and the Bezier curve subjected to smoothing is closer to the position node; when the position node is closer to the parking point, smaller adaptive weight needs to be given to the position node, so that the Bezier curve subjected to smoothing processing is smoother and smoother at the position close to the parking point.
To this end, the weights of all the position nodes on all the collision-free paths of the unmanned vehicle are obtained.
And S005, acquiring a smooth path from the unmanned vehicle to a parking point of an idle parking space according to weights and positions of all position nodes on all collision-free paths of the unmanned vehicle, acquiring the degree of confusion and the length of curvature change of the smooth path according to the smooth path, acquiring the path goodness of the smooth path according to the degree of confusion and the length of curvature change of the smooth path, and realizing intelligent parking of the unmanned vehicle according to the path goodness of the smooth path.
And drawing a Bezier curve according to the weight and the position of each position node on the collision-free path, and obtaining a smooth path from the unmanned vehicle to the parking point of the idle parking space. Will beParking spot of individual free parking spaces>Is marked as +.>. The process of drawing the bezier curve is a known technology and will not be described in detail.
In a smooth pathAnd uniformly setting k sample points, wherein the empirical value of k is 200. The radius of curvature at each sample point is obtained. The radius of curvature of the points on the obtained curve is a known technology, and is not described in detail. Taking the sum of the information entropy of the curvature radii of all sample points as a smooth path +.>Is a degree of confusion of curvature variation. The greater the degree of confusion of curvature variation, the vehicle is moving on a smooth pathThe larger the direction change amplitude is, the more chaotic the direction change is.
The length of the smooth path is obtained using an arc length integral formula. The length of the curve calculated by arc length integration is a known technique, and will not be described in detail.
According to a smooth pathIs to calculate a smooth path +.>Path goodness->。
Wherein,for smooth path->Path goodness of (3); />For smooth path->Is a degree of confusion of curvature variation;for smooth path->Is a length of (2); />The first weight is the weight of the degree of disorder of curvature change, and the empirical value is 0.6; />Is an exponential function based on natural constants.
The value range of the first weight is a constant which is larger than 0.5 and smaller than 1, and the constant is used for comprehensively acquiring the path goodness of the curvature change disorder degree and the length of the smooth path, and meanwhile, the influence of the curvature change disorder degree of the smooth path on the path goodness is larger.
The smaller the length and curvature variation degree of the smooth path, the greater the path goodness of the smooth path, and at this time, the more suitable the smooth path as a parking path of the unmanned vehicle.
And obtaining the path goodness of the smooth paths from the unmanned vehicle to all the idle parking spaces, taking the path with the maximum path goodness as the optimal path, and taking the idle parking space corresponding to the optimal path as the optimal parking space.
And controlling the vehicle to reach the parking point of the optimal parking space along the optimal path through the vehicle control system and parking.
Thus, intelligent parking of the unmanned vehicle is completed.
The foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.

Claims (10)

1. An intelligent parking method for an unmanned vehicle, which is characterized by comprising the following steps:
collecting parking points of idle parking spaces of the unmanned vehicle, and obtaining collision-free paths from the unmanned vehicle to the parking points of all the idle parking spaces;
acquiring position nodes on a collision-free path, acquiring a position node sequence, acquiring direction angles of the position nodes according to the position node sequence, acquiring a direction angle sequence, acquiring the distribution probability of the direction angles according to the direction angle sequence, and acquiring the node direction mutation rate according to the distribution probability of the direction angles and the adjacent position nodes of the direction angles in the position node sequence;
acquiring the angle difference of the position node according to the adjacent position node of the position node, acquiring the neighborhood node of the position node in the position node sequence, acquiring the neighborhood direction change difference of the position node according to the angle difference of the neighborhood node of the position node, and acquiring the continuous mutation consistency of the direction of the position node according to the neighborhood direction change difference of the position node and the node direction mutation rate;
acquiring position nodes of parking point positions on a collision-free path, setting direction control nodes according to the position nodes of the parking point positions, setting weights of the position nodes of the parking point positions and the direction control nodes, and acquiring self-adaptive weights of the position nodes on the collision-free path according to the continuous abrupt change consistency of the directions of the position nodes and the distances between the position nodes and the direction control nodes;
according to the weights and positions of all position nodes on all collision-free paths of the unmanned vehicle, obtaining a smooth path from the unmanned vehicle to a parking point of an idle parking space, obtaining the degree of confusion and the length of curvature change of the smooth path according to the smooth path, obtaining the path goodness of the smooth path according to the degree of confusion and the length of curvature change of the smooth path, and realizing intelligent parking of the unmanned vehicle according to the path goodness of the smooth path.
2. The intelligent parking method of an unmanned vehicle according to claim 1, wherein the acquiring the sequence of location nodes comprises the following specific steps:
and arranging the position nodes on the collision-free path according to the direction from the unmanned vehicle to the parking point of the idle parking space, and acquiring a position node sequence.
3. The intelligent parking method of an unmanned vehicle according to claim 1, wherein the obtaining the direction angle of the location node according to the location node sequence, obtaining the direction angle sequence, comprises the following specific steps:
each position node in the position node sequence is respectively used as a position node to be analyzed;
the last position node in the position node sequence where the position node to be analyzed is positioned is marked as a first adjacent position node of the position node to be analyzed;
the clockwise included angle between the connecting line between the position node to be analyzed and the first adjacent position node of the position node to be analyzed and the initial direction is recorded as the direction angle of the position node to be analyzed;
and arranging the direction angles of all the position nodes in order from small to large to obtain a direction angle sequence.
4. A method for intelligent parking of an unmanned vehicle according to claim 3, wherein the obtaining the distribution probability of the direction angle according to the direction angle sequence comprises the following specific steps:
dividing the direction angle sequence average into a first division threshold value number of direction angle sequence intervals;
the ratio of the total number of the repeated occurrence of direction angles contained in the direction angle sequence section to the number of the position nodes is recorded as the distribution probability of all direction angles contained in the direction angle sequence section.
5. The intelligent parking method of an unmanned vehicle according to claim 3, wherein the obtaining the angular difference of the location node according to the neighboring location nodes of the location node comprises the following specific steps:
and recording the difference between the direction angle of the position node to be analyzed and the direction angle of the first adjacent position node of the position node to be analyzed as the angle difference of the position node to be analyzed.
6. The intelligent parking method of an unmanned vehicle according to claim 3, wherein the obtaining the neighborhood node of the position node in the position node sequence obtains the neighborhood direction change difference of the position node according to the angle difference of the neighborhood node of the position node, and the specific method comprises the following steps:
marking a second division threshold position node adjacent to the position node to be analyzed in the position node sequence as a neighborhood node of the position node to be analyzed;
and carrying out anomaly detection on the angle differences of all the position nodes on the same collision-free path, obtaining outlier factors of each position node, and recording the average value of the outlier factors of the angle differences of the neighborhood nodes of the position node to be analyzed as the neighborhood direction change difference of the position node to be analyzed.
7. The intelligent parking method of an unmanned vehicle according to claim 1, wherein the setting of the direction control node according to the position node of the parking spot position comprises the following specific steps:
setting a direction control node in the parking space direction of a position node corresponding to the parking point of the idle parking space, wherein the distance between the position node corresponding to the parking point of the idle parking space and the direction control node is the average value of the distances between all adjacent position nodes on a collision-free path where the position node corresponding to the parking point of the idle parking space is located.
8. The intelligent parking method of an unmanned vehicle according to claim 1, wherein the weights of the position node and the direction control node for setting the parking point position comprise the following specific methods:
and setting the weights of the position nodes and the direction control nodes corresponding to the parking points of the idle parking spaces as a first preset threshold.
9. The intelligent parking method of an unmanned vehicle according to claim 1, wherein the obtaining the degree and length of confusion of curvature change of the smooth path according to the smooth path comprises the following specific steps:
uniformly setting a third division threshold value on the smooth path, acquiring the curvature radius of each sample point, and taking the sum of the information entropy of the curvature radius of all sample points as the curvature change chaotic degree of the smooth path;
the length of the smooth path is obtained using an arc length integral formula.
10. The intelligent parking method of an unmanned vehicle according to claim 1, wherein the intelligent parking of the unmanned vehicle is realized according to the path goodness of the smooth path, comprising the following specific steps:
and taking the path with the maximum path goodness as the optimal path, taking the idle parking space corresponding to the optimal path as the optimal parking space, controlling the vehicle to reach the parking point of the optimal parking space along the optimal path by using the vehicle control system, and parking, thereby realizing intelligent parking of the unmanned vehicle.
CN202311737049.0A 2023-12-18 2023-12-18 Intelligent parking method for unmanned vehicle Active CN117416342B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311737049.0A CN117416342B (en) 2023-12-18 2023-12-18 Intelligent parking method for unmanned vehicle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311737049.0A CN117416342B (en) 2023-12-18 2023-12-18 Intelligent parking method for unmanned vehicle

Publications (2)

Publication Number Publication Date
CN117416342A CN117416342A (en) 2024-01-19
CN117416342B true CN117416342B (en) 2024-03-08

Family

ID=89523381

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311737049.0A Active CN117416342B (en) 2023-12-18 2023-12-18 Intelligent parking method for unmanned vehicle

Country Status (1)

Country Link
CN (1) CN117416342B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113932823A (en) * 2021-09-23 2022-01-14 同济大学 Unmanned multi-target-point track parallel planning method based on semantic road map
CN114771509A (en) * 2022-05-24 2022-07-22 杨春萍 Intelligent parking path planning control method for driving assistance multi-environment
CN115062261A (en) * 2022-06-24 2022-09-16 西安交通大学 Curvature continuous autonomous parking path planning method and system based on reverse search
CN115817455A (en) * 2022-11-09 2023-03-21 同济大学 Automatic parking path planning method oriented to multiple parking space scenes
CN115938154A (en) * 2022-11-08 2023-04-07 吉林大学 Method for setting autonomous parking system of large electric truck based on field-side cooperation
CN116373851A (en) * 2023-03-29 2023-07-04 深圳市欧冶半导体有限公司 Automatic parking path planning method, automatic parking method and related device
CN116620264A (en) * 2023-05-10 2023-08-22 北京理工大学 Parking task allocation and track planning system based on multi-agent reinforcement learning

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10012984B2 (en) * 2015-12-14 2018-07-03 Mitsubishi Electric Research Laboratories, Inc. System and method for controlling autonomous vehicles
US11767035B2 (en) * 2020-09-04 2023-09-26 Mitsubishi Electric Research Laboratories, Inc. Autonomous parking with hybrid exploration of parking space

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113932823A (en) * 2021-09-23 2022-01-14 同济大学 Unmanned multi-target-point track parallel planning method based on semantic road map
CN114771509A (en) * 2022-05-24 2022-07-22 杨春萍 Intelligent parking path planning control method for driving assistance multi-environment
CN115062261A (en) * 2022-06-24 2022-09-16 西安交通大学 Curvature continuous autonomous parking path planning method and system based on reverse search
CN115938154A (en) * 2022-11-08 2023-04-07 吉林大学 Method for setting autonomous parking system of large electric truck based on field-side cooperation
CN115817455A (en) * 2022-11-09 2023-03-21 同济大学 Automatic parking path planning method oriented to multiple parking space scenes
CN116373851A (en) * 2023-03-29 2023-07-04 深圳市欧冶半导体有限公司 Automatic parking path planning method, automatic parking method and related device
CN116620264A (en) * 2023-05-10 2023-08-22 北京理工大学 Parking task allocation and track planning system based on multi-agent reinforcement learning

Also Published As

Publication number Publication date
CN117416342A (en) 2024-01-19

Similar Documents

Publication Publication Date Title
CN112099493B (en) Autonomous mobile robot trajectory planning method, system and equipment
WO2021175313A1 (en) Automatic driving control method and device, vehicle, and storage medium
CN110749333B (en) Unmanned vehicle motion planning method based on multi-objective optimization
CN113204236B (en) Intelligent agent path tracking control method
CN110395250B (en) Method and system for vehicle curve speed limiting
CN109947136B (en) Collaborative active sensing method for unmanned aerial vehicle group rapid target search
CN109765887B (en) Automatic driving control method
Magid et al. Spline-based robot navigation
CN109885883A (en) A kind of control method of the unmanned vehicle transverse movement based on GK clustering algorithm model prediction
CN110766220A (en) Local path planning method for structured road
CN110146085B (en) Unmanned aerial vehicle real-time avoidance rescheduling method based on graph building and rapid random tree exploration
CN113859266B (en) Method and system for predicting lane change track of unstructured road target vehicle
CN113721637B (en) Intelligent vehicle dynamic obstacle avoidance path continuous planning method and system and storage medium
CN112965476B (en) High-speed unmanned vehicle trajectory planning system and method based on multi-window model
CN112506199B (en) Local path planning method based on dynamic window method and suitable for Ackerman model robot
CN110146087B (en) Ship path planning method based on dynamic planning idea
CN110789530B (en) Four-wheel independent steering-independent driving vehicle trajectory tracking method and system
CN112415999B (en) Processing method, device and equipment for automatic driving path point and storage medium
CN116185014A (en) Intelligent vehicle global optimal track planning method and system based on dynamic planning
CN116576857A (en) Multi-obstacle prediction navigation obstacle avoidance method based on single-line laser radar
CN113886764A (en) Intelligent vehicle multi-scene track planning method based on Frenet coordinate system
CN114906173A (en) Automatic driving decision-making method based on two-point preview driver model
CN115861968A (en) Dynamic obstacle removing method based on real-time point cloud data
CN113895463A (en) Path planning method suitable for automatic driving vehicle turning
CN116182884A (en) Intelligent vehicle local path planning method based on transverse and longitudinal decoupling of frenet coordinate system

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