CN110187707A - Planing method, device and the unmanned equipment of unmanned equipment running track - Google Patents
Planing method, device and the unmanned equipment of unmanned equipment running track Download PDFInfo
- Publication number
- CN110187707A CN110187707A CN201910462102.8A CN201910462102A CN110187707A CN 110187707 A CN110187707 A CN 110187707A CN 201910462102 A CN201910462102 A CN 201910462102A CN 110187707 A CN110187707 A CN 110187707A
- Authority
- CN
- China
- Prior art keywords
- moment
- constraint condition
- time
- planning
- sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000013528 artificial neural network Methods 0.000 claims abstract description 20
- 230000004888 barrier function Effects 0.000 claims description 57
- 230000006870 function Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000008901 benefit Effects 0.000 abstract description 5
- 230000001419 dependent effect Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 8
- 230000001133 acceleration Effects 0.000 description 4
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 210000004218 nerve net Anatomy 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
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
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
- Traffic Control Systems (AREA)
Abstract
The application provides planing method, device and the unmanned equipment of a kind of unmanned equipment running track, and a specific embodiment of the method includes: sequence corresponding first constraint condition at the time of obtaining object run planning；Obtain corresponding second constraint condition of each object time in the moment sequence in addition to last moment；It for each object time, executes following operation: being based on first constraint condition and corresponding second constraint condition of the object time, execute the first operation or the second operation；Wherein, first operation is executed based on target depth neural network trained in advance, executes second operation based on preset rules；Based on determining multiple planned trajectory points, the planned trajectory of the object run planning is generated.The embodiment can not only play the advantage of preset rules, in turn avoid the problem due to flexibility difference caused by being too dependent on preset rules, also, the result of operation planning is made to have more reasonability.
Description
Technical field
This application involves unmanned technical field, in particular to a kind of planning side of unmanned equipment running track
Method, device and unmanned equipment.
Background technique
For at present, for the operation planning of unmanned equipment, it is normally based on the experience of people, according to unmanned
The information of equipment local environment presets some rules.When carrying out operation planning, need to acquire in real time current unmanned
The information of equipment local environment, and according to the information and preset rule of current unmanned equipment local environment, it is raw
At a large amount of tracks, then therefrom choose target trajectory of the track as operation planning.But it is run through the above way
Planning, is too dependent on preset rule, flexibility is poor, reduces the reasonability of the result of operation planning.
Summary of the invention
One of in order to solve the above-mentioned technical problem, the application provides a kind of planning side of unmanned equipment running track
Method, device and unmanned equipment.
According to the embodiment of the present application in a first aspect, provide a kind of planing method of unmanned equipment running track, wrap
It includes:
Sequence corresponding first constraint condition at the time of obtaining object run planning；And
Obtain corresponding second constraint condition of each object time in the moment sequence in addition to last moment；
For each object time, executes following operation: being based on first constraint condition and the object time pair
The second constraint condition answered, execute first operation or second operation, in the determination moment sequence object time it is next
Moment corresponding planned trajectory point；Wherein, first operation is executed based on target depth neural network trained in advance, be based on
Preset rules execute second operation；
Based on determining multiple planned trajectory points, the planned trajectory of the object run planning is generated.
Optionally, first constraint condition includes:
The motion state parameters at target device end moment in the moment sequence；
The dimensional parameters for the barrier that the target device detects；
For the either objective moment, corresponding second constraint condition of the object time includes:
Motion state parameters of the target device in the object time；
Motion state parameters of the barrier in the object time；And
The motion state parameters at the barrier each moment after the object time in the moment sequence.
Optionally, first constraint condition further include: the dimensional parameters of the target device.
Optionally, before obtaining first constraint condition and obtaining second constraint condition, further includes:
The target device and the current motion state parameters of the barrier are determined respectively；
Based on the current motion state parameters of the target device and the barrier, estimate respectively the target device and
The motion state parameters at the barrier each moment in the moment sequence.
It is optionally, described to execute the first operation or the second operation, comprising:
According to preset execution probability, the first operation or the second operation are executed；Wherein, the execution probability includes holding
First probability of row first operation and the second probability for executing second operation；First probability is general with described second
The sum of rate is 1.
Optionally, it for the either objective moment, is held in the following way based on target depth neural network trained in advance
Row first operation:
Standardized normal distribution random number is taken, as stochastic variable；
First constraint condition, corresponding second constraint condition of the object time and the stochastic variable are input to institute
Target depth neural network is stated, the result of the target depth neural network output is obtained.
Optionally, described based on determining multiple planned trajectory points, generate the planned trajectory of the object run planning, packet
It includes:
A plurality of alternate trajectory is obtained by the way of polynomial curve interpolation based on determining multiple planned trajectory points；
Using the cost value of every alternate trajectory of preset cost function calculation；
Choose the planned trajectory that the smallest alternate trajectory of cost value is planned as object run.
According to the second aspect of the embodiment of the present application, a kind of device for planning of unmanned equipment running track is provided, is wrapped
It includes:
First obtains module, sequence corresponding first constraint condition at the time of for obtaining object run planning；And
Second obtains module, for obtaining each object time in the moment sequence in addition to last moment corresponding the
Two constraint conditions；
Execution module executes following operation for being directed to each object time: based on first constraint condition and
Corresponding second constraint condition of the object time executes the first operation or the second operation, with the mesh in the determination moment sequence
Mark the corresponding planned trajectory point of subsequent time at moment；Wherein, based on described in target depth neural network execution trained in advance
First operation executes second operation based on preset rules；
Generation module, for generating the planned trajectory of the object run planning based on determining multiple planned trajectory points.
According to the third aspect of the embodiment of the present application, a kind of computer readable storage medium is provided, the storage medium is deposited
Computer program is contained, the computer program realizes side described in any one of above-mentioned first aspect when being executed by processor
Method.
According to the fourth aspect of the embodiment of the present application, a kind of unmanned equipment is provided, including memory, processor and deposit
On a memory and the computer program that can run on a processor, when processor execution described program, realizes above-mentioned for storage
Method described in any one of one side.
The technical solution that embodiments herein provides can include the following benefits:
The planning method and device for the unmanned equipment running track that embodiments herein provides, by obtaining target
Sequence corresponding first constraint condition at the time of operation planning, when obtaining each target in the moment sequence in addition to last moment
Carve corresponding second constraint condition.For above-mentioned each object time, following operation is executed: based on first constraint condition and being somebody's turn to do
Corresponding second constraint condition of object time executes the first operation or the second operation, when determining the target in the moment sequence
The corresponding planned trajectory point of the subsequent time at quarter.Wherein, the first operation is executed based on target depth neural network trained in advance,
The second operation is executed based on preset rules.Based on determining multiple planned trajectory points, the planned trajectory of object run planning is generated.
It is operated due to the present embodiment by the first operation executed based on target depth neural network and based on preset rules execute second
It combines, when carrying out operation planning, the advantage of preset rules can not only be played, meanwhile, it in turn avoids due to excessively relying on
The problem of flexibility difference caused by preset rules.Also, the present embodiment is when determining planned trajectory point, it is contemplated that moment sequence
Corresponding second constraint condition of each object time in corresponding first constraint condition and moment sequence is arranged, so that operation rule
The result drawn has more reasonability.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not
The application can be limited.
Detailed description of the invention
The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the application
Example, and together with specification it is used to explain the principle of the application.
Fig. 1 is a kind of the application planing method of unmanned equipment running track shown according to an exemplary embodiment
Flow chart；
Fig. 2 is the planning side of the unmanned equipment running track of the application another kind shown according to an exemplary embodiment
The flow chart of method；
Fig. 3 is the planning side of the unmanned equipment running track of the application another kind shown according to an exemplary embodiment
The flow chart of method；
Fig. 4 is a kind of the application device for planning of unmanned equipment running track shown according to an exemplary embodiment
Block diagram；
Fig. 5 is the planning dress of the unmanned equipment running track of the application another kind shown according to an exemplary embodiment
The block diagram set；
Fig. 6 is the planning dress of the unmanned equipment running track of the application another kind shown according to an exemplary embodiment
The block diagram set；
Fig. 7 is a kind of the application structural schematic diagram of unmanned equipment shown according to an exemplary embodiment.
Specific embodiment
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent all embodiments consistent with the application.On the contrary, they be only with it is such as appended
The example of the consistent device and method of some aspects be described in detail in claims, the application.
It is only to be not intended to be limiting the application merely for for the purpose of describing particular embodiments in term used in this application.
It is also intended in the application and the "an" of singular used in the attached claims, " described " and "the" including majority
Form, unless the context clearly indicates other meaning.It is also understood that term "and/or" used herein refers to and wraps
It may be combined containing one or more associated any or all of project listed.
It will be appreciated that though various information, but this may be described using term first, second, third, etc. in the application
A little information should not necessarily be limited by these terms.These terms are only used to for same type of information being distinguished from each other out.For example, not departing from
In the case where the application range, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as
One information.Depending on context, word as used in this " if " can be construed to " ... when " or " when ...
When " or " in response to determination ".
As shown in Figure 1, Fig. 1 is a kind of planning of unmanned equipment running track shown according to an exemplary embodiment
The flow chart of method, this method can be applied in unmanned equipment.It will be understood by those skilled in the art that this is unmanned
Equipment can include but is not limited to unmanned vehicle, unattended robot, unmanned plane, unmanned boat etc..This method includes following step
It is rapid:
In a step 101, sequence corresponding first constraint condition at the time of obtaining object run planning.
In general, unmanned equipment at runtime, needs the information previously according to its current environment, planning is not
The operating path (e.g., planning since current time, the operating path in N seconds) come in preset period of time, as operation planning
Planned trajectory.Driving Decision-making is carried out according still further to the planned trajectory of operation planning.It specifically, usually can first ring according to locating for it
The information in border predicts tracing point corresponding to multiple predetermined times in the following preset period of time, then, then above-mentioned tracing point is connected
At track, the planned trajectory of operation planning is obtained.
In the present embodiment, object run is planned to the operation planning currently carried out for unmanned equipment, can obtain
To the operating path in the following preset period of time, which be can be since current time to the period at N seconds ends.Its
In, N can be any reasonable positive integer, and the application does not limit the specific value aspect of N.Sequence at the time of object run is planned
Be classified as the sequence that a series of predetermined times in the future preset period of time are constituted, the moment sequence may include initial time (i.e.
Current time), multiple intermediate times and end moment (i.e. at the time of N seconds end).
In the present embodiment, corresponding first constraint condition of sequence can characterize the moment sequence at the time of object run is planned
Arrange the influence factor planned object run.When first constraint condition may include that target device is last a in the moment sequence
The dimensional parameters for the barrier that the motion state parameters and target device at quarter detect.
Wherein, target device is the unmanned equipment for carrying out object run planning.The motion state of target device is joined
The relative displacement that can include but is not limited to target device along road longitudinal direction, speed, acceleration are counted, target device is along road
Carriageway type locating for the relative displacement of transverse direction, speed, acceleration and target device etc..What target device detected
The dimensional parameters of barrier are each barrier in one or more barriers that target device is currently detected by sensor
Dimensional parameters.The dimensional parameters of barrier can include but is not limited to the length parameter of barrier, width parameter and height
Parameter etc..
Optionally, movement of first constraint condition in addition to end moment that may include target device in the moment sequence
Other than the dimensional parameters for the barrier that state parameter and target device detect, the ruler of target device can further include
Very little parameter.Wherein, the dimensional parameters of target device can include but is not limited to the length parameter of target device, width parameter and
Height parameter etc..Certain influence is generated since the dimensional parameters of target device can plan object run, carrying out mesh
When marking operation planning, if it is contemplated that the influence that the dimensional parameters of target device plan object run, can to run
The result of planning is more reasonable.
In a step 102, corresponding second constraint of each object time in the moment sequence in addition to last moment is obtained
Condition.
In the present embodiment, object time is any time in above-mentioned moment sequence, in addition to last moment, can be obtained
Take corresponding second constraint condition of each object time.For any one object time, the object time corresponding second is about
Beam condition can characterize the influence factor that the object time plans object run.Corresponding second constraint condition of the object time
It may include motion state parameters of the target device in the object time, the barrier that target device detects is in the object time
Each moment in the moment sequence after the object time of motion state parameters and the barrier motion state ginseng
Number.
For example, above-mentioned moment sequence may include initial time A, intermediate time B, C, D, E, and end moment F.The then moment
A, B, C, D, E are object time, each moment corresponding second constraint condition in available moment A, B, C, D, E.Wherein,
Corresponding second constraint condition of moment A may include motion state parameters of the target device in moment A, and barrier is moment A's
In moment B, C, D, E, F, (moment B, C, D, E, F are moment A in the moment sequence respectively for motion state parameters and the barrier
Each moment later) motion state parameters.Corresponding second constraint condition of moment B may include target device in moment B
Motion state parameters, barrier is in the motion state parameters of moment B and the barrier respectively in the movement of moment C, D, E, F
State parameter.And so on ... ... corresponding second constraint condition of moment E may include movement shape of the target device in moment E
State parameter, barrier moment E motion state parameters and the barrier moment F motion state parameters.
Wherein, the motion state parameters of barrier can include but is not limited to barrier along road longitudinal direction speed,
Acceleration, speed, acceleration of the barrier along road transverse direction, the distance between barrier and target device, barrier with
Carriageway type locating for relative angle and barrier between target device etc..
It should be noted that target can be obtained first before obtaining the first constraint condition and the second constraint condition
The current motion state parameters of equipment and the current motion state parameters of barrier.Then, target device and barrier are based on
Current motion state parameters, the motion state ginseng at each moment of estimating target device and barrier respectively in moment sequence
Number, to determine the first constraint condition and the second constraint condition.For example, above-mentioned motion state ginseng can be estimated using preset algorithm
Number can also estimate above-mentioned motion state parameters using the method for machine learning.It is appreciated that as known in the art and incites somebody to action
It can be applied to the application any method that can estimate above-mentioned motion state parameters for being likely to occur, the application is to estimating
It is not limited in terms of the concrete mode of above-mentioned motion state parameters.
In step 103, it for above-mentioned each object time, executes following operation: based on first constraint condition and being somebody's turn to do
Corresponding second constraint condition of object time executes the first operation or the second operation, when determining the target in the moment sequence
The corresponding planned trajectory point of the subsequent time at quarter.
It in the present embodiment, can be based on each of corresponding first constraint condition of the moment sequence and the moment sequence
Corresponding second constraint condition of object time, determines the corresponding planned trajectory of the subsequent time of each object time in moment sequence
Point obtains multiple planned trajectory points.It is directed to above-mentioned each object time, is based on first constraint condition and the object time pair
The second constraint condition answered determines the corresponding planned trajectory point of the subsequent time of the object time in the moment sequence.
For example, moment sequence may include initial time A, intermediate time B, C, D, E, and end moment F.It can then be based on
Corresponding first constraint condition of the moment sequence and corresponding second constraint condition of moment A, determine the corresponding planning rail of moment B
Mark point.Based on corresponding first constraint condition of the moment sequence and corresponding second constraint condition of moment B, C pairs of the moment is determined
The planned trajectory point answered.And so on ... ... based on corresponding first constraint condition of the moment sequence and moment E corresponding the
Two constraint conditions determine the corresponding planned trajectory point of moment F.
Specifically, for any one object time, when can be determined as follows the target in moment sequence
The corresponding planned trajectory point of the subsequent time at quarter: based on the corresponding second constraint item of first constraint condition and the object time
Part, by executing the first corresponding planned trajectory point of subsequent time for operating the object time in determining moment sequence, Huo Zhetong
It crosses and executes the second corresponding planned trajectory point of subsequent time for operating the object time in determining moment sequence.Wherein it is possible to base
The first operation is executed in target depth neural network trained in advance, executes the second operation based on preset rules.
In the present embodiment, any operation in the first operation and the second operation can be executed at random, it can also be according to pre-
If rule, compartment of terrain execute first operation and second operation, can also according to preset probability execute first operation and
Any operation in second operation.It is appreciated that the first operation or the second operation can be executed using any reasonable manner, this
Application to not limiting in this respect.
It optionally, can be in the following way based on target depth nerve net trained in advance for the either objective moment
Network executes the first operation: taking standardized normal distribution random number, constrains item as stochastic variable, and by moment sequence corresponding first
Corresponding second constraint condition of part, the object time and above-mentioned stochastic variable are input to target depth neural network, obtain target
The result of deep neural network output.Executing the first operation through the above way makes obtained rule due to introducing stochastic variable
It is more accurate to draw tracing point.
It is alternatively possible to rule of thumb set preset rules, rule function is obtained, the independent variable of the rule function can be
Constraint condition, variable can be planned trajectory point.For the either objective moment, can be held in the following way based on preset rules
Row second operates: using corresponding first constraint condition of moment sequence and corresponding second constraint condition of the object time as certainly
Variable is input to above-mentioned rule function, obtains the corresponding planned trajectory of subsequent time of the object time of rule function output
Point.
At step 104, based on determining multiple planned trajectory points, the planned trajectory of object run planning is generated.
In the present embodiment, object run can be generated according to moment each in moment sequence corresponding planned trajectory point
The planned trajectory of planning.Specifically, it is possible, firstly, to be based on multiple planned trajectory points, by the way of polynomial curve interpolation,
Obtain a plurality of alternate trajectory.Then, using the cost value of every alternate trajectory of preset cost function calculation, and cost value is chosen
The planned trajectory that the smallest alternate trajectory is planned as object run.It is appreciated that other any reasonable sides can also be passed through
Formula generates the planned trajectory of object run planning.It is as known in the art and what is be likely to occur in the future any can generate target
The method of the planned trajectory of operation planning can be applied to the application, and the application is to the planned trajectory for generating object run planning
Concrete mode in terms of do not limit.
Although should be noted that in the embodiment of above-mentioned Fig. 1, the operation of the application method is described with particular order,
It is that this does not require that or implies must execute these operations in this particular order, or have to carry out shown in whole
Operation is just able to achieve desired result.On the contrary, the step of describing in flow chart can change and execute sequence.For example, step 101 can
It to be executed before step 102, can also execute, can also be performed simultaneously with step 102 after step 102.Additionally or
It is alternatively possible to omit certain steps, multiple steps are merged into a step and are executed, and/or a step is decomposed into more
A step executes.
The planing method of the unmanned equipment running track provided by the above embodiment of the application, by obtaining target fortune
Sequence corresponding first constraint condition at the time of professional etiquette is drawn obtains each object time in the moment sequence in addition to last moment
Corresponding second constraint condition.For above-mentioned each object time, executes following operation: being based on first constraint condition and the mesh
Moment corresponding second constraint condition is marked, the first operation or the second operation are executed, to determine the object time in the moment sequence
The corresponding planned trajectory point of subsequent time.Wherein, the first operation, base are executed based on target depth neural network trained in advance
The second operation is executed in preset rules.Based on determining multiple planned trajectory points, the planned trajectory of object run planning is generated.By
In the second operation phase that the present embodiment is operated executed based on target depth neural network first and executed based on preset rules
In conjunction with, when carrying out operation planning, the advantage of preset rules can not only be played, meanwhile, it in turn avoids due to being too dependent on
The problem of flexibility difference caused by preset rules.Also, the present embodiment is when determining planned trajectory point, it is contemplated that moment sequence
Corresponding second constraint condition of each object time in corresponding first constraint condition and moment sequence, so that operation planning
Result have more reasonability.
As shown in Fig. 2, the planning of the unmanned equipment running track of Fig. 2 another kind shown according to an exemplary embodiment
The flow chart of method, This embodiment describes the determination process of target device and the motion state parameters of barrier, this method can
To be applied in unmanned equipment, comprising the following steps:
In step 201, target device and the current motion state parameters of barrier are determined respectively.
In the present embodiment, can by devices such as the sensor being installed on target device or inertial navigation systems,
Determine the current motion state parameters of target device.By devices such as the radar being installed on target device or sensors, really
Determine the current motion state parameters of barrier.It is appreciated that the movement shape that the application is current to determining target device and barrier
It is not limited in terms of the concrete mode of state parameter.
In step 202, the motion state parameters current based on target device and barrier, estimate respectively target device and
The motion state parameters at barrier each moment in sequence at the time of object run is planned.
In the present embodiment, above-mentioned motion state parameters can be estimated using preset algorithm, machine learning can also be used
Method estimate above-mentioned motion state parameters.It is appreciated that it is as known in the art and in the future be likely to occur it is any can
The method for estimating above-mentioned motion state parameters can be applied to the application, and the application is to the tool for estimating above-mentioned motion state parameters
It is not limited in terms of body mode.
In step 203, sequence corresponding first constraint condition at the time of obtaining object run planning, the first constraint item
Part includes the motion state parameters at target device end in the moment sequence a moment and the barrier that target device detects
Dimensional parameters.
In step 204, corresponding second constraint of each object time in the moment sequence in addition to last moment is obtained
Condition, corresponding second constraint condition of any one object time include the movement of target device and barrier in the object time
The motion state parameters at each moment of state parameter and the barrier in moment sequence after the object time.
In step 205, it for above-mentioned each object time, executes following operation: based on first constraint condition and being somebody's turn to do
Corresponding second constraint condition of object time executes the first operation or the second operation, when determining the target in the moment sequence
The corresponding planned trajectory point of the subsequent time at quarter.
In step 206, based on determining multiple planned trajectory points, the planned trajectory of object run planning is generated.
It should be noted that no longer going to live in the household of one's in-laws on getting married in above-mentioned Fig. 2 embodiment for the step identical with Fig. 1 embodiment
It states, related content can be found in Fig. 1 embodiment.
The planing method of the unmanned equipment running track provided by the above embodiment of the application determines that target is set respectively
The standby and current motion state parameters of barrier, and the motion state parameters current based on target device and barrier, it is pre- respectively
Estimate target device and barrier object run planning at the time of sequence in each moment motion state parameters.Obtain target fortune
Sequence corresponding first constraint condition at the time of professional etiquette is drawn, first constraint condition include that target device is last in the moment sequence
The dimensional parameters of the motion state parameters at a moment and the barrier detected.It obtains in the moment sequence in addition to last moment
Corresponding second constraint condition of each object time, corresponding second constraint condition of any one object time includes that target is set
Standby and barrier is in the motion state parameters of the object time and the barrier in moment sequence after the object time
The motion state parameters at each moment.For above-mentioned each object time, execute following operation: based on first constraint condition and
Corresponding second constraint condition of the object time executes the first operation or the second operation, to determine the target in the moment sequence
The corresponding planned trajectory point of the subsequent time at moment.Based on determining multiple planned trajectory points, the rule of object run planning are generated
Draw track.To further improve the reasonability of the result of operation planning.
In some optional embodiments, the first operation or the second operation can be executed in the following way: according to preparatory
The execution probability of setting executes the first operation or the second operation.Wherein, executing probability includes the first probability for executing the first operation
The second probability operated with execution second, the sum of the first probability and the second probability are 1.
In the present embodiment, it can in advance for the first operation and the second operation, be set separately corresponding according to the actual situation
Execution probability, the execution probability include execute first operation the first probability and execute second operation the second probability.For example,
The second probability set is operated as n for second as m for the first probability that the first operation is set, and m+n=1.It can also be with
Rule of thumb, the size of m and n is constantly adjusted.When carrying out operation planning, can be executed according to preset execution probability
First operation or the second operation.
Since the present embodiment is according to preset execution probability, the first operation or the second operation are executed, is efficiently controlled
Therefore the execution ratio of first operation and the second operation facilitates the advantage for playing preset rules, and avoid due to excessive
Problem dependent on flexibility difference caused by preset rules.
As shown in figure 3, the planning of the unmanned equipment running track of Fig. 3 another kind shown according to an exemplary embodiment
The process for generating the planned trajectory of object run planning is described in detail in the flow chart of method, the embodiment, and this method can answer
For in unmanned equipment, comprising the following steps:
In step 301, sequence corresponding first constraint condition at the time of obtaining object run planning.
In step 302, corresponding second constraint of each object time in the moment sequence in addition to last moment is obtained
Condition.
In step 303, it for above-mentioned each object time, executes following operation: based on first constraint condition and being somebody's turn to do
Corresponding second constraint condition of object time executes the first operation or the second operation, when determining the target in the moment sequence
The corresponding planned trajectory point of the subsequent time at quarter.
In step 304, it is obtained more by the way of polynomial curve interpolation based on determining multiple planned trajectory points
Alternate trajectory.
It is alternatively possible to obtain a plurality of alternate trajectory by the way of 5 order polynomial curve interpolations.
In step 305, using the cost value of every alternate trajectory of preset cost function calculation.
In the present embodiment, preset cost function can be any reasonable cost function, and the application is to cost function
Concrete form in terms of do not limit.
Within step 306, the planned trajectory that the smallest alternate trajectory of cost value is planned as object run is chosen.
It should be noted that for the step identical with Fig. 1 and Fig. 2 embodiment, in above-mentioned Fig. 3 embodiment no longer into
Row repeats, and related content can be found in Fig. 1 and Fig. 2 embodiment.
The planing method of the unmanned equipment running track provided by the above embodiment of the application, by based on determining
Multiple planned trajectory points obtain a plurality of alternate trajectory by the way of polynomial curve interpolation, using preset cost function meter
The cost value of every alternate trajectory is calculated, and chooses the planned trajectory that the smallest alternate trajectory of cost value is planned as object run.
To improve the completeness of planned trajectory, so that planned trajectory is more smooth, more meet mankind's driving behavior habit.
Corresponding with the planing method embodiment of aforementioned unmanned equipment running track, present invention also provides nobody to drive
Sail the embodiment of the device for planning of equipment running track.
As shown in figure 4, Fig. 4 is a kind of the application unmanned equipment running track shown according to an exemplary embodiment
Device for planning block diagram, the apparatus may include: first obtains module 401, and second obtains module 402, execution module 403 and raw
At module 404.
Wherein, first module 401 is obtained, sequence corresponding first constraint item at the time of for obtaining object run planning
Part.
Second obtains module 402, corresponds to for obtaining each object time in above-mentioned moment sequence in addition to last moment
The second constraint condition.
Execution module 403 executes following operation: being based on moment sequence corresponding first for being directed to each object time
Constraint condition and corresponding second constraint condition of the object time execute the first operation or the second operation, with the determination above-mentioned moment
The corresponding planned trajectory point of the subsequent time of the object time in sequence.Wherein, based on target depth nerve net trained in advance
Network executes the first operation, executes the second operation based on preset rules.
Generation module 404, for generating the planned trajectory of object run planning based on determining multiple planned trajectory points.
In some optional embodiments, corresponding first constraint condition of sequence be can wrap at the time of object run is planned
It includes: the motion state parameters at target device end moment in above-mentioned moment sequence；The ruler for the barrier that target device detects
Very little parameter.
For the either objective moment, corresponding second constraint condition of the object time includes: target device in the target
The motion state parameters at quarter；Motion state parameters of the barrier in the object time；And the barrier is in above-mentioned moment sequence
In column after the object time each moment motion state parameters.
In other optional embodiments, sequence corresponding first constraint condition can be at the time of object run is planned
It include: the dimensional parameters of target device.
As shown in figure 5, Fig. 5 is the unmanned equipment operation rail of the application another kind shown according to an exemplary embodiment
The device for planning block diagram of mark, on the basis of aforementioned embodiment illustrated in fig. 4, which can further include the embodiment:
Determining module 405 and estimate module 406.
Wherein it is determined that module 405, for determining target device and the current motion state parameters of barrier respectively.
Module 406 is estimated, for the motion state parameters current based on target device and barrier, target is estimated respectively and sets
Standby and barrier each moment in above-mentioned moment sequence motion state parameters.
In other optional embodiments, execution module 403 is configured for: according to preset execution probability,
Execute the first operation or the second operation.Wherein, executing probability includes the second operation of the first probability and execution for executing the first operation
The second probability, the sum of the first probability and the second probability be 1.
In other optional embodiments, for the either objective moment, execution module 403 can base in the following way
First operation is executed in target depth neural network trained in advance: being taken standardized normal distribution random number, is become as random
First constraint condition, corresponding second constraint condition of the object time and above-mentioned stochastic variable are input to target depth mind by amount
Through network, the result of target depth neural network output is obtained.
As shown in fig. 6, Fig. 6 is the unmanned equipment operation rail of the application another kind shown according to an exemplary embodiment
The device for planning block diagram of mark, for the embodiment on the basis of aforementioned embodiment illustrated in fig. 4, generation module 404 may include: determination
Submodule 601, computational submodule 602 and selection submodule 603.
Wherein it is determined that submodule 601, for based on determining multiple planned trajectory points, using polynomial curve interpolation
Mode obtains a plurality of alternate trajectory.
Computational submodule 602, for the cost value using preset every alternate trajectory of cost function calculation.
Choose submodule 603, the planned trajectory planned for choosing the smallest alternate trajectory of cost value as object run.
It should be appreciated that above-mentioned apparatus can be set in advance in unmanned equipment, the modes such as downloading can also be passed through
It is loaded into unmanned equipment.Corresponding module in above-mentioned apparatus can cooperate with the module in unmanned equipment with
Realize the programme of unmanned equipment running track.
For device embodiment, since it corresponds essentially to embodiment of the method, so related place is referring to method reality
Apply the part explanation of example.The apparatus embodiments described above are merely exemplary, wherein described be used as separation unit
The unit of explanation may or may not be physically separated, and component shown as a unit can be or can also be with
It is not physical unit, it can it is in one place, or may be distributed over multiple network units.It can be according to actual
The purpose for needing to select some or all of the modules therein to realize application scheme.Those of ordinary skill in the art are not paying
Out in the case where creative work, it can understand and implement.
The embodiment of the present application also provides a kind of computer readable storage medium, which is stored with computer journey
Sequence, computer program can be used for executing the planning for the unmanned equipment running track that above-mentioned Fig. 1 is provided to Fig. 3 any embodiment
Method.
Corresponding to the planing method of above-mentioned unmanned equipment running track, the embodiment of the present application also proposed Fig. 7 institute
The schematic configuration diagram of the unmanned equipment for the exemplary embodiment according to the application shown.Referring to FIG. 7, in hardware layer
Face, which includes processor, internal bus, network interface, memory and nonvolatile memory, may be used also certainly
It can include hardware required for other business.Processor reads corresponding computer program to memory from nonvolatile memory
In then run, the device for planning of unmanned equipment running track is formed on logic level.Certainly, in addition to software realization side
Except formula, other implementations, such as logical device or the mode of software and hardware combining etc. is not precluded in the application, that is,
It says that the executing subject of following process flow is not limited to each logic unit, is also possible to hardware or logical device.
Those skilled in the art after considering the specification and implementing the invention disclosed here, will readily occur to its of the application
Its embodiment.This application is intended to cover any variations, uses, or adaptations of the application, these modifications, purposes or
Person's adaptive change follows the general principle of the application and including the undocumented common knowledge in the art of the application
Or conventional techniques.The description and examples are only to be considered as illustrative, and the true scope and spirit of the application are by following
Claim is pointed out.
It should be understood that the application is not limited to the precise structure that has been described above and shown in the drawings, and
And various modifications and changes may be made without departing from the scope thereof.Scope of the present application is only limited by the accompanying claims.
Claims (10)
1. a kind of planing method of unmanned equipment running track, which is characterized in that the described method includes:
Sequence corresponding first constraint condition at the time of obtaining object run planning；And
Obtain corresponding second constraint condition of each object time in the moment sequence in addition to last moment；
For each object time, following operation is executed: corresponding based on first constraint condition and the object time
Second constraint condition executes the first operation or the second operation, with the subsequent time of the object time in the determination moment sequence
Corresponding planned trajectory point；Wherein, first operation is executed based on target depth neural network trained in advance, based on default
Rule executes second operation；
Based on determining multiple planned trajectory points, the planned trajectory of the object run planning is generated.
2. the method according to claim 1, wherein first constraint condition includes:
The motion state parameters at target device end moment in the moment sequence；
The dimensional parameters for the barrier that the target device detects；
For the either objective moment, corresponding second constraint condition of the object time includes:
Motion state parameters of the target device in the object time；
Motion state parameters of the barrier in the object time；And
The motion state parameters at the barrier each moment after the object time in the moment sequence.
3. according to the method described in claim 2, it is characterized in that, first constraint condition further include: the target device
Dimensional parameters.
4. according to the method described in claim 2, it is characterized in that, obtaining first constraint condition and obtaining described the
Before two constraint conditions, further includes:
The target device and the current motion state parameters of the barrier are determined respectively；
Based on the current motion state parameters of the target device and the barrier, the target device and described is estimated respectively
The motion state parameters at barrier each moment in the moment sequence.
5. method according to any one of claims 1-4, which is characterized in that it is described to execute the first operation or the second operation,
Include:
According to preset execution probability, the first operation or the second operation are executed；Wherein, the execution probability includes executing institute
It states the first probability of the first operation and executes the second probability of second operation；First probability and second probability it
Be 1.
6. method according to any one of claims 1-4, which is characterized in that be directed to the either objective moment, pass through such as lower section
Formula executes first operation based on target depth neural network trained in advance:
Standardized normal distribution random number is taken, as stochastic variable；
First constraint condition, corresponding second constraint condition of the object time and the stochastic variable are input to the mesh
Deep neural network is marked, the result of the target depth neural network output is obtained.
7. method according to any one of claims 1-4, which is characterized in that described based on determining multiple planned trajectories
Point generates the planned trajectory of the object run planning, comprising:
A plurality of alternate trajectory is obtained by the way of polynomial curve interpolation based on determining multiple planned trajectory points；
Using the cost value of every alternate trajectory of preset cost function calculation；
Choose the planned trajectory that the smallest alternate trajectory of cost value is planned as object run.
8. a kind of device for planning of unmanned equipment running track, which is characterized in that described device includes:
First obtains module, sequence corresponding first constraint condition at the time of for obtaining object run planning；And
Second obtains module, for obtaining each object time corresponding second in the moment sequence in addition to last moment about
Beam condition；
Execution module executes following operation: being based on first constraint condition and the mesh for being directed to each object time
Moment corresponding second constraint condition is marked, the first operation or the second operation are executed, in the determination moment sequence when target
The corresponding planned trajectory point of the subsequent time at quarter；Wherein, described first is executed based on target depth neural network trained in advance
Operation executes second operation based on preset rules；
Generation module, for generating the planned trajectory of the object run planning based on determining multiple planned trajectory points.
9. a kind of computer readable storage medium, which is characterized in that the storage medium is stored with computer program, the calculating
Method described in any one of the claims 1-7 is realized when machine program is executed by processor.
10. a kind of unmanned equipment, can run on a memory and on a processor including memory, processor and storage
Computer program, which is characterized in that the processor is realized described in any one of the claims 1-7 when executing described program
Method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910462102.8A CN110187707B (en) | 2019-05-30 | 2019-05-30 | Unmanned equipment running track planning method and device and unmanned equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910462102.8A CN110187707B (en) | 2019-05-30 | 2019-05-30 | Unmanned equipment running track planning method and device and unmanned equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110187707A true CN110187707A (en) | 2019-08-30 |
| CN110187707B CN110187707B (en) | 2022-05-10 |
Family
ID=67718883
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910462102.8A Active CN110187707B (en) | 2019-05-30 | 2019-05-30 | Unmanned equipment running track planning method and device and unmanned equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110187707B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079721A (en) * | 2020-03-23 | 2020-04-28 | 北京三快在线科技有限公司 | Method and device for predicting track of obstacle |
| CN111123927A (en) * | 2019-12-20 | 2020-05-08 | 北京三快在线科技有限公司 | Trajectory planning method and device, automatic driving equipment and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107357168A (en) * | 2017-06-01 | 2017-11-17 | 同济大学 | A kind of unmanned vehicle barrier-avoiding method based on Chance-constrained Model PREDICTIVE CONTROL |
| US20170364083A1 (en) * | 2016-06-21 | 2017-12-21 | Baidu Online Network Technology (Beijing) Co., Ltd. | Local trajectory planning method and apparatus for smart vehicles |
| CN108068815A (en) * | 2016-11-14 | 2018-05-25 | 百度（美国）有限责任公司 | System is improved for the decision-making based on planning feedback of automatic driving vehicle |
| JP2018177074A (en) * | 2017-04-18 | 2018-11-15 | 国立大学法人 東京大学 | Autonomous type underwater robot and control method for the same |
| CN108983813A (en) * | 2018-07-27 | 2018-12-11 | 长春草莓科技有限公司 | A kind of unmanned plane during flying preventing collision method and system |
| US20180374359A1 (en) * | 2017-06-22 | 2018-12-27 | Bakhi.com Times Technology (Beijing) Co., Ltd. | Evaluation framework for predicted trajectories in autonomous driving vehicle traffic prediction |
| US20190086925A1 (en) * | 2017-09-18 | 2019-03-21 | Baidu Usa Llc | Path optimization based on constrained smoothing spline for autonomous driving vehicles |
| CN109684702A (en) * | 2018-12-17 | 2019-04-26 | 清华大学 | Driving Risk Identification based on trajectory predictions |
| CN109782779A (en) * | 2019-03-19 | 2019-05-21 | 电子科技大学 | AUV paths planning method under ocean current environment based on population meta-heuristic algorithms |
-
2019
- 2019-05-30 CN CN201910462102.8A patent/CN110187707B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170364083A1 (en) * | 2016-06-21 | 2017-12-21 | Baidu Online Network Technology (Beijing) Co., Ltd. | Local trajectory planning method and apparatus for smart vehicles |
| CN108068815A (en) * | 2016-11-14 | 2018-05-25 | 百度（美国）有限责任公司 | System is improved for the decision-making based on planning feedback of automatic driving vehicle |
| JP2018177074A (en) * | 2017-04-18 | 2018-11-15 | 国立大学法人 東京大学 | Autonomous type underwater robot and control method for the same |
| CN107357168A (en) * | 2017-06-01 | 2017-11-17 | 同济大学 | A kind of unmanned vehicle barrier-avoiding method based on Chance-constrained Model PREDICTIVE CONTROL |
| US20180374359A1 (en) * | 2017-06-22 | 2018-12-27 | Bakhi.com Times Technology (Beijing) Co., Ltd. | Evaluation framework for predicted trajectories in autonomous driving vehicle traffic prediction |
| US20190086925A1 (en) * | 2017-09-18 | 2019-03-21 | Baidu Usa Llc | Path optimization based on constrained smoothing spline for autonomous driving vehicles |
| CN108983813A (en) * | 2018-07-27 | 2018-12-11 | 长春草莓科技有限公司 | A kind of unmanned plane during flying preventing collision method and system |
| CN109684702A (en) * | 2018-12-17 | 2019-04-26 | 清华大学 | Driving Risk Identification based on trajectory predictions |
| CN109782779A (en) * | 2019-03-19 | 2019-05-21 | 电子科技大学 | AUV paths planning method under ocean current environment based on population meta-heuristic algorithms |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111123927A (en) * | 2019-12-20 | 2020-05-08 | 北京三快在线科技有限公司 | Trajectory planning method and device, automatic driving equipment and storage medium |
| CN111079721A (en) * | 2020-03-23 | 2020-04-28 | 北京三快在线科技有限公司 | Method and device for predicting track of obstacle |
| WO2021190484A1 (en) * | 2020-03-23 | 2021-09-30 | 北京三快在线科技有限公司 | Trajectory prediction method and apparatus for obstacle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110187707B (en) | 2022-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110136481B (en) | Parking strategy based on deep reinforcement learning | |
| GB2595620A (en) | Combined prediction and path planning for autonomous objects using neural networks | |
| CN106104656B (en) | Map information generating systems, method and program | |
| Bakker et al. | Traffic light control by multiagent reinforcement learning systems | |
| CN110018689A (en) | A kind of more virtual target point overall situation active path planning algorithms based on dynamic window | |
| CN110189547A (en) | A kind of obstacle detection method, device and vehicle | |
| Kallmann et al. | Geometric and discrete path planning for interactive virtual worlds | |
| CN110646009A (en) | DQN-based vehicle automatic driving path planning method and device | |
| CN109213153B (en) | Automatic vehicle driving method and electronic equipment | |
| Cheng et al. | CoSIGN: A parallel algorithm for coordinated traffic signal control | |
| CN110187707A (en) | Planing method, device and the unmanned equipment of unmanned equipment running track | |
| Gupta et al. | Sensor scheduling algorithms requiring limited computation [vehicle sonar range-finder example] | |
| Shiroshita et al. | Behaviorally diverse traffic simulation via reinforcement learning | |
| CN107525507B (en) | The determination method and device of yaw | |
| Wahl et al. | Approximate dynamic programming methods applied to far trajectory planning in optimal control | |
| CN113391633A (en) | Urban environment-oriented mobile robot fusion path planning method | |
| Liu et al. | Impact of sharing driving attitude information: A quantitative study on lane changing | |
| Zhao et al. | Hybrid navigation method for multiple robots facing dynamic obstacles | |
| JP7058761B2 (en) | Mobile control device, mobile control learning device, and mobile control method | |
| CN109492835B (en) | Method for determining vehicle control information, method for training model and related device | |
| CN115373384A (en) | Vehicle dynamic path planning method and system based on improved RRT | |
| CN108121347A (en) | For the method, apparatus and electronic equipment of control device movement | |
| Evans et al. | Reward signal design for autonomous racing | |
| Randlov | Learning macro-actions in reinforcement learning | |
| WO2021008798A1 (en) | Training of a convolutional neural network |
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 |