CN111427368A - Improved multi-target collision-prevention driving method for unmanned intelligent vehicle - Google Patents
Improved multi-target collision-prevention driving method for unmanned intelligent vehicle Download PDFInfo
- Publication number
- CN111427368A CN111427368A CN202010485077.8A CN202010485077A CN111427368A CN 111427368 A CN111427368 A CN 111427368A CN 202010485077 A CN202010485077 A CN 202010485077A CN 111427368 A CN111427368 A CN 111427368A
- Authority
- CN
- China
- Prior art keywords
- intelligent vehicle
- unmanned intelligent
- subgroup
- target
- collision avoidance
- 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.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 230000008569 process Effects 0.000 claims abstract description 17
- 230000002068 genetic effect Effects 0.000 claims abstract description 9
- 230000006870 function Effects 0.000 claims description 22
- 238000013508 migration Methods 0.000 claims description 20
- 230000005012 migration Effects 0.000 claims description 20
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 5
- 210000002569 neuron Anatomy 0.000 claims description 5
- 230000004888 barrier function Effects 0.000 claims description 4
- 238000012544 monitoring process Methods 0.000 claims description 4
- 206010063385 Intellectualisation Diseases 0.000 abstract description 2
- 230000036544 posture Effects 0.000 description 6
- 238000011160 research Methods 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- QZXCCPZJCKEPSA-UHFFFAOYSA-N chlorfenac Chemical compound OC(=O)CC1=C(Cl)C=CC(Cl)=C1Cl QZXCCPZJCKEPSA-UHFFFAOYSA-N 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- LFULEKSKNZEWOE-UHFFFAOYSA-N propanil Chemical compound CCC(=O)NC1=CC=C(Cl)C(Cl)=C1 LFULEKSKNZEWOE-UHFFFAOYSA-N 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0257—Control of position or course in two dimensions specially adapted to land vehicles using a radar
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0276—Control of position or course in two dimensions specially adapted to land vehicles using signals provided by a source external to the vehicle
- G05D1/0278—Control of position or course in two dimensions specially adapted to land vehicles using signals provided by a source external to the vehicle using satellite positioning signals, e.g. GPS
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Aviation & Aerospace Engineering (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Theoretical Computer Science (AREA)
- Genetics & Genomics (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Physiology (AREA)
- Electromagnetism (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention belongs to the technical field of intelligent vehicle path planning, and particularly relates to an improved multi-target unmanned intelligent vehicle collision avoidance driving method. The invention comprises the following steps: acquiring the position, attitude angle and obstacle position of the unmanned intelligent vehicle; calculating the collision risk of the unmanned intelligent vehicle; and planning a collision avoidance path of the multi-target unmanned intelligent vehicle. According to the unmanned intelligent vehicle, the collision avoidance countermeasures are formulated by collecting and analyzing the information of the obstacles, a series of processes of avoiding the obstacles are completed, the intellectualization of obstacle avoidance of the unmanned intelligent vehicle can be effectively improved, and the workload of drivers is reduced; the invention plans the driving path of the unmanned intelligent vehicle by using the distributed genetic algorithm, overcomes the problem that the traditional algorithm is easy to fall into the local optimal solution, and improves the quality of the planned path.
Description
Technical Field
The invention belongs to the technical field of intelligent vehicle path planning, and particularly relates to an improved multi-target unmanned intelligent vehicle collision avoidance driving method.
Background
The unmanned intelligent vehicle is a robot system which can sense the surrounding environment and the self state of the vehicle in a certain mode, can realize the autonomous movement facing a set target in a road section with obstacles or other environments, and further can complete the set operation function. To date, the development of unmanned smart vehicle systems over long periods of research and experimentation, with the efforts of researchers, has achieved surprising results and preoccupation experience. In an environment which can be determined in advance, researches related to intelligent vehicle navigation and obstacle avoidance strategies have achieved considerable results and applications, but in an unknown environment, the results obtained by the related researches cannot achieve the expected targets of people. The generation of new problems and new requirements can never stop, the problems of the intelligent vehicle in practical application still have no better solution due to the limitations of some basic theories and technologies, a complete theoretical system is not formed, only less prior knowledge is needed, and a plurality of key theories and experiments are needed to be researched and verified. Therefore, the intelligent vehicle still has various defects in the driving process of an unknown environment and an unstructured road. On the basis of research on the intelligent vehicle collision avoidance risk degree and the intelligent vehicle collision avoidance rule, the idea of planning an intelligent vehicle collision avoidance path by using an improved distributed genetic algorithm is proposed by using a GSP (global system for plant) or the like. The distributed genetic algorithm is a parallel algorithm, has the capability of fast random search, and is generally accepted by scholars at home and abroad after being proposed. Therefore, the invention hopes to explore a barrier detection method of the unmanned intelligent vehicle in a low-speed environment and realize verification of an optimization scheme. The intelligent vehicle is in line with the development direction of the current intelligent vehicle, provides experience for the application of the intelligent vehicle in related fields and has profound significance.
Disclosure of Invention
The invention aims to provide an improved multi-target unmanned intelligent vehicle collision avoidance driving method which enables an unmanned intelligent vehicle to effectively realize multi-target unmanned intelligent vehicle collision avoidance path planning.
The purpose of the invention is realized as follows:
the improved multi-target unmanned intelligent vehicle collision avoidance driving method comprises the following steps:
(1) acquiring the position, attitude angle and obstacle position of the unmanned intelligent vehicle;
(2) calculating the collision risk of the unmanned intelligent vehicle;
(3) and planning a collision avoidance path of the multi-target unmanned intelligent vehicle.
The position, the attitude angle and the obstacle position of the unmanned intelligent vehicle are obtained, and the specific process is as follows:
the method comprises the steps of planning an optimal path for the unmanned intelligent vehicle to run for a known surrounding road environment, running the unmanned intelligent vehicle in the road environment according to the optimal path when the collision avoidance requirement is not met, scanning and monitoring the surrounding road environment to judge whether collision avoidance targets exist or not, and respectively measuring the position, the attitude angle and the position of an obstacle of the unmanned intelligent vehicle in running by using a sensor of the unmanned intelligent vehicle if other targets exist on the road during running.
Calculating the collision risk of the unmanned intelligent vehicle, and the specific process is as follows:
and taking the nearest meeting distance and the nearest meeting time as the input of the BP neural network, and taking the collision risk of the unmanned intelligent vehicle as the network output. And obtaining the connection weight and the threshold of each neuron through a BP neural network, and outputting the collision risk of the unmanned intelligent vehicle.
The specific process for planning the collision avoidance path of the multi-target unmanned intelligent vehicle comprises the following steps:
3.1 when the unmanned intelligent vehicle runs according to the planned optimal path before starting, detecting other targets existing on the road in an iterative mode, and judging the risk degree of the targets according to target parameters;
3.2 if the risk degree of the target is higher than 0.2, entering a collision avoidance procedure; otherwise, the unmanned intelligent vehicle continues to run along the set path;
3.3 judging the meeting attitude angle of the unmanned intelligent vehicle and the target;
3.4 planning a reasonable collision avoidance path according to the information collected by the system;
3.5 if the risk degree of all the targets is less than 0.3, returning to the original path to continue driving;
3.6 output the best feasible path to avoid obstacles.
The detection of other objects present on the road in an iterative manner includes in particular:
3.1.1: the target is encoded.
3.1.2: the target is taken as a population and the population is initialized.
3.1.3: generating subgroups.
3.1.4: and executing the original genetic algorithm in the subgroup when the migration period T is not reached, and executing 3.1.5 after the integral multiple of the migration period T is reached.
3.1.5: and obtaining the optimal individual of each subgroup, comparing to obtain the optimal individual of the subgroup, judging whether the maximum iteration times is reached, outputting a result if the maximum iteration times is reached, otherwise, performing a migration operation, and executing 3.1.6.
3.1.6: and calculating the distribution space of each subgroup according to a subgroup distribution formula, if the distribution space is larger than the original space of the subgroup, comparing and judging whether the space of the subgroup i is larger than that of the subgroup j, if the space of the subgroup i is larger than that of the subgroup j, copying the optimal individual of the subgroup j into the subgroup i, otherwise, copying the optimal individual of the subgroup i into the subgroup j. If the allocated space is smaller than the original space of the subgroup, part of individuals are discarded at will.
3.1.7: regrouping the subgroup, returning to step 3.1.3.
The migration operation comprises the following steps:
3.1.5.1 subgroup reassignment: in the process of target migration, calculating the fitness of the optimal individual in each subgroup and queuing each subgroup according to the sequence of the fitness from large to small, wherein the space size obtained by the next evolution of the subgroup is in a linear relation with the current fitness thereof:
qjthe number of individuals assigned to the next jth subgroup, f (q)j) Is the fitness value of the best individual of the jth subgroup,
is the sum of fitness values of all subgroups and q is the individual number of the jth subgroup.
3.1.5.2 migration of individuals: selecting individuals in the sub-population according to the size of the fitness value as a migration object, namely selecting the individuals with high fitness function values, wherein the sub-population with the fitness function lower than a threshold value migrates the local optimal solution of the sub-population to the sub-population with the fitness function higher than the threshold value, and the following formula is shown:
Mqj＝{qi|f(qj)≤f(qi)}
wherein M isqjAnd i represents a subgroup number different from j, which represents a subgroup number of jth subgroup individuals having a smaller fitness value than j.
The invention has the following beneficial effects:
according to the unmanned intelligent vehicle, the collision avoidance countermeasures are formulated by collecting and analyzing the information of the obstacles, a series of processes of avoiding the obstacles are completed, the intellectualization of obstacle avoidance of the unmanned intelligent vehicle can be effectively improved, and the workload of drivers is reduced; the invention plans the driving path of the unmanned intelligent vehicle by using the distributed genetic algorithm, overcomes the problem that the traditional algorithm is easy to fall into the local optimal solution, and improves the quality of the planned path. The invention sets different fitness functions by detecting different meeting postures, and is beneficial to planning a safe and economic route for the unmanned intelligent vehicle. According to the invention, the collision avoidance strategy is formulated and implemented according to the risk of the obstacles and the road collision avoidance rule, so that the safety of the unmanned intelligent vehicle during running is ensured.
Drawings
Fig. 1 is a flow chart of a collision avoidance driving method of a multi-target unmanned intelligent vehicle.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The invention relates to an improved multi-target collision-prevention driving method for an unmanned intelligent vehicle. And a distributed genetic algorithm is improved, so that the unmanned intelligent vehicle can independently search for obstacles in the driving process, and a collision avoidance strategy is implemented according to the distribution condition of the obstacles. The purpose of the invention is realized by the following steps:
1. acquiring the position and attitude angle of the obstacle and the position of the unmanned intelligent vehicle:
the position information of the unmanned intelligent vehicle and the obstacle is measured by position reference systems such as a satellite, a laser and a radar, and the attitude angle information such as the direction of the unmanned intelligent vehicle is measured by an electric compass, a motion reference unit and the like. And filtering and time-space alignment are carried out on the obtained attitude angle and position information to obtain the accurate position attitude of the unmanned intelligent vehicle and the position of the obstacle.
2. And establishing a collision risk degree model of the unmanned intelligent vehicle, and calculating the collision risk degree of the ship.
3. And determining a target function and planning a multi-target collision avoidance path of the unmanned intelligent vehicle.
The specific details include:
an improved multi-target unmanned intelligent vehicle collision avoidance driving method comprises the following steps:
(1) acquiring the position, attitude angle and obstacle position of the unmanned intelligent vehicle; the position information of the unmanned intelligent vehicle and the obstacle is measured by a satellite, a laser or a radar, and the direction and attitude angle information of the unmanned intelligent vehicle is measured by an electric compass or a motion reference unit; filtering and time-space alignment are carried out on the obtained attitude angle and position information to obtain the accurate position attitude of the unmanned intelligent vehicle and the position of the obstacle;
(2) calculating the collision risk of the unmanned intelligent vehicle;
(3) and planning a collision avoidance path of the multi-target unmanned intelligent vehicle.
The position, the attitude angle and the obstacle position of the unmanned intelligent vehicle are obtained, and the specific process is as follows:
the method comprises the steps of planning an optimal path for the unmanned intelligent vehicle to run for a known surrounding road environment, running the unmanned intelligent vehicle in the road environment according to the optimal path when the collision avoidance requirement is not met, scanning and monitoring the surrounding road environment to judge whether collision avoidance targets exist or not, and respectively measuring the position, the attitude angle and the position of an obstacle of the unmanned intelligent vehicle in running by using a sensor of the unmanned intelligent vehicle if other targets exist on the road during running.
Calculating the collision risk of the unmanned intelligent vehicle, and the specific process is as follows:
taking the nearest meeting distance and the nearest meeting time as the input of a BP neural network, and taking the collision risk of the unmanned intelligent vehicle as the network output; and obtaining the connection weight and the threshold of each neuron through a BP neural network, and outputting the collision risk of the unmanned intelligent vehicle.
The specific process for planning the collision avoidance path of the multi-target unmanned intelligent vehicle comprises the following steps:
3.1 when the unmanned intelligent vehicle runs according to the planned optimal path before starting, detecting other targets existing on the road in an iterative mode, and judging the risk degree of the targets according to target parameters;
3.2 if the risk degree of the target is higher than 0.2, entering a collision avoidance procedure; otherwise, the unmanned intelligent vehicle continues to run along the set path;
3.3 judging the meeting attitude angle of the unmanned intelligent vehicle and the target;
3.4 planning a reasonable collision avoidance path according to the information collected by the system;
3.5 if the risk degree of all the targets is less than 0.3, returning to the original path to continue driving;
3.6 output the best feasible path to avoid obstacles.
Detecting other objects present on the road in an iterative manner, including in particular:
3.1.1: encoding the target;
3.1.2: taking the target as a population and initializing the population;
3.1.3: generating a subgroup;
3.1.4: executing an original genetic algorithm in the subgroup when the migration period T is not reached, and executing 3.1.5 after the integral multiple of the migration period T is reached;
3.1.5: obtaining the optimal individuals of each subgroup, comparing the optimal individuals of the subgroups to obtain the optimal individuals of the population, judging whether the maximum iteration times is reached, if the maximum iteration times is reached, outputting a result, and if the maximum iteration times is not reached, performing migration operation and executing 3.1.6;
3.1.6: calculating the distribution space of each subgroup according to a subgroup distribution formula, if the distribution space is larger than the original space of the subgroup, comparing and judging whether the space of the subgroup i is larger than that of the subgroup j, if the space of the subgroup i is larger than that of the subgroup j, copying the optimal individual of the subgroup j into the subgroup i, otherwise, copying the optimal individual of the subgroup i into the subgroup j; if the distribution space is smaller than the original space of the subgroup, part of individuals are discarded at will;
3.1.7: regrouping the subgroup, returning to step 3.1.3.
The migration operation comprises the following steps:
3.1.5.1 subgroup reassignment: in the process of target migration, calculating the fitness of the optimal individual in each subgroup and queuing each subgroup according to the sequence of the fitness from large to small, wherein the space size obtained by the next evolution of the subgroup is in a linear relation with the current fitness thereof:
qjthe number of individuals assigned to the next jth subgroup, f (q)j) Is the fitness value of the best individual of the jth subgroup,
is the sum of fitness values of all subgroups, q is the individual number of the jth subgroup;
3.1.5.2 migration of individuals: selecting individuals in the sub-population according to the size of the fitness value as a migration object, namely selecting the individuals with high fitness function values, wherein the sub-population with the fitness function lower than a threshold value migrates the local optimal solution of the sub-population to the sub-population with the fitness function higher than the threshold value, and the following formula is shown:
Mqj＝{qi|f(qj)≤f(qi)}
wherein M isqjAnd i represents a subgroup number different from j, which represents a subgroup number of jth subgroup individuals having a smaller fitness value than j.
When the size of the sub-population is reduced, the part of individuals with the minimum fitness is discarded; when the size of the sub-population increases, in addition to absorbing the optimal individuals from the outside, one can randomly replicate themselves to fill the space. The sub-population with the highest fitness does not migrate any individual to other sub-populations, while the sub-population with the lowest fitness does not migrate any individual and loses a part of the individuals
Filtering and time-space aligning the acquired attitude angle and position information comprises:
the driving direction of the unmanned intelligent vehicle is as follows:
in the formula vOx，vOyAre respectively unmannedThe speed of the vehicle on the x and y axes of the reference system, β0Setting an initial angle for the unmanned intelligent vehicle;
according to the geographic coordinates of the unmanned intelligent vehicle and the obstacle, the relative distance between the unmanned intelligent vehicle and the obstacle is calculated as follows:
in the formula, xT，yTIs the horizontal and vertical coordinates of the obstacle; x is the number ofO，yOIs the horizontal and vertical coordinates of the unmanned intelligent vehicle.
The true orientation of the obstacle relative to the unmanned intelligent vehicle is thetaT：
The true orientation of the unmanned intelligent vehicle relative to the barrier is theta0：
The phase orientation of the obstacle is αT：
Relative driving direction of obstacle relative to unmanned intelligent vehicle
And acquiring the position and the posture of the obstacle and the unmanned intelligent vehicle.
The method comprises the steps of planning an optimal path for the unmanned intelligent vehicle to run for a known surrounding road environment, running the unmanned intelligent vehicle in the road environment according to the optimal path when the collision avoidance requirement is not met, scanning and monitoring the surrounding road environment to judge whether collision avoidance targets exist or not, and respectively measuring the position, the attitude angle and the position of an obstacle of the unmanned intelligent vehicle in running by using a sensor of the unmanned intelligent vehicle if other targets exist on the road during running.
The driving direction of the unmanned intelligent vehicle is as follows:
in the formula vOx，vOyThe speed of the unmanned intelligent vehicle on the x axis and the y axis of the reference system is β0The initial angle of the unmanned intelligent vehicle is set.
According to the geographic coordinates of the unmanned intelligent vehicle and the obstacle, the relative distance between the unmanned intelligent vehicle and the obstacle is calculated as follows:
in the formula, xT，yTIs the horizontal and vertical coordinates of the obstacle; x is the number ofO，yOIs the horizontal and vertical coordinates of the unmanned intelligent vehicle.
The true orientation of the obstacle relative to the unmanned intelligent vehicle is thetaT：
The true orientation of the unmanned intelligent vehicle relative to the barrier is theta0：
The phase orientation of the obstacle is αT：
Relative driving direction of obstacle relative to unmanned intelligent vehicle
And establishing a collision avoidance risk degree model of the unmanned intelligent vehicle, and calculating a collision risk degree value of the target and the unmanned intelligent vehicle.
In the driving process, if other targets exist on the road, the risk degree of the unmanned intelligent vehicle is calculated by using target driving data measured by a measuring system on the unmanned intelligent vehicle.
Nearest meeting distance:
recent times encountered:
in the research of the collision risk of the unmanned intelligent vehicle, DCPA and TCPA are two important reasons for influencing the risk. In order to make the calculation speed faster, the values of the DCPA and the TCPA are used as the input of a BP neural network, and the collision risk of the unmanned intelligent vehicle is used as the network output. And obtaining the connection weight and the threshold of each neuron by learning the expert data.
According to the traffic rules, aiming at the specific meeting posture of the unmanned intelligent vehicle and the target, the system should make different collision avoidance instructions. That is, different fitness functions are set under different meeting postures. In order to obtain a safe and economical ride, the fitness function should consist of two parts.
The first part is a safety evaluation function of the path, a target with the highest risk degree is found firstly, the shortest distance between the target and the unmanned intelligent vehicle is obtained, and the safety of the path is evaluated according to the preset safety distance. The security evaluation function is as follows:
wherein, x is an individual and represents a flight path, i is any node on the path, the value of i is from 1 to N-1, N is the total number of nodes in the individual, that is to say, each path has N-1 small segments
When the unmanned intelligent vehicle runs along the collision avoidance path, different meeting situations can be met on different small sections, when the unmanned intelligent vehicle runs to one node, all surrounding targets can be detected, the collision risk degree of each target is calculated, and the target with the largest collision risk degree with the unmanned intelligent vehicle is found. giThe minimum distance between the target with the maximum collision risk degree with the unmanned intelligent vehicle and the unmanned intelligent vehicle is shown from the ith node to the (i + 1) th node. d represents the set minimum safe distance. Since the minimum value is found for the target, g is the minimum value in a safe situationiWhen > d, clear (x)i) The value of (A) will be small, in case of insecurity, the opposite is clear (x)i) The value of (a) is large. Where k and h are proportionality coefficients, and k is 0.01 and h is 10.
When the unmanned intelligent vehicle runs on an avoidance path, meeting postures of different nodes and targets are changed, so that the algorithm is difficult to realize, and corresponding simplification is performed. The method comprises the steps of firstly finding a target with the largest collision risk degree with the unmanned intelligent vehicle at a certain node, judging the meeting posture of the unmanned intelligent vehicle, analyzing the position condition of the unmanned intelligent vehicle from the current node to the next node, and judging the steering condition of the unmanned intelligent vehicle. The traffic rule evaluation function is as follows:
in the formula:
setting T _ Cost (x)i)∈(0,1]The smaller the value of the function is, the more the traffic rule is obeyed by the section of the path. Firstly, sorting the collision risk degrees of each target and the unmanned intelligent vehicle on each section of route so as to find a target j with the maximum collision risk degree with the unmanned intelligent vehicle, and judging the conformity of each route to the traffic rules according to a meeting pattern formed by the target j and the unmanned intelligent vehicle.
And (3) adding a proper weight according to the two evaluation functions to form an objective function:
f(x)＝0.5S(x)+0.5T(x)
because the minimum problem is to be solved, according to the basic knowledge of the genetic algorithm, the fitness function is determined as follows:
according to the fitness function, when the evolution degree of the population meets the termination condition, the individual with the maximum fitness value in the last generation of individuals is the optimal solution required by the problem.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. An improved multi-target unmanned intelligent vehicle collision avoidance driving method is characterized by comprising the following steps:
(1) acquiring the position, attitude angle and obstacle position of the unmanned intelligent vehicle; the position information of the unmanned intelligent vehicle and the obstacle is measured by a satellite, a laser or a radar, and the direction and attitude angle information of the unmanned intelligent vehicle is measured by an electric compass or a motion reference unit; filtering and time-space alignment are carried out on the obtained attitude angle and position information to obtain the accurate position attitude of the unmanned intelligent vehicle and the position of the obstacle;
(2) calculating the collision risk of the unmanned intelligent vehicle;
(3) and planning a collision avoidance path of the multi-target unmanned intelligent vehicle.
2. The improved multi-target collision avoidance driving method for the unmanned intelligent vehicle as claimed in claim 1, wherein the position, attitude angle and obstacle position of the unmanned intelligent vehicle are obtained by the following specific processes:
the method comprises the steps of planning an optimal path for the unmanned intelligent vehicle to run for a known surrounding road environment, running the unmanned intelligent vehicle in the road environment according to the optimal path when the collision avoidance requirement is not met, scanning and monitoring the surrounding road environment to judge whether collision avoidance targets exist or not, and respectively measuring the position, the attitude angle and the obstacle position of the running unmanned intelligent vehicle by using a sensor of the unmanned intelligent vehicle if the collision avoidance targets exist on the road during running.
3. The improved multi-target collision avoidance driving method for the unmanned intelligent vehicle as claimed in claim 2, wherein the collision risk of the unmanned intelligent vehicle is calculated by the following specific processes:
taking the nearest meeting distance and the nearest meeting time as the input of a BP neural network, and taking the collision risk of the unmanned intelligent vehicle as the network output; and obtaining the connection weight and the threshold of each neuron through a BP neural network, and outputting the collision risk of the unmanned intelligent vehicle.
4. The improved multi-target unmanned intelligent vehicle collision avoidance driving method according to claim 3, wherein the specific process of planning the multi-target unmanned intelligent vehicle collision avoidance path is as follows:
3.1 when the unmanned intelligent vehicle runs according to the planned optimal path before starting, detecting other targets existing on the road in an iterative mode, and judging the risk degree of the targets according to target parameters;
3.2 if the risk degree of the target is higher than 0.2, entering a collision avoidance procedure; otherwise, the unmanned intelligent vehicle continues to run along the set path;
3.3 judging the meeting attitude angle of the unmanned intelligent vehicle and the target;
3.4 planning a reasonable collision avoidance path according to the information collected by the system;
3.5 if the risk degree of all the targets is less than 0.3, returning to the original path to continue driving;
3.6 output the best feasible path to avoid obstacles.
5. The improved multi-target unmanned intelligent vehicle collision avoidance driving method according to claim 4, wherein other targets existing on the road are detected in an iterative manner, and the method specifically comprises the following steps:
3.1.1: encoding the target;
3.1.2: taking the target as a population and initializing the population;
3.1.3: generating a subgroup;
3.1.4: executing an original genetic algorithm in the subgroup when the migration period T is not reached, and executing 3.1.5 after the integral multiple of the migration period T is reached;
3.1.5: obtaining the optimal individuals of each subgroup, comparing the optimal individuals of the subgroups to obtain the optimal individuals of the population, judging whether the maximum iteration times is reached, if the maximum iteration times is reached, outputting a result, and if the maximum iteration times is not reached, performing migration operation and executing 3.1.6;
3.1.6: calculating the distribution space of each subgroup according to a subgroup distribution formula, if the distribution space is larger than the original space of the subgroup, comparing and judging whether the space of the subgroup i is larger than that of the subgroup j, if the space of the subgroup i is larger than that of the subgroup j, copying the optimal individual of the subgroup j into the subgroup i, otherwise, copying the optimal individual of the subgroup i into the subgroup j; if the distribution space is smaller than the original space of the subgroup, part of individuals are discarded at will;
3.1.7: regrouping the subgroup, returning to step 3.1.3.
6. The improved multi-target unmanned intelligent vehicle collision avoidance driving method according to claim 5, wherein the transferring operation comprises:
3.1.5.1 subgroup reassignment: in the process of target migration, calculating the fitness of the optimal individual in each subgroup and queuing each subgroup according to the sequence of the fitness from large to small, wherein the space size obtained by the next evolution of the subgroup is in a linear relation with the current fitness thereof:
qjthe number of individuals assigned to the next jth subgroup, f (q)j) Is the fitness value of the best individual of the jth subgroup,
is the sum of fitness values of all subgroups, q is the individual number of the jth subgroup;
3.1.5.2 migration of individuals: selecting individuals in the sub-population according to the size of the fitness value as a migration object, namely selecting the individuals with high fitness function values, wherein the sub-population with the fitness function lower than a threshold value migrates the local optimal solution of the sub-population to the sub-population with the fitness function higher than the threshold value, and the following formula is shown:
Mqj＝{qi|f(qj)≤f(qi)}
wherein M isqjAnd i represents a subgroup number different from j, which represents a subgroup number of jth subgroup individuals having a smaller fitness value than j.
7. The improved multi-target unmanned intelligent vehicle collision avoidance driving method according to claim 5, wherein the filtering and the spatial-temporal alignment of the acquired attitude angle and position information comprises:
the driving direction of the unmanned intelligent vehicle is as follows:
in the formula vOx，vOyThe speed of the unmanned intelligent vehicle on the x axis and the y axis of the reference system is β0Setting an initial angle for the unmanned intelligent vehicle;
according to the geographic coordinates of the unmanned intelligent vehicle and the obstacle, the relative distance between the unmanned intelligent vehicle and the obstacle is calculated as follows:
in the formula, xT，yTIs the horizontal and vertical coordinates of the obstacle; x is the number ofO，yOIs the horizontal and vertical coordinates of the unmanned intelligent vehicle.
The true orientation of the obstacle relative to the unmanned intelligent vehicle is thetaT：
The true orientation of the unmanned intelligent vehicle relative to the barrier is theta0：
The phase orientation of the obstacle is αT：
Relative driving direction of obstacle relative to unmanned intelligent vehicle
8. The improved multi-target unmanned intelligent vehicle collision avoidance driving method as claimed in claim 3, wherein the shortest meeting distance is:
recent times encountered:
and B, obtaining the connection weight and the threshold of each neuron through learning of expert data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010485077.8A CN111427368A (en) | 2020-06-01 | 2020-06-01 | Improved multi-target collision-prevention driving method for unmanned intelligent vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010485077.8A CN111427368A (en) | 2020-06-01 | 2020-06-01 | Improved multi-target collision-prevention driving method for unmanned intelligent vehicle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111427368A true CN111427368A (en) | 2020-07-17 |
Family
ID=71557249
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010485077.8A Withdrawn CN111427368A (en) | 2020-06-01 | 2020-06-01 | Improved multi-target collision-prevention driving method for unmanned intelligent vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111427368A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112612266A (en) * | 2020-12-04 | 2021-04-06 | 湖南大学 | Unstructured road global path planning method and system |
| CN116107328A (en) * | 2023-02-09 | 2023-05-12 | 陕西科技大学 | Optimal automatic obstacle avoidance method for ornithopter based on improved genetic algorithm |
-
2020
- 2020-06-01 CN CN202010485077.8A patent/CN111427368A/en not_active Withdrawn
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112612266A (en) * | 2020-12-04 | 2021-04-06 | 湖南大学 | Unstructured road global path planning method and system |
| CN112612266B (en) * | 2020-12-04 | 2022-04-01 | 湖南大学 | Unstructured road global path planning method and system |
| CN116107328A (en) * | 2023-02-09 | 2023-05-12 | 陕西科技大学 | Optimal automatic obstacle avoidance method for ornithopter based on improved genetic algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111240319B (en) | Outdoor multi-robot cooperative operation system and method thereof | |
| CN106873599A (en) | Unmanned bicycle paths planning method based on ant group algorithm and polar coordinate transform | |
| CN105717929A (en) | Planning method for mixed path of mobile robot under multi-resolution barrier environment | |
| Fu et al. | On trajectory homotopy to explore and penetrate dynamically of multi-UAV | |
| Zhao et al. | A path planning method based on multi-objective cauchy mutation cat swarm optimization algorithm for navigation system of intelligent patrol car | |
| CN112577506B (en) | Automatic driving local path planning method and system | |
| CN111338356A (en) | Multi-target unmanned ship collision avoidance path planning method for improving distributed genetic algorithm | |
| Chen et al. | Tracking with UAV using tangent-plus-Lyapunov vector field guidance | |
| CN111427368A (en) | Improved multi-target collision-prevention driving method for unmanned intelligent vehicle | |
| Ming et al. | A survey of path planning algorithms for autonomous vehicles | |
| CN113391633A (en) | Urban environment-oriented mobile robot fusion path planning method | |
| Xin et al. | Coordinated motion planning of multiple robots in multi-point dynamic aggregation task | |
| CN115826586A (en) | Path planning method and system fusing global algorithm and local algorithm | |
| Wang et al. | Deep reinforcement learning-aided autonomous navigation with landmark generators | |
| Fu et al. | Collision-free and kinematically feasible path planning along a reference path for autonomous vehicle | |
| CN113296519A (en) | Mecanum wheel-based mobile robot motion planning method and system | |
| WO2022246802A1 (en) | Driving strategy determination method and apparatus, device, and vehicle | |
| CN114895676A (en) | Method for realizing high-speed running of ground automatic driving vehicle based on space intelligent system | |
| Zhong et al. | FGP-Astar Algorithm Based on Path Planning for Mobile Robots | |
| Yu | The Impact of Path Planning Model Based on Improved Ant Colony Optimization Algorithm on Green Traffic Management. | |
| CN117782097B (en) | Cloud platform-based robot inspection path planning method and system | |
| CN116088577B (en) | Unmanned cluster autonomous exploration method, unmanned cluster autonomous exploration system, electronic equipment and medium | |
| Zhao et al. | Multi-Lane Lane Change Decision and Trajectory Planning of Vehicle Based on Logic Regression Algorithm and Gaussian Probability Density Model | |
| Bi et al. | Autonomous Path Planning | |
| Murthy et al. | Human Behaviour-Based Trajectory Planning for Autonomous Overtaking Maneuver |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200717 |
|
| WW01 | Invention patent application withdrawn after publication |