CN106355880A - Unmanned vehicle control parameter calibrating method for vehicle-following safety - Google Patents

Abstract

The invention discloses an unmanned vehicle control parameter calibrating method for vehicle-following safety, comprising the following steps: extracting a vehicle-following data set of vehicles before and after a manned vehicle and calculating safety indexes of the manned vehicle by collecting trajectory data information of the manned vehicle; utilizing data of the manned vehicle to calculate a data set of an unmanned vehicle, and calculating the safety indexes of the unmanned vehicle; establishing a target safety function according to the safety indexes of the manned vehicle and the unmanned vehicle; detecting control parameters of the unmanned vehicle; and finally determining the parameter which minimizes the target safety function as the control parameter of the unmanned vehicle. The unmanned vehicle control parameter calibrating method for the vehicle-following safety disclosed by the invention has the beneficial effect that safety of the unmanned vehicle is ensured by excavating existing data of the manned vehicle in combination with a safe vehicle following target of the unmanned vehicle.

Description

A kind of towards with car safety automatic driving vehicle control parameter scaling method

Technical field

The present invention relates to technical field of control over intelligent traffic, especially a kind of towards the automatic driving vehicle control with car safety Parameter calibration method processed.

Background technology

Fast-developing with national economy, motorization level improves constantly, and the traffic problems of China are becoming increasingly acute, and are brought Traffic safety problem especially prominent.According to statistics, only 2010, just there is vehicle accident 210821 in the whole nation, accident causes extremely The number of dying reaches as many as 6.5 ten thousand.In numerous vehicle accidents, occupy very great ratio with car rear-end collision.Due to front Rear vehicle hypotelorism, if rear car driver can not be adjusted in time during the mutation of front truck driving condition, will result in and chase after with car Tail accident.

With the fast development of intelligent transport technology, automatic driving vehicle technology gets the attention.Automatic driving car By vehicle-mounted awareness apparatus, front truck is detected, and carry out timely feedback operation, chase after with car therefore, it is possible to preferable reduction The risk of tail accident.However, existing automatic driving vehicle control system parameter be determined to become a big bottleneck.Automatic driving vehicle Manufacturer is typically based on driving habit to carry out parameter determination, and is adjusted by the method for tentative calculation, lacks systematic science Control parameter scaling method, thus the safety of the automatic driving vehicle of design cannot be ensured.

Content of the invention

The technical problem to be solved is, provides a kind of control towards the automatic driving vehicle with car safety to join Number scaling method, can gather the safety index of pilot steering vehicle and automatic driving vehicle, set up targeted security function and come really Determine the control parameter of automatic driving vehicle, ensure the safety of automatic driving vehicle.

For solve above-mentioned technical problem, the present invention provide a kind of towards with car safety automatic driving vehicle control parameter mark Determine method, comprise the steps:

(1) gather pilot steering track of vehicle data: the video that pilot steering vehicle runs is obtained by unmanned plane, Carry out pilot steering track of vehicle data by Video processing software to extract, track data include before and after pilot steering vehicle car with Car data collection d_lAnd d_f；Wherein, data set d_lThe position of the front truck followed in the kth second including n-th pilot steering vehicle The speed of the front truck that n-th pilot steering vehicle was followed in the kth secondN-th pilot steering vehicle is before the kth second follows The length of wagon of carData set d_fIncluding n-th pilot steering vehicle in the position of kth secondN-th pilot steering Vehicle is in the speed of kth secondThe duration t that n-th pilot steering vehicle runs_n, n=1,2 ..., n, n and k are more than 0 Positive integer；

(2) according to the information in step (1), calculate the safety index in the kth second for n-th pilot steering vehicleFor:

${\mathrm{ttc}}_n^k=\frac{x_{l,n}^k-x_{f,n}^k-l_{l,n}^k}{v_{f,n}^k-v_{l,n}^k}$

(3) according to the information in step (1), automatic driving vehicle is iterated to calculate by automatic driving vehicle Controlling model Data set u_f；

$e_n^{k,i}=x_{l,n}^k-x_{f,n}^{*k,i}-t_{hw}^i{v}_{f,n}^{*k,i}$

$v_{f,n}^{*k+1,i}=v_{f,n}^{*k,i}+k_p^i{e}_n^{k,i}+k_d^i{\stackrel{\·}{e}}_n^{k,i}$

$x_{f,n}^{*k+1,i}=x_{f,n}^{*k,i}+v_{f,n}^{*k+1,i}*1$

Wherein, parameterFor i-th group of parameter value, i=1,2 ..., m, m areNumber；

For model parameter it isWhen calculate n-th automatic driving vehicle of gained kth second and front truck spacing by mistake Difference；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the position of kth second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle the kth second speed；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth position of+1 second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth speed of+1 second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle kth second and front truck interval error Derivative；

The initial value of iteration is set to: k=1,Data set u_fIncludingWith

(4) according to the data set u in step (3)_f, computation model parameter isWhen calculate gained n-th automatic driving car The kth second safety indexFor:

$ttc{*}_n^{k,i}=\frac{x_{l,n}^k-x_{f,n}^{*k,i}-l_{l,n}^k}{v_{f,n}^{*k,i}-v_{l,n}^k}$

(5) according to the information in step (1), step (2) and step (4), targeted security function y, the computing formula of y are set up As follows:

$y=y\left({\stackrel{\→}{\mathrm{\β}}}_i\right)=\underset{n=1}{\overset{n}{\σ}}\frac{\sqrt{\underset{k=1}{\overset{t_n}{\σ}}{(ttc{*}_n^{k,i}-{\mathrm{ttc}}_n^k)}^2}}{t_n}$

(6) according to the information in step (5), judge the parameter in automatic driving vehicle Controlling modelWhether all examine Survey, that is, whether i is equal to m, if so, then proceeds to step (7)；Otherwise, i=i+1, proceeds to step (3)；

(7) output makes the minimum parameter value of targeted security function yThat is:

$y\left({\stackrel{\→}{\mathrm{\β}}}^{*}\right)=min\{y\left({\stackrel{\→}{\mathrm{\β}}}_1\right),y\left({\stackrel{\→}{\mathrm{\β}}}_2\right),\mathrm{...},y\left({\stackrel{\→}{\mathrm{\β}}}_i\right),\mathrm{...},y\left({\stackrel{\→}{\mathrm{\β}}}_m\right)\}.$

The invention has the benefit that by the track data information gathering pilot steering vehicle, extracting pilot steering car In front and back car is with car data collection, and calculates pilot steering vehicle safety index, utilizes pilot steering vehicle data to calculate no simultaneously People drives vehicle data collection, and calculates automatic driving vehicle safety index, by pilot steering vehicle and automatic driving vehicle Safety index, to set up targeted security function, detects automatic driving vehicle control parameter, and final determination makes targeted security function Minimum parameter is as the control parameter of automatic driving vehicle.The data of artificial vehicle is excavated, is driven in conjunction with nobody The safety sailing vehicle, with car target, ensures the safety of automatic driving vehicle.

Brief description

Fig. 1 is method of the present invention schematic flow sheet.

Specific embodiment

As shown in figure 1, a kind of towards the automatic driving vehicle control parameter scaling method with car safety, include following step Rapid: (1) gathers pilot steering track of vehicle data: obtains, by unmanned plane, the video that pilot steering vehicle runs, passes through Video processing software carry out pilot steering track of vehicle data extract, track data include pilot steering vehicle before and after car with car number According to collection d_lAnd d_f；Wherein, data set d_lThe position of the front truck followed in the kth second including n-th pilot steering vehicleN-th The speed of the front truck that pilot steering vehicle was followed in the kth secondThe car of the front truck that n-th pilot steering vehicle was followed in the kth second Body lengthData set d_fIncluding n-th pilot steering vehicle in the position of kth secondN-th pilot steering vehicle exists The speed of kth secondThe duration t that n-th pilot steering vehicle runs_n, n=1,2 ..., n, n and k are the positive integer more than 0；

(2) according to the information in step (1), calculate the safety index in the kth second for n-th pilot steering vehicleFor:

${\mathrm{ttc}}_n^k=\frac{x_{l,n}^k-x_{f,n}^k-l_{l,n}^k}{v_{f,n}^k-v_{l,n}^k}$

(3) according to the information in step (1), automatic driving vehicle is iterated to calculate by automatic driving vehicle Controlling model Data set u_f；

$e_n^{k,i}=x_{l,n}^k-x_{f,n}^{*k,i}-t_{hw}^i{v}_{f,n}^{*k,i}$

$v_{f,n}^{*k+1,i}=v_{f,n}^{*k,i}+k_p^i{e}_n^{k,i}+k_d^i{\stackrel{\·}{e}}_n^{k,i}$

$x_{f,n}^{*k+1,i}=x_{f,n}^{*k,i}+v_{f,n}^{*k+1,i}*1$

Wherein, parameterFor i-th group of parameter value, i=1,2 ..., m, m areNumber；

For model parameter it isWhen calculate gained n-th automatic driving vehicle kth second and front truck interval error；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the position of kth second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle the kth second speed；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth position of+1 second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth speed of+1 second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle kth second and front truck interval error Derivative；

The initial value of iteration is set to: k=1,Data set u_fIncludingWith

(4) according to the data set u in step (3)_f, computation model parameter isWhen calculate gained n-th automatic driving car The kth second safety indexFor:

$ttc{*}_n^{k,i}=\frac{x_{l,n}^k-x_{f,n}^{*k,i}-l_{l,n}^k}{v_{f,n}^{*k,i}-v_{l,n}^k}$

(5) according to the information in step (1), step (2) and step (4), targeted security function y, the computing formula of y are set up As follows:

$y=y\left({\stackrel{\→}{\mathrm{\β}}}_i\right)=\underset{n=1}{\overset{n}{\σ}}\frac{\sqrt{\underset{k=1}{\overset{t_n}{\σ}}{(ttc{*}_n^{k,i}-{\mathrm{ttc}}_n^k)}^2}}{t_n}$

(6) according to the information in step (5), judge the parameter in automatic driving vehicle Controlling modelWhether all examine Survey, that is, whether i is equal to m, if so, then proceeds to step (7)；Otherwise, i=i+1, proceeds to step (3)；

(7) output makes the minimum parameter value of targeted security function yThat is:

$y\left({\stackrel{\→}{\mathrm{\β}}}^{*}\right)=min\{y\left({\stackrel{\→}{\mathrm{\β}}}_1\right),y\left({\stackrel{\→}{\mathrm{\β}}}_2\right),...,y\left({\stackrel{\→}{\mathrm{\β}}}_i\right),...,y\left({\stackrel{\→}{\mathrm{\β}}}_m\right)\}.$

A specific embodiment is given below.Using the present invention, automatic driving vehicle control parameter is demarcated, this calculation In example, for the sake of simplicity, only consider n=2 and m=2, but n and m can get arbitrary positive integer in actual applications.

Step 1, gathers pilot steering track of vehicle data: obtain regarding of pilot steering vehicle operation by unmanned plane Frequently, carry out pilot steering track of vehicle data by Video processing software to extract, before and after track data includes pilot steering vehicle Car is with car data collection d_lAnd d_f；Wherein, data set d_lThe position of the front truck followed in the kth second including n-th pilot steering vehicleThe speed of the front truck that n-th pilot steering vehicle was followed in the kth secondN-th pilot steering vehicle was followed in the kth second Front truck length of wagonData set d_fIncluding n-th pilot steering vehicle in the position of kth secondN-th artificial Drive the speed in the kth second for the vehicleThe duration t that n-th pilot steering vehicle runs_n, n=1,2 ..., n, take n=2, the k to be Positive integer more than 0,WithUnit be rice,Unit be rice,WithUnit be meter per second, t_nList Position is the second；

Wherein, data set d_lAs shown in the table:

The corresponding data of n=1 is:

The corresponding data of n=2 is:

Data set d_fAs shown in the table:

The corresponding data of n=1 is:

The corresponding data of n=2 is:

Step 2, according to the information in step 1, calculates the safety index in the kth second for n-th pilot steering vehicle For:

${\mathrm{ttc}}_n^k=\frac{x_{l,n}^k-x_{f,n}^k-l_{l,n}^k}{v_{f,n}^k-v_{l,n}^k}$

Computed information is:

The corresponding data of n=1 is:

The corresponding data of n=2 is:

Step 3, according to the information in step 1, iterates to calculate automatic driving car by automatic driving vehicle Controlling model Data set u_fIt may be assumed that

$e_n^{k,i}=x_{l,n}^k-x_{f,n}^{*k,i}-t_{hw}^i{v}_{f,n}^{*k,i}$

$v_{f,n}^{*k+1,i}=v_{f,n}^{*k,i}+k_p^i{e}_n^{k,i}+k_d^i{\stackrel{\·}{e}}_n^{k,i}$

$x_{f,n}^{*k+1,i}=x_{f,n}^{*k,i}+v_{f,n}^{*k+1,i}*1$

Wherein, parameterFor i-th group of parameter value, i=1,2 ..., m, take m=2,

For model parameter it isWhen calculate gained n-th automatic driving vehicle kth second and front truck interval error；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the position of kth second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle the kth second speed；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth position of+1 second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth speed of+1 second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle kth second and front truck interval error Derivative；

The initial value of iteration is set to: k=1,

Data set u_fIncludingWithWork as i=1,When calculate data be:

The corresponding data of n=1 is:

The corresponding data of n=2 is:

Step 4, according to the data set u in step 3_f, computation model parameter isWhen calculate n-th of gained unmanned Vehicle is in the safety index of kth secondFor:

$ttc{*}_n^{k,i}=\frac{x_{l,n}^k-x_{f,n}^{*k,i}-l_{l,n}^k}{v_{f,n}^{*k,i}-v_{l,n}^k}$

Computed information is:

The corresponding data of n=1 is:

The corresponding data of n=2 is:

Step 5, according to step 1, the information in step 2 and step 4, set up targeted security function y, the computing formula of y is such as Under:

$y=y\left({\stackrel{\→}{\mathrm{\β}}}_1\right)=\underset{n=1}{\overset{2}{\σ}}\frac{\sqrt{\underset{k=1}{\overset{t_n}{\σ}}{(ttc{*}_n^{k,1}-{\mathrm{ttc}}_n^k)}^2}}{t_n}=0.33$

Step 6, according to step 5, in information, judge the parameter in automatic driving vehicle Controlling modelWhether all examine Survey, that is, whether i is equal to m, now i=1, and m=2, i are not equal to m, i=i+1=2, proceed to step 3；Repeat step 3 to step 6, As i=2, calculate gainedNow i=m, proceeds to step 7；

Step 7, output makes the minimum parameter value of targeted security function yThat is:

$y\left({\stackrel{\→}{\mathrm{\β}}}^{*}\right)=min\{y\left({\stackrel{\→}{\mathrm{\β}}}_1\right),y\left({\stackrel{\→}{\mathrm{\β}}}_2\right)\}=min\{0.33,0.35\}$

${\stackrel{\→}{\mathrm{\β}}}^{*}={\stackrel{\→}{\mathrm{\β}}}_1=(0.4,0.45,0.2)$

Although the present invention is illustrated with regard to preferred implementation and has been described, it is understood by those skilled in the art that Without departing from scope defined by the claims of the present invention, variations and modifications can be carried out to the present invention.

Claims (1)

1. a kind of towards with car safety automatic driving vehicle control parameter scaling method it is characterised in that comprising the steps:

(1) gather pilot steering track of vehicle data: the video that pilot steering vehicle runs is obtained by unmanned plane, passes through Video processing software carry out pilot steering track of vehicle data extract, track data include pilot steering vehicle before and after car with car number According to collection d_lAnd d_f；Wherein, data set d_lThe position of the front truck followed in the kth second including n-th pilot steering vehicleN-th The speed of the front truck that pilot steering vehicle was followed in the kth secondThe car of the front truck that n-th pilot steering vehicle was followed in the kth second Body lengthData set d_fIncluding n-th pilot steering vehicle in the position of kth secondN-th pilot steering vehicle is The speed of k secondThe duration t that n-th pilot steering vehicle runs_n, n=1,2 ..., n, n and k are the positive integer more than 0；

(2) according to the information in step (1), calculate the safety index in the kth second for n-th pilot steering vehicleFor:

${\mathrm{ttc}}_n^k=\frac{x_{l,n}^k-x_{f,n}^k-l_{l,n}^k}{v_{f,n}^k-v_{l,n}^k}$

(3) according to the information in step (1), iterate to calculate automatic driving vehicle data by automatic driving vehicle Controlling model Collection u_f；

$e_n^{k,i}=x_{l,n}^k-x_{f,n}^{*k,i}-t_{hw}^i{v}_{f,n}^{*k,i}$

$v_{f,n}^{*k+1,i}=v_{f,n}^{*k,i}+k_p^i{e}_n^{k,i}+k_d^i{\stackrel{\·}{e}}_n^{k,i}$

$x_{f,n}^{*k+1,i}=x_{f,n}^{*k,i}+v_{f,n}^{*k+1,i}*1$

Wherein, parameterFor i-th group of parameter value, i=1,2 ..., m, m areNumber；

For model parameter it isWhen calculate gained n-th automatic driving vehicle kth second and front truck interval error；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the position of kth second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle the kth second speed；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth position of+1 second；

For model parameter it isWhen calculate gained n-th automatic driving vehicle in the kth speed of+1 second；

For model parameter it isWhen calculate n-th automatic driving vehicle leading in kth second and the interval error of front truck of gained Number；

The initial value of iteration is set to: k=1,Data set u_fIncludingWith(4) root According to the data set u in step (3)_f, computation model parameter isWhen calculate gained n-th automatic driving vehicle the kth second peace All referring to markFor:

$ttc{*}_n^{k,i}=\frac{x_{l,n}^k-x_{f,n}^{*k,i}-l_{l,n}^k}{v_{f,n}^{*k,i}-v_{l,n}^k}$

(5) according to the information in step (1), step (2) and step (4), set up targeted security function y, the computing formula of y is such as Under:

$y=y\left({\stackrel{\→}{\mathrm{\β}}}_i\right)=\underset{n=1}{\overset{n}{\σ}}\frac{\sqrt{\underset{k=1}{\overset{t_n}{\σ}}{(ttc{*}_n^{k,i}-{\mathrm{ttc}}_n^k)}^2}}{t_n}$

(6) according to the information in step (5), judge the parameter in automatic driving vehicle Controlling modelWhether all detect, i.e. i Whether it is equal to m, if so, then proceed to step (7)；Otherwise, i=i+1, proceeds to step (3)；

(7) output makes the minimum parameter value of targeted security function yThat is:

$y\left({\stackrel{\→}{\mathrm{\β}}}^{*}\right)=min\{y\left({\stackrel{\→}{\mathrm{\β}}}_1\right),y\left({\stackrel{\→}{\mathrm{\β}}}_2\right),...,y\left(\stackrel{\→}{{\mathrm{\β}}_i}\right),...,y\left({\stackrel{\→}{\mathrm{\β}}}_m\right)\}.$