CN107632308A - A kind of vehicle front barrier profile testing method based on recurrence superposition algorithm - Google Patents
A kind of vehicle front barrier profile testing method based on recurrence superposition algorithm Download PDFInfo
- Publication number
- CN107632308A CN107632308A CN201710733511.8A CN201710733511A CN107632308A CN 107632308 A CN107632308 A CN 107632308A CN 201710733511 A CN201710733511 A CN 201710733511A CN 107632308 A CN107632308 A CN 107632308A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- value
- mover
- scan data
- 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
Landscapes
- Traffic Control Systems (AREA)
- Optical Radar Systems And Details Thereof (AREA)
Abstract
The invention discloses a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm, this method includes：The geometrical relationship on radar and ground calculates barrier profile elevations h；History scans the coordinate matching with Current Scan；Consider that there is radar light beam the characteristics of normal distribution to introduce probability density function；Realize the quasi-continuous estimation of barrier profile；Accurate barrier profile elevations h is tried to achieve by way of history scanning is superimposed with the continuous recurrence of Current Scan data；By returning computed altitude deviation and pitch angle deviation；The fusion of new and old scanning.The measuring method that the present invention uses is simple and feasible, the general outline height of road ahead barrier can accurately and efficiently be detected, and overlay region constantly carries out recurrence superposition caused by new and old scanning, can remove influences caused by some interference, greatly strengthen information density.Accurate vehicle front barrier profile elevations h can be obtained.
Description
Technical field
The invention belongs to automobile intelligent driving and Radar Technology field, it is related to recognition methods of the radar to target, specifically relates to
And a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm, can not be quickly accurate to solve existing radar
Really identify road obstacle contouring problem.
Background technology
Laser radar is a kind of conventional distance measuring sensor, small etc. excellent due to being disturbed with high resolution, by environmental factor
Point, it is widely used in various fields.Laser radar is divided into three kinds of single line laser radar, multi-line laser radar and face battle array radar.It is single
Line laser radar scans every time produces a scan line, its ranging speed is fast, data volume is few, small volume, it is in light weight, be adapted to it is quick
Processing, because of high performance-price ratio, single line laser radar is most widely used at present.
Laser radar is a kind of sensor that range measurement is carried out using light wave.Laser radar is frequently with pulse propagation time
The distance of method measurement object.During work, it can launch laser pulse in the form of assembling light beam, and measure transmitting and be received back to
Propagation time between ripple, and because laser pulse is with light velocity propagation, it is possible to it is determined that distance.Laser generator is to target list
The distance of individual pip is：D=ct/2, wherein, c is the light velocity；T is to launch laser beam to receiving back from ground laser radar
The time difference of ripple signal；D is distance of the generating laser to target pip.
Laser beam can be assembled by optical system well, therefore be not only distance, and target is relative to sensor
Accurate side and upright position can also be determined.It can realize that millimetre-sized distance is surveyed using this measuring principle
Amount, therefore it is especially suitable for detecting the minor variations of road roughness, so as to build whole ground height profile.
Laser radar light beam can cause distance values to calculate mistake when projecting grit, raindrop surface.And due to vehicle-mounted
There are some external interferences in radar so that radar pulse echo is more uncontrollable during motion state, it is therefore necessary to have
Intelligent analysis process just can reach practical.
Currently for the existing more research of measurement of vehicle front barrier profile：
Pertinent literature 1：Application number 201310063898.2, Ricoh Co., Ltd Chen Chao, Shi Zhongchao utilize binocular camera
A kind of method and system that pavement-height shape is estimated in road scene is provided, this method includes：Obtain road scene
Disparity map；Road surface region of interest is detected based on the disparity map；Determine that multiple road surfaces are interested based on the road surface region of interest
Point；And pavement-height shape is estimated based on the multiple road surface point-of-interest.Because usual road scene is extremely complex, bag
Include pedestrian, vehicle, barrier etc. so that the amount of calculation of algorithm is very big, generally requires processed offline after scan data, it is difficult to real
When on-line checking.
Pertinent literature 2：F.Moosmann etc. proposes a kind of image recognition algorithm with obstacle recognition ability.It is this
Algorithm uses three-dimensional laser radar, and ground and barrier are identified into segmentation.Because three-dimensional laser radar is also not up to measured
It is the production stage, expensive, limit the application of this method.
Pertinent literature 3：Application number 201610804686.9, Song Wei, Zhou little Long, Wu Bin disclose one kind and utilize convolutional Neural
Network C NN obstacle recognition method, using deep learning algorithm, the identification of barrier is carried out based on bionical eye system, and carried
The method for having supplied to configure interface in identification process, strengthen the communication process of identification process and bionical eye system.This method has
There is higher discrimination, but effective execution of algorithm has to rely on the neural metwork training of great amount of images model library, once
Obstacle information lacks in model library, it will greatly influences recognition result.
The content of the invention
Present invention aims at a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm is provided, lead to
Cross repeatedly continuous scanning recurrence matching repeatedly to analyze laser pulse echo-signal, increase the information density of signal with this,
Vehicle front barrier profile elevations h accurately is obtained, can be applied without restriction in road traffic.
The purpose of the present invention is realized by following scheme：
A kind of vehicle front barrier profile testing method based on recurrence superposition algorithm, laser radar mount scheme：
Laser radar be arranged on front part of vehicle headlight height and position, can since the position that bumper terminates measuring channels
Road, the light beam of laser radar more obliquely can be incident upon on track.Shortcoming caused by more flat scan angle is with being arranged on vehicle
When the vehicle lost before several meters of measurement length compare, it is smaller, therefore, to realize it is pre- take aim at function, laser radar is installed
It is proper in headlight height and position；
The inventive method comprises the following steps：
Step 1: calculating barrier profile elevations h by the geometrical relationship on laser radar and ground, polar coordinate system is established, will
The barrier profile elevations h initial data that laser radar collects is converted into polar coordinates by trigonometric function coordinate transform；
Step 2: the coordinate matching of past scan data and present scan data：By trigonometric function coordinate transform barrier
Hinder thing profile elevations h equation to represent to be converted into cartesian coordinate system from polar coordinate system, twice sweep is included by same seat with this
Under mark system；
Step 3: consider that there is radar light beam the characteristics of normal distribution to introduce probability density function：Introduce Gaussian normal point
Cloth, the normpdf of each measurement point is obtained, radar surveying point hot spot is symbolized by probability density function
The true distribution situation of the barrier profile elevations h obtained；
Step 4: realize the quasi-continuous estimation of barrier profile：The probability density of each scanning element is completed by the step 3
Distribution situation, the quasi-continuous estimation of profile elevations h can be carried out by probability density curve；Establish a coordinate system, abscissa generation
The measurement point of table laser radar light beam scanning is to the distance of radar, and ordinate represents barrier profile elevations h value, on the horizontal scale
A shift register with equidistant sampled point is introduced, the height value scanned every time is inputted by the abscissa of a quantization,
Barrier profile elevations h value and probability density are calculated in the register of rasterizing by the distance of laser radar to scanning element；
Using n probability density function summation of single pass as unified standard, the obstacle of each equidistant points is scanned each time
The quasi-continuous estimation of thing profile；
Step 5: try to achieve accurate obstacle by way of past scan data is superimposed with the continuous recurrence of present scan data
Thing profile elevations h, while coefficient correlation is introduced, the degree of correlation that evaluation twice sweep influences on real barrier profile elevations h；
Step 6: pass through linear regression computed altitude value deviation and pitch angle deviation：Showed by the step 5
New relation equation between scan data and in the past scan data, is answered by linear regression, can be obtained multigroup
Barrier profile elevations h value deviation and pitch angle deviation, height value deviation optimum value can be determined using least square method and bowed
Elevation deflection optimum value, it can now calculate the height correction value of new scan data；
Step 7: the fusion of new and old scan data：Height correction value using the step 6 superposition will be swept now
Retouch data and scan data merges before, you can obtain accurate barrier profile elevations h value；Update this barrier profile simultaneously
Height value, recurrence superposition next time is carried out, is superimposed repeatedly, you can obtain accurately vehicle front barrier profile.
The present invention provides a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm, only using only one
The cheap single line laser radar of platform, measurement cost is relatively low, and recurrence of the present invention superposition scan matching algorithm,
The all boundary condition applied in real vehicle is considered, can be applied without restriction in road traffic.The present invention
It completely independent can efficiently be performed independent of the constraint of the conditions such as any extraneous model library, and can realize that real-time online detects
Processing, overcome the problems present in the current field.
Brief description of the drawings
Fig. 1 is principle of the invention conception figure
Fig. 2 is flow chart of the present invention
Fig. 3 is radar and the geometrical relationship calculating barrier profile elevations h on ground
Fig. 4 is the shift register with equidistant sampled point
Fig. 5 is the quasi-continuous estimation for the barrier profile for scanning each equidistant points each time
Embodiment
Technical scheme is discussed in detail below in conjunction with accompanying drawing：
It is a kind of based on the vehicle front barrier profile testing method of recurrence superposition algorithm, it is necessary to which single line laser radar is pacified
Mounted in the headlight height and position of front part of vehicle, measurement road, laser beam can relatively incline since the position that bumper terminates
Tiltedly it is incident upon on track, the inventive method comprises the following steps, as shown in Figure 2：
Step 1: barrier profile elevations h is calculated by the geometrical relationship on laser radar and ground；
Current obstacle height is calculated using the laser beam of the laser radar installed in front part of vehicle, as shown in Figure 3.
While laser radar rotates, laser beam also fan provides measurement point i distance in the range of measurement angle
Numerical value.The pulsed light beam that laser radar is sent relative to road inclination angle n0Expression formula be：Wherein, ncGeneration
Pitching angular variation of the table laser radar in installation site, nLThe relative angle of pitch between car body and wheel is represented,Represent and work as
Preceding lidar measurement light beam relative to sensor outer housing angle.
The scope of lidar measurement angle is 0~90 °.First measurement point position of single pass is in from horizontal line to road
Road about rotates down 45 ° of position.By trigonometric function coordinate transform, the setting height(from bottom) and laser beam of laser radar are utilized
Dip angle parameter, laser radar can be calculated to the absolute upright height z on ground0Arrive sensing in the direction of the x axis with measurement point
The distance x of device0, calculation formula is as follows：
x0=d0*cos(n0)
z0=z-d0*sin(n0)
Original vertical range z is calculated by following formula between laser radar and track：
Z=zcz+zzd-xs*sin(nL)+ys*sin(wL)
Barrier profile elevations h value z can then be derived0Calculation formula it is as follows：
In formula, ZczRepresent vertical shift Z of the laser radar in installation sitecz；Zzd、nL、wLCar body and car are represented respectively
Relative motion between wheel is vibrated, jolts and wave；xsAnd ysVehicle is respectively described in lengthwise position and lateral attitude, vehicle
The distance between center of gravity and laser radar；d0Represent the distance between laser radar and measurement point.
Step 2: the coordinate matching of past scan data and present scan data
During vehicle actual travel, vehicle has the motion on vertical and horizontal certainly, and car body also has in itself
Relative to the relative motion of road, the exercise parameter during this can be described as：Car body longitudinal driving speed vx, car body and
Relative motion vibration z between wheelzd, sensor jolts nL, sensor wave wL。
It is assumed that the transport condition of vehicle is, it is known that can thus overlap twice sweep.Assuming that known front and rear twice sweep number
According to, equation is converted into from polar coordinate representation by trigonometric function coordinate conversion relation by cartesian coordinate system and represented, it is specific to turn
It is as follows to change formula：
Past scan data：
Present scan data：
In formula, past scanning is represented with " past ", is annotated with subscript " p ",Represent in scan data in the past, in x-axis
On direction certain measurement point to sensor distance,Represent and scan obtained barrier profile elevations h value in the past；Scanning now
Represented with " now ", annotated with subscript " n ",Represent in present scan data, certain measurement point is to sensor on x-axis direction
Distance,Represent the barrier profile elevations h value that scanning is obtained now.
Step 3: consider that there is radar light beam the characteristics of normal distribution to introduce probability density function
In the equation of above-mentioned steps two, measurement point is all considered as a point to be studied, and laser thunder
A barrier profile elevations h value is just corresponded to up to each distance values measured.But real situation is, each measurement
Point reality is all formal distribution with hot spot, rather than a point.In a hot spot, height number is entered with certain probability
Row distribution.Using laser radar light beam there is the characteristics of normal distribution to introduce normpdf, measurement point
Probability density can passes through a continuously distributed function, Gauss normal distribution function approximation：
In above equation, x is a random variable of continuous type, can be understood as laser radar and measurement in a model
Horizontal range between point, σ is standard deviation (or variance).
The algorithm that above step is realized all is that such hypothesis is too based on the infinitesimal of measurement point is propagated
Idealization, the height profile of barrier can not really be described.If twice sweep has identical distance basis, then returns
Analysis is returned to realize the superposition of twice sweep.Based on this reason, the coordinate system shown in a Fig. 4, abscissa generation are established
For the measurement point of table laser radar light beam scanning to the distance of radar, ordinate represents barrier profile elevations h value.On the horizontal scale
It is introduced into a shift register (can be understood as an array in algorithm routine) with equidistant sampled point, the height scanned every time
Angle value is used as input by the abscissa value of a quantization, and the diffusion problem of such measurement point is just addressed.
Shift register with equidistant sampled point is with advantages below：The plane distribution of measurement point is considered basically；
The matching of Multiple-Scan is with a common distance basis.
It is exactly the shift register exemplary applications of measurement point in single pass shown in Fig. 4.In a shift register
Some abscissa value has corresponded to a discrete height value, because measurement point is showed in the form of hot spot in truth
, in the range of this hot spot, discrete measured values may occur with some statistics probability, so can is to measure across subjects
The probability density occurred in the range of measurement point distribution.
Introduce after shift register, by Laser Radar Scanning point to laser radar installation site apart from this ginseng
Number, you can to obtain barrier profile elevations h value and the probability density corresponding to this parameter respectively.
In a shift register, every distance be Δ x1Equidistant sampled point carry out abscissa segmentation, equivalent to swash
Optical radar scanning element has carried out rasterizing processing to the distance between radar installation site.Can clearly it be drawn by Fig. 4：
The abscissa of shift register covers the whole measurement range of laser radar signal, according to raster width and maximum
Scanning distance scope a, it can be deduced that shift register with m+1 discrete equidistant sampled points.Citing, if measurement model
Enclose for 0-20m, raster width 10cm, then the sampled point quantity that shift register has is just m+1=201.In register
In, the height value and probability density distribution of each measurement point are inputted by abscissa value x.Result is shown in form 1：
Form 1 is used for the register with equidistant sampled point for preserving scan data
Assuming that one group of scanning has k measurement point, then probability density of the different measurement hot spots in each sampled point 0...m
For：
The probability density function can characterize the order of accuarcy of obstacle height measured in glossing up, probability density peak
Value is bigger, and probability distribution is more concentrated, and the accuracy of measurement is higher.Measurement data can be carried out by the probability density function
Continuous treatment, obtain more dense barrier profile elevations h curve.
Step 4: realize the quasi-continuous estimation of barrier profile
When new scanning produces every time, k height value will be obtained, they are above represented in the form of discrete in distance.But
It is in actual conditions, there can be a corresponding height value on the position that each probability of shift register is not zero.It is assuming that existing
In scanning be made up of k measurement point, then the current estimate can of height value by m+1 in shift register from
Grid point is dissipated, is then calculated according to the vector of k height value of sum of products of probability density matrix (n probability of single pass
Density function summation is as unified standard)：
It is ξ in the probability density value of the 1st sampled point0,1、ξ0,2…ξ0,3, the estimate can standard of respective heights value
Change processing is calculated：
Wherein, the weighted sum of scan data is：
The quasi-continuous estimation of the probability density value and respective heights value of 2nd, 3 ... individual sampled points can also be by standardization at
Reason obtains.
It is ξ in the probability density value of m-th of sampled pointm,1、ξm,2…ξm,k, the estimate can standard of respective heights value
Change processing is calculated：
Wherein, the weighted sum of scan data is：
The quasi-continuous estimation of the barrier profile of each equidistant points can be scanned each time according to algorithm above.
Step 5: try to achieve accurate obstacle by way of past scan data is superimposed with the continuous recurrence of present scan data
Thing profile elevations h
The data of all scannings, including scanning now and the data scanned in the past are used by the recursive call of algorithm,
The signal quality of barrier can be greatly improved.Recursive call scanning in the past and the scan matching scanned now in each scanning
Algorithm, so that it may to realize this target.Recurrence superposition algorithm can be sketched simply with below equation：
The recursive call scanned now：
The recursive call of past scanning：
In formula,The calculated value of height value in present scan data is represented,First sampled point is represented present
Probability density function summation in scan data；The calculated value of height value in scan data in the past is represented,Represented
Go the probability density function summation of first sampled point in scan data；Represent in present scan data in the direction of the x axis
Measurement point to sensor distance,Represent distance of the measurement point to sensor in the direction of the x axis in scan data in the past.
Recursive call algorithmic function f is the fusion process of new and old scan data in the step 7.
Step 6: by returning computed altitude value deviation and pitch angle deviation
In recurrence superposition algorithm, also have to consider the calculating of correlative factor.To newly be scanned now by the Return Law and
Past it is old scanning superposition when, it is necessary to consider that the error of road profile height value.Offsets in height in shift register or
Person's height error can be expressed as：
In formula,Represent the barrier profile elevations h value of scanning in the past；Represent the barrier profile scanned now
Height value；Represent the offsets in height or height error in shift register.
Want to carry out superposition to new scanning and old scanning by linear regression, just must determine will be with great weight
Consider the barrier profile elevations h value error corresponding to each abscissa value in shift register.In simple terms, only swept now
Retouch and pass by the position with high standardization probability density of scanning, just need to consider the error of height value.Therefore, two
Regression analysis is just considered in the probability density distribution minimum intersection range of secondary scanning, because only that the lap of twice sweep is
With correlation.In view of this consideration, introduces relative coefficient R, can be from the minimum of recapitulative probability density function
Correlation R is calculated in standard, specific formula for calculation is as follows：
In formula,Represent respectively present scan data and in the past in scan data probability density function and.
A parameter is added in recurrence superposition algorithm again：Relative coefficient R, this is also indicated that, it is determined that scanning now
When the barrier profile elevations h value deviation and pitch angle deviation that were scanned with the past, it is also considered that the hot spot of laser measurement point is put down
EDS maps, and the factor such as height value probability density distribution corresponding to the measurement point occurred therewith.In present scan data and mistake
Go between scan data, it can be deduced that following new relation：
In formula, R represents relative coefficient；Represent the measurement point in the shift register with equidistant sampled point and swash
Distance in optical radar X-direction；Δ n, Δ z represent pitch angle deviation between new and old scan data respectively and height value is inclined
Difference；
This equation is an over-determined systems, similar to Ax=b.Above-mentioned equation can be solved by linear regression
Answer.Build the generalized inverse matrix A+ of matrix A：
Multigroup barrier profile elevations h value deviation and pitch angle deviation Δ n can be obtained according to above formula, using least square
Method can determine height value deviation optimum value Δ z and pitch angle deviation optimum value Δ n, can now calculate new scan data
Height correction value：
z0,cy,xz=z0,cy,n+Δn*x0,cy+Δz
In formula, new and old scanning is overlapped, z0,cy,xzRepresent the height correction value of new scan data.
Step 7: the fusion of new and old scan data
By the realization of above step, amendment before can now utilizing obtained by recurrence superposition will scan now
Data and the in the past data fusion of scanning.After the data newly scanned are added in the data preserved of scanning in the past, own
The generality probability density scanned before will increase the probability density newly scanned：
∑ξ0,sum=∑ ξ0,n+∑ξ0,p
In formula, ∑ ξ0,sumRepresent the generality probability density for all scannings that first sampled point is included；∑ξ0,pRepresent
The generality probability density summation of past scan data；∑ξ0,nRepresent the generality probability density summation of present scan data.
On the premise of new probability density is considered, the average height value after the renewal of road profile can be calculated：
This average height value z0,cy,sumThe accurate barrier profile elevations h value exactly finally given, uses z0,cy,sumGeneration
For the barrier profile elevations h value of present scan, the recurrence for then carrying out scanning now again with new round scanning is superimposed, under calculating
The height value of secondary scanning, recurrence superposition repeatedly, you can obtain accurately vehicle front barrier profile.
The present invention has advantages below:
New recurrence superposition barrier profile information Processing Algorithm considers all boundary applied in real vehicle
Condition, it can apply without restriction in road traffic.
The information of single sweep operation is excessively imperfect and inaccurate.This algorithm, which is made full use of in continuous scanning, part range
The fact that be overlapping, " superposition " will be being scanned to a certain degree and is improving information density.
By probability density function and the introducing of quasi-continuous estimation, recurrence superposition algorithm can make radar scanning data
" infinitely approaching " actual value.
In order that those skilled in the art more fully understand the present invention, below with MATLAB algorithm simulating flow to this hair
It is bright to make further example explanation.
Powerful array operation ability is had based on MATLAB in itself, for the reliability of check algorithm, can be utilized
MATLAB builds the algorithm of this subject study, carries out fail-safe analysis.It is the algorithm for building shaping below, it is assumed that respectively obtain sharp
Two groups of data of optical radar twice sweep.First, data are subjected to coordinate axis transform and realize the function of equidistant shift register,
The matching of twice sweep is realized using interpl functions afterwards；Next, will newly it be scanned with polyfit and polyval functions
To data be fitted to obtain new scan data after new recurrence, the standard of probability density function in algorithm is realized with this
Change；Then, relative coefficient R, the matrix A comprising coefficient correlation are drawn, x is tried to achieve with this, calculates height value deviation and pitching
Angular displacement.Data fusion is finally obtained into corrected new height value, completes the realization of the algorithm.
Claims (8)
1. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm, it is characterised in that by single line laser
Radar be arranged on front part of vehicle headlight height and position, can since the position that bumper terminates measurement road, laser beam
Inclination is incident upon on track, and the inventive method comprises the following steps：
Step 1: calculating barrier profile elevations h by the geometrical relationship on laser radar and ground, polar coordinate system is established, by laser
The barrier profile elevations h initial data that radar collects is converted into polar coordinates by trigonometric function coordinate transform；
Step 2: the coordinate matching of past scan data and present scan data：By trigonometric function coordinate transform barrier
Profile elevations h equation is converted into cartesian coordinate system from polar coordinate system expression, and twice sweep is included into same coordinate system with this
Under；
Step 3: consider that there is radar light beam the characteristics of normal distribution to introduce probability density function：Gauss normal distribution is introduced, is obtained
To the normpdf of each measurement point, radar surveying point hot spot is symbolized by probability density function and obtained
Barrier profile elevations h true distribution situation；
Step 4: realize the quasi-continuous estimation of barrier profile：The probability density distribution of each scanning element is completed by the step 3
Situation, the quasi-continuous estimation of profile elevations h can be carried out by probability density curve；A coordinate system is established, abscissa, which represents, to swash
To the distance of radar, ordinate represents barrier profile elevations h value, introduces on the horizontal scale the measurement point of optical radar light beam scanning
One shift register with equidistant sampled point, the height value scanned every time are inputted by the abscissa of a quantization, passed through
The distance of laser radar to scanning element calculates barrier profile elevations h value and probability density in the register of rasterizing；One
K probability density function summation of secondary scanning is scanned the barrier wheel of each equidistant points as unified standard each time
Wide quasi-continuous estimation；
Step 5: try to achieve accurate barrier wheel by way of past scan data is superimposed with the continuous recurrence of present scan data
Wide height, while coefficient correlation is introduced, the degree of correlation that evaluation twice sweep influences on real barrier profile elevations h；
Step 6: pass through linear regression computed altitude value deviation and pitch angle deviation：Obtained by the step 5 and swept now
The new relation equation between data and in the past scan data is retouched, is answered by linear regression, multigroup obstacle can be obtained
Thing profile elevations h value deviation and pitch angle deviation, height value deviation optimum value and the angle of pitch can be determined using least square method
Deviation optimum value, it can now calculate the height correction value of new scan data；
Step 7: the fusion of new and old scan data：Number will be scanned now using the height correction value of the step 6 superposition
Merged according to scan data before, you can obtain accurate barrier profile elevations h value；Update this barrier profile elevations h simultaneously
Value, recurrence superposition next time is carried out, is superimposed repeatedly, you can obtain accurately vehicle front barrier profile.
2. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm as claimed in claim 1, it is special
Sign is that the step 1 calculates barrier profile elevations h by the geometrical relationship on laser radar and ground includes procedure below：
The pulsed light beam that laser radar is sent relative to road inclination angle n0Expression formula be：
Wherein, ncRepresent pitching angular variation of the laser radar in installation site, nLRepresent the relative pitching between car body and wheel
Angle,Represent angle of the present laser radar surveying light beam relative to sensor outer housing；
By trigonometric function coordinate transform, using the setting height(from bottom) of laser radar and the dip angle parameter of laser beam, can calculate
Go out barrier profile elevations h value z0Distance x of the measurement point to sensor in the direction of the x axis0, calculation formula is as follows：
x0=d0*cos(n0)
z0=z-d0*sin(n0)
Original vertical range z is calculated by following formula between laser radar and track：
Z=zcz+zzd-xs*sin(nL)+ys*sin(wL)
Barrier profile elevations h value z can then be derived0Calculation formula it is as follows：
In formula, ZczRepresent vertical shift Z of the laser radar in installation sitecz；Zzd、nL、wLRepresent respectively car body and wheel it
Between relative motion vibration, jolt and wave；xsAnd ysVehicle is respectively described in lengthwise position and lateral attitude, vehicle's center of gravity
The distance between laser radar；d0Represent the distance between laser radar and measurement point.
3. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm as claimed in claim 1, it is special
Sign is that the step 2 past scan data and the process of the coordinate matching of present scan data are：
It is assumed that the transport condition of vehicle is, it is known that can thus overlap twice sweep；
Assuming that known forward and backward twice sweep data, sit barrier profile elevations h from pole by trigonometric function coordinate conversion relation
Mark expression is converted into cartesian coordinate system expression, and specific conversion formula is as follows：
Past scan data：
Present scan data：
In formula,Represent in scan data in the past, in the direction of the x axis the distance of certain measurement point to sensor,Representative is swept in the past
Retouch obtained barrier profile elevations h value；Represent in present scan data, certain measurement point arrives sensor on x-axis direction
Distance,Represent the barrier profile elevations h value that scanning is obtained now.
4. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm as claimed in claim 1, it is special
Sign is that the step 3 considers that there is the characteristics of normal distribution to introduce the process of probability density function and be for radar light beam：
Using laser radar light beam there is the characteristics of normal distribution to introduce normpdf, the probability of measurement point is close
Degree can passes through a continuously distributed function, Gauss normal distribution function approximation：
<mrow>
<mi>&xi;</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mi>&sigma;</mi>
<msqrt>
<mrow>
<mn>2</mn>
<mi>&pi;</mi>
</mrow>
</msqrt>
</mrow>
</mfrac>
<msup>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mi>x</mi>
<mo>-</mo>
<msub>
<mi>x</mi>
<mrow>
<mi>r</mi>
<mi>e</mi>
<mi>f</mi>
</mrow>
</msub>
</mrow>
<mi>&sigma;</mi>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
</msup>
</mrow>
X is a random variable of continuous type in formula, represents the horizontal range between laser radar and measurement point；σ is standard deviation；
A coordinate system is established, abscissa represents the measurement point of laser radar light beam scanning and represented to the distance of radar, ordinate
Barrier profile elevations h value；
A shift register with equidistant sampled point is introduced on the horizontal scale, and the height value scanned every time passes through a quantization
Abscissa value as input：By Laser Radar Scanning point to laser radar installation site apart from this parameter, you can with
Barrier profile elevations h value and the probability density corresponding to this parameter are obtained respectively；
In a shift register, every distance be Δ x1Equidistant sampled point carry out abscissa segmentation, equivalent to laser radar
Scanning element has carried out rasterizing processing to the distance between radar installation site：
<mrow>
<msub>
<mover>
<mi>x</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>=</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mn>0</mn>
</msub>
<mo>,</mo>
<msub>
<mi>x</mi>
<mn>0</mn>
</msub>
<mo>+</mo>
<mn>1</mn>
<mo>*</mo>
<mi>&Delta;</mi>
<mi>x</mi>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mn>0</mn>
</msub>
<mo>+</mo>
<mi>m</mi>
<mo>*</mo>
<msub>
<mi>&Delta;x</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>,</mo>
<mn>...</mn>
<mo>,</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>m</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>0....</mn>
<mi>m</mi>
<mo>,</mo>
<msub>
<mover>
<mi>x</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>&Element;</mo>
<msup>
<mi>R</mi>
<mrow>
<mi>m</mi>
<mo>+</mo>
<mn>1</mn>
</mrow>
</msup>
</mrow>
The abscissa of shift register covers the whole measurement range of laser radar signal, according to raster width and maximum scan
Distance range a, it can be deduced that shift register with m+1 discrete equidistant sampled points：I.e. 0,1...m
In a register, barrier profile elevations h value and probability density distribution pass through abscissa value corresponding to each measurement point
X inputs are entered；
Assuming that one group of scanning has k measurement point, then probability density of the different measurement hot spots in each sampled point 0...m is：
<mrow>
<msub>
<mi>&xi;</mi>
<mi>j</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<msub>
<mi>&sigma;</mi>
<mrow>
<mi>c</mi>
<mi>l</mi>
<mi>d</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<msqrt>
<mrow>
<mn>2</mn>
<mi>&pi;</mi>
</mrow>
</msqrt>
</mrow>
</mfrac>
<msup>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<msub>
<mi>x</mi>
<mi>j</mi>
</msub>
</mrow>
<mrow>
<msub>
<mi>&sigma;</mi>
<mrow>
<mi>c</mi>
<mi>l</mi>
<mi>d</mi>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>j</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
</msup>
<mo>,</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mn>...</mn>
<mi>k</mi>
<mo>;</mo>
<mi>j</mi>
<mo>=</mo>
<mn>0</mn>
<mo>,</mo>
<mn>1...</mn>
<mi>m</mi>
</mrow>
Above-mentioned probability density function can characterize the order of accuarcy of obstacle height measured in glossing up, pass through probability density
Function can carry out quasi-continuousization estimation processing to measurement data, obtain more dense barrier profile elevations h curve.
5. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm as claimed in claim 1, it is special
Sign is that the quasi-continuous estimation of barrier profile is realized according to step 4 includes procedure below：
It is ξ in the probability density value of the 1st sampled point0,1、ξ0,2…ξ0,3, at the estimate can standardization of respective heights value
Reason is calculated：
<mrow>
<msub>
<mover>
<mi>z</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<mo>&Sigma;</mo>
<mrow>
<mo>(</mo>
<msub>
<mover>
<mi>&xi;</mi>
<mo>&RightArrow;</mo>
</mover>
<mn>0</mn>
</msub>
<mo>*</mo>
<msub>
<mover>
<mi>z</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mo>&Sigma;</mo>
<msub>
<mover>
<mi>&xi;</mi>
<mo>&RightArrow;</mo>
</mover>
<mn>0</mn>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2...</mn>
<mi>k</mi>
</mrow>
Wherein, the weighted sum of scan data is：
The quasi-continuous estimation of the probability density value and respective heights value of 2nd, 3 ... individual sampled points can also be obtained by standardization
Arrive.
It is ξ in the probability density value of m-th of sampled pointm,1、ξm,2…ξm,k, at the estimate can standardization of respective heights value
Reason is calculated：
Wherein, the weighted sum of scan data is：
The quasi-continuous estimation of the barrier profile of each sampled point can be scanned each time according to algorithm above.
6. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm as claimed in claim 1, it is special
Sign is that the step 5 tries to achieve accurate barrier by way of past scan data is superimposed with the continuous recurrence of present scan data
Thing profile elevations h is hindered to include procedure below：
By taking first sampled point as an example, recursive call scans the scan matching algorithm with scanning now, recurrence superposition algorithm in the past
Sketched with below equation：
The recursive call of present scan data：
The recursive call of past scan data：
In formula,The calculated value of height value in present scan data is represented,First sampled point is represented in scanning number now
Probability density function summation in；The calculated value of height value in scan data in the past is represented,Represent and scan in the past
The probability density function summation of first sampled point in data；Represent in present scan data measurement point in the direction of the x axis
To the distance of sensor,Represent distance of the measurement point to sensor in the direction of the x axis in scan data in the past.
7. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm as claimed in claim 1, it is special
Sign is that the step 6 includes procedure below by returning computed altitude value deviation and pitch angle deviation：
Offsets in height in the step 4 in shift register can be expressed as：
<mrow>
<mover>
<mi>&epsiv;</mi>
<mo>&RightArrow;</mo>
</mover>
<msub>
<mi>z</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mover>
<mi>x</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mover>
<mi>z</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>p</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mover>
<mi>z</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>n</mi>
</mrow>
</msub>
</mrow>
In formula,Represent the barrier profile elevations h value of scan data in the past；Represent the barrier of present scan data
Profile elevations h value；Represent the offsets in height or height error in shift register；
Introduce relative coefficient R：
<mrow>
<mi>R</mi>
<mo>=</mo>
<mo>|</mo>
<mo>|</mo>
<mo>&Sigma;</mo>
<msub>
<mover>
<mi>&xi;</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>n</mi>
</mrow>
</msub>
<mo>,</mo>
<mo>&Sigma;</mo>
<msub>
<mover>
<mi>&xi;</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>p</mi>
</mrow>
</msub>
<mo>|</mo>
<msub>
<mo>|</mo>
<mi>min</mi>
</msub>
</mrow>
In formula,Respectively represent now scanning and in the past scan in probability density function and；
The degree of correlation influenceed by relative coefficient R evaluation twice sweeps on real barrier profile elevations h；
Therefore between present scan data and in the past scan data, it can be deduced that following new governing equation：
<mrow>
<mo>(</mo>
<mi>R</mi>
<mo>,</mo>
<mi>R</mi>
<mo>*</mo>
<msub>
<mover>
<mi>x</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>)</mo>
<mo>*</mo>
<mfenced open = "(" close = ")">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>&Delta;</mi>
<mi>z</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>&Delta;</mi>
<mi>n</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<mi>R</mi>
<mo>*</mo>
<mover>
<mi>&epsiv;</mi>
<mo>&RightArrow;</mo>
</mover>
<msub>
<mi>z</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
</mrow>
In formula, R represents relative coefficient；Represent measurement point and laser thunder in the shift register with equidistant sampled point
Up to the distance in X-direction；Δ n, Δ z represent the pitch angle deviation and height value deviation between new and old scan data respectively；
Above-mentioned equation can be answered by linear regression, build the generalized inverse matrix A+ of matrix A：
<mrow>
<mover>
<mi>x</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<msup>
<mi>A</mi>
<mo>+</mo>
</msup>
<mo>*</mo>
<mi>b</mi>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>A</mi>
<mi>T</mi>
</msup>
<mo>*</mo>
<mi>A</mi>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mo>*</mo>
<msup>
<mi>A</mi>
<mi>T</mi>
</msup>
<mo>*</mo>
<mi>b</mi>
</mrow>
<mrow>
<mfenced open = "(" close = ")">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>&Delta;</mi>
<mi>z</mi>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>&Delta;</mi>
<mi>n</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>=</mo>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mrow>
<mo>(</mo>
<mrow>
<mi>R</mi>
<mo>,</mo>
<mi>R</mi>
<msub>
<mover>
<mi>x</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
</mrow>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<mo>&CenterDot;</mo>
<mo>(</mo>
<mrow>
<mi>R</mi>
<mo>,</mo>
<mi>R</mi>
<msub>
<mover>
<mi>x</mi>
<mo>&RightArrow;</mo>
</mover>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
</mrow>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mo>&CenterDot;</mo>
<msup>
<mrow>
<mo>(</mo>
<mi>R</mi>
<mo>,</mo>
<msub>
<mi>Rx</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<mo>&CenterDot;</mo>
<mrow>
<mo>(</mo>
<mi>R</mi>
<mo>&CenterDot;</mo>
<mover>
<mi>&epsiv;</mi>
<mo>&RightArrow;</mo>
</mover>
<msub>
<mi>z</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
</mrow>
Multigroup barrier profile elevations h value deviation and pitch angle deviation can be obtained according to above formula, can be true using least square method
Determine height value deviation optimum value Δ z and pitch angle deviation optimum value Δ n, can now calculate the height correction of new scan data
Value：
z0,cy,xz=z0,cy,n+Δn*x0,cy+Δz
In formula, new and old scanning is overlapped, z0,cy,xzRepresent the height correction value of new scan data.
8. a kind of vehicle front barrier profile testing method based on recurrence superposition algorithm as claimed in claim 1, it is special
Sign is that the fusion of the new and old scan data of step 7 includes procedure below：
After the data scanned now are added in the data preserved of scanning in the past, the generality probability of all scannings before
Density will increase the probability density newly scanned：
∑ξ0,sum=∑ ξ0,n+∑ξ0,p
In formula, ∑ ξ0,sumRepresent the generality probability density for all scan datas that first sampled point is included；∑ξ0,pRepresent
The generality probability density summation of past scan data；∑ξ0,nRepresent the generality probability density summation of present scan data；
On the premise of new probability density is considered, the average height value after the renewal of road profile can be calculated：
<mrow>
<msub>
<mi>z</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>s</mi>
<mi>u</mi>
<mi>m</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>z</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>p</mi>
</mrow>
</msub>
<mo>*</mo>
<mo>&Sigma;</mo>
<msub>
<mi>&xi;</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>p</mi>
</mrow>
</msub>
<mo>+</mo>
<msub>
<mi>z</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>c</mi>
<mi>y</mi>
<mo>,</mo>
<mi>x</mi>
<mi>z</mi>
</mrow>
</msub>
<mo>*</mo>
<mo>&Sigma;</mo>
<msub>
<mi>&xi;</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>n</mi>
</mrow>
</msub>
</mrow>
<mrow>
<mo>&Sigma;</mo>
<msub>
<mi>&xi;</mi>
<mrow>
<mn>0</mn>
<mo>,</mo>
<mi>s</mi>
<mi>u</mi>
<mi>m</mi>
</mrow>
</msub>
</mrow>
</mfrac>
</mrow>
This average height value z0,cy,sumThe accurate barrier profile elevations h value exactly finally given, uses z0,cy,sumInstead of this
The barrier profile elevations h value of scanning, the recurrence for then carrying out scanning now again with new round scanning are superimposed, and calculate scanning next time
Height value, repeatedly recurrence be superimposed, you can obtain accurately vehicle front barrier profile.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710733511.8A CN107632308B (en) | 2017-08-24 | 2017-08-24 | Method for detecting contour of obstacle in front of vehicle based on recursive superposition algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710733511.8A CN107632308B (en) | 2017-08-24 | 2017-08-24 | Method for detecting contour of obstacle in front of vehicle based on recursive superposition algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107632308A true CN107632308A (en) | 2018-01-26 |
| CN107632308B CN107632308B (en) | 2021-02-05 |
Family
ID=61100143
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710733511.8A Active CN107632308B (en) | 2017-08-24 | 2017-08-24 | Method for detecting contour of obstacle in front of vehicle based on recursive superposition algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107632308B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108445503A (en) * | 2018-03-12 | 2018-08-24 | 吉林大学 | The unmanned path planning algorithm merged with high-precision map based on laser radar |
| CN109447943A (en) * | 2018-09-21 | 2019-03-08 | 中国科学院深圳先进技术研究院 | A kind of object detection method, system and terminal device |
| CN109633685A (en) * | 2018-11-22 | 2019-04-16 | 浙江中车电车有限公司 | A kind of method and system based on laser radar obstruction detection state |
| CN109633676A (en) * | 2018-11-22 | 2019-04-16 | 浙江中车电车有限公司 | A kind of method and system based on the laser radar obstruction detection direction of motion |
| CN111505623A (en) * | 2020-04-24 | 2020-08-07 | 中南大学 | Method and system for detecting obstacle in driving process of unmanned vehicle and vehicle |
| WO2020221123A1 (en) | 2019-04-28 | 2020-11-05 | 郑州宇通客车股份有限公司 | Vehicle control system based on height of obstacle, and vehicle |
| CN112947454A (en) * | 2021-02-25 | 2021-06-11 | 浙江理工大学 | Fire fighting evaluation method, device, equipment and storage medium for warehouse |
| CN113296117A (en) * | 2020-04-22 | 2021-08-24 | 追创科技(苏州)有限公司 | Obstacle recognition method, obstacle recognition device and storage medium |
| CN113777616A (en) * | 2021-07-27 | 2021-12-10 | 武汉市异方体科技有限公司 | Distance measuring method for moving vehicle |
| CN114485855A (en) * | 2022-01-26 | 2022-05-13 | 吉林大学 | Vehicle front surface ponding depth detection system based on self-vehicle sensor |
| CN114994638A (en) * | 2022-08-04 | 2022-09-02 | 之江实验室 | Automatic driving automobile obstacle identification method based on elliptic envelope curve set |
| CN116755056A (en) * | 2023-07-31 | 2023-09-15 | 中国人民解放军国防科技大学 | Distance measurement resolution significantly enhancing method and system for hyperspectral laser radar |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010185769A (en) * | 2009-02-12 | 2010-08-26 | Toyota Motor Corp | Object detector |
| CN102721397A (en) * | 2012-06-07 | 2012-10-10 | 江苏科技大学 | Method for extracting road surface characteristic parameters based on modern time series of vertical dynamic load |
| US20130141580A1 (en) * | 2011-12-06 | 2013-06-06 | Mobileye Technologies Limited | Road vertical contour detection |
| CN103197323A (en) * | 2013-04-17 | 2013-07-10 | 清华大学 | Scanning data matching method and device for laser distance measuring machine |
| CN103669183A (en) * | 2013-12-02 | 2014-03-26 | 黑龙江科技大学 | Time sequence model of road surface evenness |
| CN104233935B (en) * | 2014-08-28 | 2016-05-11 | 吉林大学 | A kind of pavement quality grade discrimination method based on profile of road information |
-
2017
- 2017-08-24 CN CN201710733511.8A patent/CN107632308B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010185769A (en) * | 2009-02-12 | 2010-08-26 | Toyota Motor Corp | Object detector |
| US20130141580A1 (en) * | 2011-12-06 | 2013-06-06 | Mobileye Technologies Limited | Road vertical contour detection |
| CN102721397A (en) * | 2012-06-07 | 2012-10-10 | 江苏科技大学 | Method for extracting road surface characteristic parameters based on modern time series of vertical dynamic load |
| CN103197323A (en) * | 2013-04-17 | 2013-07-10 | 清华大学 | Scanning data matching method and device for laser distance measuring machine |
| CN103669183A (en) * | 2013-12-02 | 2014-03-26 | 黑龙江科技大学 | Time sequence model of road surface evenness |
| CN104233935B (en) * | 2014-08-28 | 2016-05-11 | 吉林大学 | A kind of pavement quality grade discrimination method based on profile of road information |
Non-Patent Citations (1)
| Title |
|---|
| 张黎光: "基于激光的路面平整度自动检测***研究与实现", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108445503A (en) * | 2018-03-12 | 2018-08-24 | 吉林大学 | The unmanned path planning algorithm merged with high-precision map based on laser radar |
| CN109447943A (en) * | 2018-09-21 | 2019-03-08 | 中国科学院深圳先进技术研究院 | A kind of object detection method, system and terminal device |
| CN109447943B (en) * | 2018-09-21 | 2020-08-14 | 中国科学院深圳先进技术研究院 | Target detection method, system and terminal equipment |
| CN109633685A (en) * | 2018-11-22 | 2019-04-16 | 浙江中车电车有限公司 | A kind of method and system based on laser radar obstruction detection state |
| CN109633676A (en) * | 2018-11-22 | 2019-04-16 | 浙江中车电车有限公司 | A kind of method and system based on the laser radar obstruction detection direction of motion |
| WO2020221123A1 (en) | 2019-04-28 | 2020-11-05 | 郑州宇通客车股份有限公司 | Vehicle control system based on height of obstacle, and vehicle |
| JP2023500994A (en) * | 2020-04-22 | 2023-01-17 | 追▲べき▼創新科技（蘇州）有限公司 | Obstacle recognition method, device, autonomous mobile device and storage medium |
| JP7383828B2 (en) | 2020-04-22 | 2023-11-20 | 追▲べき▼創新科技（蘇州）有限公司 | Obstacle recognition method, device, autonomous mobile device and storage medium |
| CN113296117A (en) * | 2020-04-22 | 2021-08-24 | 追创科技(苏州)有限公司 | Obstacle recognition method, obstacle recognition device and storage medium |
| WO2021212986A1 (en) * | 2020-04-22 | 2021-10-28 | 追觅创新科技(苏州)有限公司 | Obstacle identification method and apparatus, self-moving device and storage medium |
| CN113296117B (en) * | 2020-04-22 | 2023-08-08 | 追觅创新科技(苏州)有限公司 | Obstacle recognition method, obstacle recognition device and storage medium |
| CN111505623A (en) * | 2020-04-24 | 2020-08-07 | 中南大学 | Method and system for detecting obstacle in driving process of unmanned vehicle and vehicle |
| CN112947454A (en) * | 2021-02-25 | 2021-06-11 | 浙江理工大学 | Fire fighting evaluation method, device, equipment and storage medium for warehouse |
| CN113777616A (en) * | 2021-07-27 | 2021-12-10 | 武汉市异方体科技有限公司 | Distance measuring method for moving vehicle |
| CN114485855A (en) * | 2022-01-26 | 2022-05-13 | 吉林大学 | Vehicle front surface ponding depth detection system based on self-vehicle sensor |
| CN114994638A (en) * | 2022-08-04 | 2022-09-02 | 之江实验室 | Automatic driving automobile obstacle identification method based on elliptic envelope curve set |
| CN116755056A (en) * | 2023-07-31 | 2023-09-15 | 中国人民解放军国防科技大学 | Distance measurement resolution significantly enhancing method and system for hyperspectral laser radar |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107632308B (en) | 2021-02-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107632308A (en) | A kind of vehicle front barrier profile testing method based on recurrence superposition algorithm | |
| CN1940591B (en) | System and method of target tracking using sensor fusion | |
| US8331653B2 (en) | Object detector | |
| CN109085570A (en) | Automobile detecting following algorithm based on data fusion | |
| CN110531376A (en) | Detection of obstacles and tracking for harbour automatic driving vehicle | |
| CN106405555A (en) | Obstacle detecting method and device used for vehicle-mounted radar system | |
| CN110738121A (en) | front vehicle detection method and detection system | |
| CN108132025A (en) | A kind of vehicle three-dimensional outline scans construction method | |
| CN103578117A (en) | Method for determining poses of camera relative to environment | |
| US11994531B2 (en) | Method of determining an uncertainty estimate of an estimated velocity | |
| Thormann et al. | Extended target tracking using Gaussian processes with high-resolution automotive radar | |
| CN104021676A (en) | Vehicle positioning and speed measuring method based on dynamic video feature of vehicle | |
| CN110758379B (en) | Method and device for detecting inclined parking space and automatic parking method and system | |
| CN112346463B (en) | Unmanned vehicle path planning method based on speed sampling | |
| CN112781599B (en) | Method for determining the position of a vehicle | |
| CN104700414A (en) | Rapid distance-measuring method for pedestrian on road ahead on the basis of on-board binocular camera | |
| CN107796373B (en) | Distance measurement method based on monocular vision of front vehicle driven by lane plane geometric model | |
| Kellner et al. | Road curb detection based on different elevation mapping techniques | |
| CN109901193A (en) | The light of short distance barrier reaches arrangement for detecting and its method | |
| CN115079143B (en) | Multi-radar external parameter quick calibration method and device for double-bridge steering mine card | |
| Kamann et al. | Object tracking based on an extended Kalman filter in high dynamic driving situations | |
| TWI680898B (en) | Light reaching detection device and method for close obstacles | |
| CN116736322B (en) | Speed prediction method integrating camera image and airborne laser radar point cloud data | |
| Zhu et al. | A real-time road boundary detection algorithm based on driverless cars | |
| Barsi et al. | An offline path planning method for autonomous vehicles |
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 |