CN101639978A - Method capable of dynamically partitioning traffic control subregion - Google Patents

Abstract

The invention discloses a method capable of dynamically partitioning traffic control subregion, including the following steps: acquisition of relevant data for associativity analysis; determining degree of association of adjacent crossings by utilizing the acquired data; further determining combination degree of association of multiple crossings; traversing all possible control subregion partitionschemes by adopting subregion partition layer diffusion algorithm and selecting the control subregion partition scheme with maximum performance index by comparation; and partitioning a new traffic control subregion according to the maximum performance index scheme obtained in real time. The invention provides the definition of the degree of association of adjacent crossings and analysis algorithmthereof and realizes effective combination of association key elements of the adjacent crossings; and the provided coordination control subregion partitioning method based on analysis of the degree of association considers various factors influencing the control subregion partition and realizes mensuration, standardization and systematization of the control subregion partition.

Description

The method in a kind of dynamic partitioning traffic control subarea

Technical field

The present invention relates to regional traffic and coordinate control, particularly relate to the method in a kind of dynamic partitioning traffic control subarea.

Background technology

When coordinating control of traffic signals is carried out in a bigger zone of scope, often need to be divided into several relatively independent parts, each part is carried out the control corresponding scheme according to traffic characteristics separately, and these relatively independent parts are called the control subarea.The reasonable division in control subarea will help carrying out control strategy flexibly, make the block of traffic characteristics difference great disparity all can obtain the Optimal Control effect, be the prerequisite that realizes effective traffic zone coordination control.Set up a cover integrated and efficient control subarea and divide index, realize measuring, standardization, systematization that the control subarea is divided, become the inherent requirement of control subarea partitioning technology research.

Factors such as the crossing traffic behavior that the foreign scholar divides influence control subarea, road section length, vehicle arrival rate are analyzed, and utilize controlling index, threshold value and the algorithm of the means antithetical phrase zoning branch of mathematical modeling to carry out correlative study.In recent years, domestic scholars is also obtaining some progress aspect the control subarea division methods research, has proposed the automatic division methods in control subarea based on " cycle principle ", " principle of flow ", " distance principle " and " saturation degree principle "; Set up the control subarea cycle of the rational change range that helps definite intersection signal cycle and divided index, and can provide the control subarea of certain quantitative basis apart from dividing index for the division of control subarea; The intelligence division of proposition by introducing similarity and setting up the cycle subarea of search model realization provided based on the subarea of genetic algorithm and divided the threshold parameter optimization method; And the notion of utilizing interconnected index between the crossing, dynamic subarea division methods based on the fuzzy clustering algorithm etc. has been proposed.Yet said method all will not influence the various factors of dividing in the control subarea and carry out effectively comprehensively, fail the distinct control subarea criteria for classifying of the system that makes.

Compare with other subarea division methods, divide model and have more scientific rationality by the factor that influences adjacent crossing correlativity being carried out control subarea that analysis-by-synthesis sets up, the present invention has provided the definition and the analytical algorithm thereof of the adjacent crossing degree of association, realized effectively comprehensive to adjacent crossing correlativity key element, the coordination control subarea division methods based on correlation analysis that proposes has been taken all factors into consideration the various factors that influence control subarea is divided, realized the measuring that the control subarea is divided, standardization and systematization are for solid foundation has been established in the research of regional coordination control technology method.

Summary of the invention

The present invention is directed to the characteristics of adjacent crossing relevance, taken all factors into consideration the various factors that influence control subarea is divided, propose the method in a kind of dynamic partitioning traffic control subarea, realized measuring, standardization and systematization that the control subarea is divided.

The present invention is achieved through the following technical solutions:

The method in a kind of dynamic partitioning traffic control subarea comprises the steps:

(1) acquisition of relevant data for associativity analysis: utilize distance measurement tools to gather adjacent crossing spacing, adjacent crossing spacing will be as the static action factor that influences relevance between the crossing; Gather the related wagon flow volume of traffic between adjacent crossing and predict the most relevance wagon flow vehicle increment of next period, as the maximum volume of traffic between the adjacent crossing, the maximum volume of traffic between the adjacent crossing will be as the dynamic action factor that influences relevance between the crossing with the related wagon flow volume of traffic between adjacent crossing and most relevance wagon flow vehicle increment sum; By communication line each intersection signal periodic transfer is arrived traffic coordinating control center;

(2) determining of the adjacent crossing degree of association: the adjacent crossing degree of association is made up of the periodic associated degree two parts of the road section traffic volume amount degree of association and crossing, wherein, the road section traffic volume amount degree of association reflects the maximum volume of traffic and the pairing influence of holding the adjacent crossing of the comparison relevance of the volume of traffic of adjacent crossing spacing between the adjacent crossing; The periodic associated degree in crossing reflects the influence of the relative deviation of signal period separately of adjacent crossing to adjacent crossing relevance;

(3) determining of the multi-intersection combination degree of association: the multi-intersection combination degree of association is made up of total road section traffic volume amount degree of association and total periodic associated degree two parts in crossing, wherein, total road section traffic volume amount degree of association is the concentrated reflection of each adjacent road section traffic volume amount relevance, and the periodic associated degree in total crossing then depends on the maximum relative deviation of each intersection signal between the cycle;

(4) coordinating the control subarea dynamically divides: adopt the subarea to divide layer broadcast algorithm and travel through all possible control subarea splitting scheme, than the control subarea splitting scheme of selecting the performance index maximum; The division principle in control subarea will be excellent with control subarea sum N less, and number equates under the situation in the control subarea, and is greatly excellent with regional total correlation degree V;

(5), carry out new traffic control subarea and divide according to the maximum performance index scheme that obtains in real time.

The method in above-mentioned dynamic partitioning traffic control subarea, in the step (1), image data comprises: adjacent crossing I _iTo crossing I _jSpacing L _{L (i → j)}, crossing I _iTo crossing I _jNumber of track-lines n _{L (i → j)}, next signal period crossing I _iTo crossing I _jMaximum volume of traffic N _{Max (i → j)}, crossing I _iWith crossing I _jSignal period.

The method in above-mentioned dynamic partitioning traffic control subarea in the step (1), is gathered adjacent crossing I by coil detection technique or video detection technology _iTo crossing I _jBetween related wagon flow volume of traffic N _{E (i → j)}, obtain adjacent crossing I by prediction _iTo crossing I _jBetween most relevance wagon flow vehicle increment N _{I (i → j)}, then adjacent crossing I _iTo crossing I _jBetween maximum volume of traffic N _{Max (i → j)}=N _{E (i → j)}+ N _{I (i → j)}Prediction most relevance wagon flow vehicle increment N _{I (i → j)}Be to open the bright moment as starting point, the most relevance wagon flow vehicle increment that prediction occurs on the highway section in the cycle at next signal, N with the let pass green light of related wagon flow phase place of crossing, upstream _{I (i → j)}Crossing, upstream by i → j directional correlation wagon flow is sailed rate q into _{In (i → j)}, the crossing, downstream rolls rate q away from _{Out (i → j)}With the let pass phase place split λ of related wagon flow of common signal cycle C, the crossing, upstream of upstream and downstream crossing _{I (i → j)}, the let pass phase place split λ of related wagon flow of crossing, downstream _{J (i → j)}And crossing, upstream let pass the phase differential O between the related wagon flow phase place of related wagon flow phase place and crossing, downstream that lets pass _{F (i → j)}Common decision.

In the method in above-mentioned dynamic partitioning traffic control subarea, work as O _{F (i → j)}〉=C * λ _{I (i → j)}, and O _{F (i → j)}+ C * λ _{J (i → j)}During≤C, N _{I (i → j)}=q _{In (i → j)}* C * λ _{I (i → j)}

The method in above-mentioned dynamic partitioning traffic control subarea, in the step (2), the road section length between the adjacent crossing is more little, number of track-lines more less, the volume of traffic is big more, the then described road section traffic volume amount degree of association is big more; Signal period between the adjacent crossing differs more little, and the periodic associated degree in then described crossing is big more.

The method in above-mentioned dynamic partitioning traffic control subarea, in the step (2), adjacent crossing I _iWith crossing I _jDegree of association Cor _{(i, j)}=Cor _{S (i, j)}+ Cor _{I (i, j)}, crossing I wherein _iWith crossing I _jRoad section traffic volume amount degree of association Cor _{S (i, j)}=max{Cor _{S (i → j)}, Cor _{S (j → i)}, crossing I _iTo crossing I _jThe road section traffic volume amount degree of association ${\mathrm{Cor}}_{S(i\→j)}=\frac{(N_{E(i\→j)}+N_{I(i\→j)})\×K_N}{N_{a(i\→j)}}\×K_{L_{1(i\→j)}}=\frac{(N_{E(i\→j)}+N_{I(i\→j)})\×L_v\×K_N}{n_{1(i\→j)}\×L_{1(i\→j)}}\×K_{L_{1(i\→j)}},$ Can hold volume of traffic N _{A (i → j)}=(n _{1 (i → j)}* L _{1 (i → j)})/L _v, L wherein _vBe average vehicle length, K _NBe the ratio amplification coefficient,

Be crossing I _iTo crossing I _jThe related penalty coefficient of the pairing road section traffic volume amount of track, direction highway section total length; ${\mathrm{Cor}}_{I(i,j)}=-\min \{\frac{C_{\max }\÷\mathrm{int}(C_{\max }\÷C_{\min })-C_{\min }}{C_{\min }},\frac{\mathrm{int}(C_{\max }\÷C_{\min }+1)\×C_{\min }-C_{\max }}{C_{\max }}\}\×K_C$ Be adjacent crossing I _iWith crossing I _jThe periodic associated degree in crossing, K _CBe the periodic associated weight coefficient of adjacent intersection signal, int is for rounding mathematical operation, and min is for getting little mathematical operation, C _MaxFor at crossing I _iWith crossing I _jSignal period among get big, C _MinFor at crossing I _iWith crossing I _jSignal period among get little.

The method in above-mentioned dynamic partitioning traffic control subarea, in the step (3), crossing I ₁, I ₂... I _mThe combination degree of association be ${\mathrm{Cor}}_{(I_1,I_2,...I_m)}={\mathrm{Cor}}_{S(I_1,I_2,...I_m)}+{\mathrm{Cor}}_{I(I_1,I_2,...I_m)},$ I wherein ₁, I ₂... I _mBe each crossing in the control subarea with m crossing, crossing I _iWith crossing I _j∈ { I ₁, I ₂... I _m;

Be crossing I ₁, I ₂... I _mTotal road section traffic volume amount degree of association, and have ${\mathrm{Cor}}_{S(I_1,I_2,...I_m)}=\underset{k=1}{\overset{n}{\mathrm{\Π}}}f\left({\mathrm{Cor}}_{S_k}\right),$ Wherein n is a related crossing logarithm, and Π calculates for connecting the multiplier student movement,

Be k to the road section traffic volume amount degree of association between the related crossing, $f\left({\mathrm{Cor}}_{S_k}\right)={\left(\min \right\{{\mathrm{Cor}}_{S_k},\mathrm{sgn}\left({\mathrm{Cor}}_{S_k}\right)\left\}\right)}^{\frac{1}{k}}$ Be road section traffic volume amount degree of association composite function, sgn is a sign function, Will by $\{{\mathrm{<figure-callout\; id="1"\; label="Cor\; S"\; filenames="A2009100422320027H1.png"\; state="\{\{state\}\}">Cor</figure-callout>}}_{S_{}},{\mathrm{Cor}}_{S_2},...,{\mathrm{Cor}}_{S_n}\}=\mathrm{sort}\{{\mathrm{Cor}}_{S(I_i,I_j)},...{\mathrm{Cor}}_{S(I_s,I_t)}\}$ Determine related crossing I _iWith crossing I _j, related crossing I _sWith crossing I _t∈ { I ₁, I ₂... I _m, sort is the ascending sort function, expression rearranges to the road section traffic volume amount degree of association between the related crossing n by from small to large order, compose successively again with ${\mathrm{Cor}}_{{\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="A2009100422320027H1.png"\; state="\{\{state\}\}">S</figure-callout>}}_1},{\mathrm{Cor}}_{S_2},...,{\mathrm{Cor}}_{S_n};{\mathrm{Cor}}_{I(I_1,I_2,...I_m)}$ Be crossing I ₁, I ₂... I _mThe periodic associated degree in total crossing, and have ${\mathrm{Cor}}_{I(I_1,I_2,...I_m)}=\min \left\{{\mathrm{Cor}}_{I(I_x,I_y)}\right|I_x,I_y\&Element;\{I_1,I_2,...,I_m\}\},$

Be crossing I _xWith crossing I _yBetween periodic associated degree, crossing I _xWith crossing I _yBe crossing I ₁, I ₂... I _mIn any two.

The method in above-mentioned dynamic partitioning traffic control subarea, in the step (4), the performance index of control subarea splitting scheme $\mathrm{PI}=-N^{k_N}+V,$ Wherein $V=\underset{1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{Cor}}_{(I_1,I_2,...I_m)},$ k _NThe weight coefficient of expression control subarea sum.

The method in above-mentioned dynamic partitioning traffic control subarea, in the step (4), subarea division layer broadcast algorithm step is as follows:

1) select a crossing the 1st layer as the 1st control subarea:

1.1) select and the 1st the 2nd layer of controlling the 1st layer of crossing that is associated, subarea as the 1st control subarea, calculate the combination degree of association of the 1st control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme,, then change next step over to and carry out the 3rd layer of diffusion in the 1st subarea if be false;

1.2) select and the 1st the 3rd layer of controlling the 2nd layer of crossing that is associated, subarea as the 1st control subarea, calculate the combination degree of association of the 1st control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme, if be false, then carry out the 4th layer of diffusion in the 1st subarea, by that analogy, behind the crossing that do not have to continue to spread, the 1st control subarea, changing step 2 over to) diffusion of carrying out next control subarea calculates;

2) in the set of remaining crossing, select a crossing the 1st layer as the 2nd control subarea;

2.1) select and the 2nd the 2nd layer of controlling the 1st layer of crossing that is associated, subarea as the 2nd control subarea, calculate the combination degree of association of the 2nd control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme,, then change next step over to and carry out the 3rd layer of diffusion in the 2nd subarea if be false;

2.2) select and the 2nd the 3rd layer of controlling the 2nd layer of crossing that is associated, subarea as the 2nd control subarea, calculate the combination degree of association of the 2nd control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme, if be false, then carry out the 4th layer of diffusion in the 2nd subarea, by that analogy, behind the crossing that do not have to continue to spread, the 2nd control subarea, change step 3) over to and carry out the diffusion in next control subarea and calculate;

3) in the set of remaining crossing, select a crossing the 1st layer, spread as the 3rd control subarea, by that analogy, up to all control subareas are traveled through.

The present invention compared with prior art has following advantage and effect:

The research of traffic control subarea division principle mainly also rests on static qualitative analysis aspect at present, does not form the subarea division methods that dynamic quantitative is analyzed as yet, and these have all limited the actual motion effect of regional coordination control system to a certain extent.The present invention takes all factors into consideration adjacent crossing spacing, the road section traffic volume amount, and the intersection signal timing parameter is to the Different Effects of adjacent crossing relevance power, defined the adjacent crossing degree of association and the multi-intersection combination degree of association, disaggregation space by definition control subarea splitting scheme, constraint condition and interpretational criteria, set up based on correlation analysis and coordinated control subarea division model, adopting the subarea to divide layer broadcast algorithm realizes controlling the assay of subarea splitting scheme, provide the complete control subarea of a cover and divided flow process, realize and existing effective combination of controlling the subarea division principle, further strengthen the scientific rationality of control subarea division methods.In addition, by embodiment as can be known, in same control subarea, each intersection signal cycle is all more approaching, the adjacent crossing degree of association is all bigger, and the adjacent crossing degree of association that is between the adjacent crossing in different controls subarea is all less, and the intersection signal cycle in difference control subarea differ bigger, these have all demonstrated fully control subarea division methods is effective.

Description of drawings

Fig. 1 a～Fig. 1 f is the adjacent crossing prediction of the highway section vehicle under six a kinds of unlike signal periodic Control increment graph respectively, figure by

The expression green light period,

Represent red signal interval.

Fig. 2 is that process flow diagram is divided in the control subarea.

Fig. 3 is that layer broadcast algorithm figure is divided in the subarea.

Fig. 4 is a control area road network structure condition diagram.

Fig. 5 is that figure is as a result divided in the control subarea in real time.

Embodiment

Below in conjunction with accompanying drawing concrete enforcement of the present invention is described further.

The method in the dynamic partitioning traffic control subarea in the present embodiment comprises the steps:

1, acquisition of relevant data for associativity analysis

Utilize distance measurement tools to gather adjacent crossing spacing, adjacent crossing spacing will be as the static action factor that influences relevance between the crossing; Gather the related wagon flow volume of traffic between adjacent crossing and predict the most relevance wagon flow vehicle increment of next period, as the maximum volume of traffic between the adjacent crossing, the maximum volume of traffic between the adjacent crossing will be as the dynamic action factor that influences relevance between the crossing with the related wagon flow volume of traffic between adjacent crossing and most relevance wagon flow vehicle increment sum; By communication line each intersection signal periodic transfer is arrived traffic coordinating control center.

(1) adjacent crossing spacing

Adjacent crossing spacing is a key factor of the adjacent crossing of decision correlativity power.On the one hand, if adjacent crossing spacing is long, the crossing, upstream is rolled fleet away from and can be come with the growth of the distance of travelling is discrete gradually, will present during crossing, approaching downstream to arrive state at random, coordinate control action and obviously weaken, show as weak relevant between the adjacent crossing; If adjacent crossing spacing is very short, the crossing, upstream is rolled away from when fleet arrives the crossing, downstream will keep good continuity, coordinate control action and obviously strengthen, and show as strong correlation between the adjacent crossing.On the other hand, the length of adjacent crossing spacing will determine the volume of traffic that holds between the crossing, if adjacent crossing spacing is very long, then there is the bigger volume of traffic that holds between the crossing, downstream road section queuing vehicle is difficult to influence the vehicle clearance of crossing, upstream, shows as weak relevant between the adjacent crossing; If adjacent crossing spacing is too short, then there is the very little volume of traffic that holds between the crossing, downstream road section queuing vehicle is easy to cause the crossing, upstream that traffic jam takes place, and shows as strong correlation between the adjacent crossing.

Adjacent crossing spacing can be gathered each adjacent crossing I by the method for field survey as a kind of static action index _iTo crossing I _jSpacing L _{1 (i → j)}, crossing I _iTo crossing I _jNumber of track-lines n _{1 (i → j)}

(2) road section traffic volume amount

The big young pathbreaker of road section traffic volume amount directly determines the congested traffic condition in highway section, is a principal element that influences adjacent crossing correlativity size in real time.When the road section traffic volume amount hour, will remain the bigger volume of traffic that holds between the crossing, the discreteness that wagon flow is travelled is stronger, has weak real-time correlativity between the adjacent crossing; When the road section traffic volume amount is big, will remain the less volume of traffic that holds between the crossing, the discreteness that wagon flow is travelled a little less than, have the hard real time correlativity between the adjacent crossing.

The road section traffic volume amount is as a kind of dynamic action index, gather the volume of traffic between adjacent crossing by coil detection technique or video detection technology, and predict the volume of traffic (Forecasting Methodology will be introduced in detail in part 2) between adjacent crossing of next period, thereby obtain next signal period crossing I _iTo crossing I _jMaximum volume of traffic N _{Max (i → j)}

(3) intersection signal timing parameter

Road section traffic volume amount and decision coordination control live effect can be effectively controlled in being provided with of intersection signal timing parameter (signal period, split and phase differential), are the another kind of principal elements of the real-time correlativity size in the adjacent crossing of decision.

1) signal period

Realize coordination signal controlling good between the adjacent crossing, requiring must keep between the adjacent crossing comparatively stable phase differential is that the equal signal period must be adopted in adjacent crossing.If the adjacent crossing signal period separately differs greatly, uneven, then be difficult to take into account the vehicle pass-through efficient of coordination control effect and each crossing self between the adjacent crossing, show as weak correlativity between the adjacent crossing at this moment; If the adjacent crossing signal period separately is close or neat, then can be by suitably adjusting to determine a common signal cycle, guarantee the vehicle pass-through efficient of coordination control effect and each crossing self between the adjacent crossing, show as strong correlation between the adjacent crossing at this moment.

2) split

The split that sails crossing, upstream, wagon flow place, highway section phase place into is poor with the split that rolls crossing, downstream, wagon flow place, highway section phase place away from, with the accumulation and the dissipation of road section traffic volume amount in decision a period of time, also be a key factor of the adjacent crossing of influence correlativity power.For example, when the split difference is timing, the road section traffic volume amount will progressively accumulate, and the correlativity between the adjacent crossing strengthens gradually; When the split difference when negative, the highway section accumulation volume of traffic will progressively dissipate, the correlativity between the adjacent crossing is attenuated to certain certain value gradually.

3) phase differential

Sail wagon flow place, highway section crossing, upstream phase place into and roll phase differential between the phase place of crossing, downstream, wagon flow place, highway section away from, with the maximum volume of traffic that determines to exist on the highway section in the signal period, also can produce certain influence to the size of the real-time correlativity in adjacent crossing.But different with the split difference, phase differential does not have build-up effect to the influence of adjacent crossing correlativity, and just brings into play snap at current demand signal in the cycle.

For the traffic control parameter, it mainly is the signal period of gathering each signalized intersections.

2, the adjacent crossing degree of association determines

The adjacent crossing degree of association is made up of the periodic associated degree two parts of the road section traffic volume amount degree of association and crossing, wherein, the road section traffic volume amount degree of association reflects the maximum volume of traffic and the pairing influence of holding the adjacent crossing of the comparison relevance of the volume of traffic of adjacent crossing spacing between the adjacent crossing; The periodic associated degree in crossing reflects the influence of the relative deviation of signal period separately of adjacent crossing to adjacent crossing relevance.

For from crossing, upstream I _iTo crossing, downstream I _jDirection (being called for short i → j direction), crossing I _iTo crossing I _jDegree of association Cor _{(i → j)}Can define calculating by following formula:

Cor _(i→j)＝Cor _S(i→j)+Cor _I(i，j)

${\mathrm{Cor}}_{S(i\&RightArrow;j)}=\frac{(N_{E(i\&RightArrow;j)}+N_{I(i\&RightArrow;j)})\&times;L_v\&times;K_N}{n_{1(i\&RightArrow;j)}\&times;L_{1(i\&RightArrow;j)}}\&times;K_{L_{1(i\&RightArrow;j)}}$

${\mathrm{Cor}}_{I(i,j)}=-\min \{\frac{C_{\max }\&divide;\mathrm{int}(C_{\max }\&divide;C_{\min })-C_{\min }}{C_{\min }},\frac{\mathrm{int}(C_{\max }\&divide;C_{\min }+1)\&times;C_{\min }-C_{\max }}{C_{\max }}\}\&times;K_C$

In the formula:

Cor _{S (i → j)}---the road section traffic volume amount degree of association of i → j direction;

Cor _{I (i → j)}---crossing I _iWith crossing I _jBetween periodic associated degree;

N _{E (i → j)}---the already present related wagon flow volume of traffic on i → j direction highway section, comprise queuing vehicle number and driving vehicle number, can obtain in real time by the road section traffic volume amount detecting device;

N _{I (i → j)}---the most relevance wagon flow vehicle increment that may occur in the next signal period on i → j direction highway section, need take all factors into consideration the road section traffic volume situation and the intersection signal controlled variable is carried out real-time estimate;

L _v---average vehicle length;

K _N---the ratio amplification coefficient;

n _{1 (i → j)}---the related wagon flow on i → j direction highway section takies number of track-lines;

L _{1 (i → j)}---i → track, j direction highway section total length;

---the related penalty coefficient of the pairing road section traffic volume amount of i → track, j direction highway section total length;

C _Max---at crossing I _iWith crossing I _jSignal period among get big;

C _Min---at crossing I _iWith crossing I _jSignal period among get little;

Int---round mathematical operation;

K _C---the periodic associated weight coefficient of adjacent intersection signal.

At this, the related wagon flow on so-called i → j direction highway section is meant decision i → j direction crossing I _iWith crossing I _jBetween the key flow of degree of association size, also is the major flow that needs to coordinate controlling Design on i → j direction usually.

For adjacent crossing I _iWith crossing I _jBetween degree of association Cor _{(i, j)}, may be defined as adjacent crossing I _iWith crossing I _jBetween the bi-directional combination degree of association

Promptly at the i → adjacent crossing of j direction degree of association Cor _{(i → j)}With the j → adjacent crossing of i direction degree of association Cor _{(j → i)}Among get big.

${\mathrm{Cor}}_{(i,j)}={\mathrm{Cor}}_{(i\&LeftRightArrow;j)}=\max \{{\mathrm{Cor}}_{(i\&RightArrow;j)},{\mathrm{Cor}}_{(j\&RightArrow;i)}\}=\max \{{\mathrm{Cor}}_{S(i\&RightArrow;j)},{\mathrm{Cor}}_{S(j-i)}\}+{\mathrm{Cor}}_{I(i,j)}={\mathrm{Cor}}_{S(i,j)}+{\mathrm{Cor}}_{I(i,j)}$

In the formula:

Cor _{S (i, j)}---crossing I _iWith crossing I _jBetween the road section traffic volume amount degree of association.

Most relevance wagon flow vehicle increment N _{I (i → j)}Be to open the bright moment as starting point, the most relevance wagon flow vehicle increment that prediction occurs on the highway section in the cycle at next signal, N with the let pass green light of related wagon flow phase place of crossing, upstream _{I (i → j)}To sail rate q into by the crossing, upstream of i → j directional correlation wagon flow _{In (i → j)}, the crossing, downstream rolls rate q away from _{Out (i → j)}With the let pass phase place split λ of related wagon flow of common signal cycle C, the crossing, upstream of upstream and downstream crossing _{I (i → j)}, the let pass phase place split λ of related wagon flow of crossing, downstream _{J (i → j)}And crossing, upstream let pass the phase differential O between the related wagon flow phase place of related wagon flow phase place and crossing, downstream that lets pass _{F (i → j)}Common decision.In addition, already present related wagon flow highway section queuing vehicle number N on i → j direction highway section _{Q (i → j)}In the time of greatly, need consider that crossing, downstream queuing vehicle launch characteristic influences the sluggishness that the highway section, upstream arrives vehicle ', at this moment N _{I (i → j)}Increase to some extent corresponding.

(1) works as N _{Q (i → j)}Smaller or equal to i → j direction highway section queuing vehicle threshold value N _{T (i → j)}The time, highway section vehicle prediction increment N _{I (i → j)}The highway section that depends primarily on related wagon flow is sailed, is rolled away from into the coherent signal timing parameter of rate and upstream and downstream crossing:

1. work as O _{F (i → j)}〉=C * λ _{I (i → j)}And O _{F (i → j)}+ C * λ _{J (i → j)}During≤C, shown in accompanying drawing 1a,

N _I(i→j)＝q _in(i→j)×C×λ _i(i→j)

2. work as O _{F (i → j)}〉=C * λ _{I (i → j)}And C＜O _{F (i → j)}+ C * λ _{J (i → j)}＜C+C * λ _{I (i → j)}The time, shown in accompanying drawing 1b,

N _I(i→j)＝max{q _in(i→j)×C×λ _i(i→j)-q _out(i→j)×(O _f(i→j)+C×λ _j(i→j)-C)，0}

3. work as O _{F (i → j)}+ C * λ _{J (i → j)}〉=C+C * λ _{I (i → j)}The time, shown in accompanying drawing 1c,

N _I(i→j)＝max{q _in(i→j)×C×λ _i(i→j)-q _out(i→j)×C×λ _i(i→j)，0}

4. work as O _{F (i → j)}+ C * λ _{J (i → j)}≤ C * λ _{I (i → j)}The time, shown in accompanying drawing 1d,

N _I(i→j)＝q _in(i→j)×O _f(i→j)

5. work as O _{F (i → j)}＜C * λ _{I (i → j)}And C * λ _{I (i → j)}＜O _{F (i → j)}+ C * λ _{J (i → j)}During≤C, shown in accompanying drawing 1e,

N _I(i→j)＝max{q _in(i→j)×O _f(i→j)，q _in(i→j)×C×λ _i(i→j)-q _out(i→j)×(C×λ _i(i→j)-O _f(i→j))}

6. work as O _{F (i → j)}＜C * λ _{I (i → j)}And O _{F (i → j)}+ C * λ _{J (i → j)}During＞C, shown in accompanying drawing 1f,

$N_{I(i\&RightArrow;j)}=\max \left\{\begin{array}{c}{q}_{\mathrm{in}(i\&RightArrow;j)}\&times;C\&CenterDot;{\mathrm{\&lambda;}}_{i(i\&RightArrow;j)}-q_{\mathrm{out}(i\&RightArrow;j)}\&times;(C\&times;{\mathrm{\&lambda;}}_{i(i\&RightArrow;j)}+C\&times;{\mathrm{\&lambda;}}_{j(i\&RightArrow;j)}-C),\\ q_{\mathrm{in}(i\&RightArrow;j)}\&times;O_{f(i\&RightArrow;j)}-q_{\mathrm{out}(i\&RightArrow;j)}\&times;(O_{f(i\&RightArrow;j)}+C\&times;{\mathrm{\&lambda;}}_{j(i\&RightArrow;j)}-C)\end{array}\right\}$

(2) work as N _{Q (i → j)}Greater than i → j direction highway section queuing vehicle threshold value N _{T (i → j)}The time, highway section vehicle prediction increment N _{I (i → j)}Estimation need also to consider that crossing, downstream queuing vehicle launch characteristic arrives the sluggishness influence of traveling state of vehicle to the highway section, upstream.Utilize the starting ripple and the parking wave propagation characteristic of traffic flow, crossing, the equivalent downstream green light that can derive the sluggish influence of a simulation highway section queuing vehicle opens bright time lag Δ t.The green light of crossing, downstream is opened bright moment postponement Δ t (green light of crossing, downstream remains unchanged the finish time), utilize above-mentioned classifying and analyzing method can calculate N _{Q (i → j)}＞N _{T (i → j)}Highway section vehicle prediction increment N under the situation _{I (i → j)}

N _{T (i → j)}To determine jointly by the factors such as starting wave propagation characteristic, road-section average travel speed and phase differential of traffic flow.

3, the multi-intersection combination degree of association determines

The multi-intersection combination degree of association is made up of total road section traffic volume amount degree of association and total periodic associated degree two parts in crossing, wherein, total road section traffic volume amount degree of association is the concentrated reflection of each adjacent road section traffic volume amount relevance, and the periodic associated degree in total crossing then depends on the maximum relative deviation of each intersection signal between the cycle.

The multi-intersection combination degree of association is one total correlation between one group of related crossing carried out the traffic parameter that quantification is described, and it is with the objective influence to the multi-intersection total correlation of the road section traffic volume operation conditions between a plurality of continuous crossings of concentrated expression and signal controlling demand difference.Based on above-mentioned analysis, and consider the Different Effects of factors such as crossing spacing, road section traffic volume amount and intersection signal timing parameter to the multi-intersection combination to the adjacent crossing degree of association.The multi-intersection combination degree of association

By total road section traffic volume amount degree of association With the periodic associated degree in total crossing Two parts are formed, wherein, and total road section traffic volume amount degree of association

Be the concentrated reflection of each adjacent road section traffic volume amount relevance, and have ${\mathrm{Cor}}_{S(I_1,I_2,...I_m)}=\underset{k=1}{\overset{n}{\mathrm{\&Pi;}}}f\left({\mathrm{Cor}}_{S_k}\right),$ The periodic associated degree in total crossing

Then depend on the maximum relative deviation between each intersection signal cycle C ${\mathrm{Cor}}_{I(I_1,I_2,...I_m)}=\min \left\{{\mathrm{Cor}}_{I(I_x,I_y)}\right|I_x,I_y\&Element;\{I_1,I_2,...,I_m\}\}.$

Practice shows, along with increasing of crossing and highway section number, the coordinating control of traffic signals effect will constantly weaken, and crossing spacing and road section traffic volume amount become the dominance factor that the control effect is coordinated in influence, and the road section traffic volume amount degree of association total between the multi-intersection should reduce gradually.Therefore, for one group of related crossing (I ₁, I ₂... I _m) between the combination degree of association

Can define calculating by following formula, and should guarantee simultaneously that generally speaking the adjacent crossing road section traffic volume amount degree of association is less than 1.

${\mathrm{Cor}}_{(I_1,I_2,...I_m)}={\mathrm{Cor}}_{S(I_1,I_2,...I_m)}+{\mathrm{Cor}}_{I(I_1,I_2,...I_m)}=\underset{k=1}{\overset{n}{\mathrm{\&Pi;}}}f\left({\mathrm{Cor}}_{S_k}\right)+\min \left\{{\mathrm{Cor}}_{I(I_x,I_y)}\right|I_x,I_y\&Element;\{I_1,I_2,...,I_m\}\}$

In the formula:

---related crossing (I ₁, I ₂... I _m) between total road section traffic volume amount degree of association;

---related crossing (I ₁, I ₂... I _m) between total periodic associated degree in crossing;

Π---connect the multiplier student movement and calculate;

N---related crossing logarithm, promptly related highway section number;

---k is to the road section traffic volume amount degree of association between the related crossing;

---road section traffic volume amount degree of association composite function, desirable $f\left({\mathrm{Cor}}_{S_k}\right)={\left(\min \right\{{\mathrm{Cor}}_{S_k},\mathrm{sgn}\left({\mathrm{Cor}}_{S_k}\right)\left\}\right)}^{\frac{1}{k}};$

---crossing I _xWith crossing I _yBetween periodic associated degree (crossing I _xWith crossing I _yCan not link to each other).

To determine by following formula:

$\{{\mathrm{<figure-callout\; id="1"\; label="Cor\; S"\; filenames="A2009100422320027H1.png"\; state="\{\{state\}\}">Cor</figure-callout>}}_{S_{}},{\mathrm{Cor}}_{S_2},...,{\mathrm{Cor}}_{S_n}\}=\mathrm{sort}\{{\mathrm{Cor}}_{S(I_i,I_j)},...{\mathrm{Cor}}_{S(I_S,I_t)}\}$

Wherein, sort is the ascending sort function, and expression rearranges to the road section traffic volume amount degree of association between the related crossing n by from small to large order, compose successively again with But connect the cumulative bad that multiplication demonstrates fully the adjacent road section traffic volume amount degree of association, get little computing and show that then the periodic associated degree in total crossing only depends on the maximum relative deviation of each intersection signal between the cycle.

4, coordinating the control subarea dynamically divides

(1) model is divided in the control subarea

For one by m signalized intersections (I ₁, I ₂..., I _m) highway section (R links to each other with the n bar ₁, R ₂..., R _n) control area formed, the disaggregation space of subarea splitting scheme is 2 ⁿKey (R separated in individual n position binary-coded character ₁R ₂R _n), R wherein _kGetting the 1 expression k bar highway section that links to each other is taken as association, gets 0 expression k bar highway section that links to each other and be taken as dereferenced.

The subarea splitting scheme need satisfy constraint condition: 1. as adjacent crossing I _fWith crossing I _gBetween the degree of association

Separate threshold value T smaller or equal to adjacent crossing _NISThe time, crossing I _fWith crossing I _gTo unconditionally be separated in different traffic controls subarea; 2. as adjacent crossing I _fWith crossing I _gBetween the degree of association

Merge threshold value T more than or equal to adjacent crossing _NICThe time, crossing I _fWith crossing I _gTo unconditionally be incorporated in same traffic control subarea; 3. each traffic control subarea combination degree of association must be separated threshold value T greater than multi-intersection _MIS, promptly ${\mathrm{Cor}}_{A_i}>T_{\mathrm{MIS}}$ (A _iBe the set of i traffic control crossing that the subarea comprises, i=1,2 ..., N), common T _NIS〉=T _MIS

The cardinal rule of estimating subarea splitting scheme quality is taken as usually: the division in subarea is excellent with control subarea sum N less; Number equates that the division in subarea is greatly excellent with regional total correlation degree V under the situation in the control subarea.Subarea splitting scheme performance index function is defined as $\mathrm{PI}=-N^{k_N}+V,$ Wherein $V=\underset{1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{Cor}}_{(I_1,I_2,...I_m)},$ k _NThe weight coefficient of expression control subarea sum, k _N〉=1.

(2) flow process is divided in the control subarea

Utilize above-mentioned coordination control subarea to divide model, in the disaggregation space, seek the optimum subarea splitting scheme that satisfies certain constraint condition.

Control subarea partiting step is as follows:

Step1. generate a kind of control subarea splitting scheme, whether check satisfies the basic constraint condition that subarea splitting scheme needs satisfy, satisfied then scheme is carried out initialization process, makes V=0, Cor (A ₁ ⁰)=0.5, S ₁=U={I ₁, I ₂..., I _n, A ₁ ⁰=Δ A ₁ ⁰={ I ₁ ⁰, I ₁ ⁰∈ S ₁, do not satisfy changing for the 5th step over to;

Step2. generate the sub-incidence matrix of subregion, further utilize the subarea to divide layer broadcast algorithm and carry out performance Index Calculation;

Step3. judge whether the subarea splitting scheme passes through, and if by would enter the 4th the step, by then change over to the 5th the step;

Step4. scheme and existing optimum division scheme is relatively more good and bad, what performance index were superior then is updated to the optimum division scheme with this scheme, changes for the 5th step over to;

Step5. judge whether to travel through all splitting schemes, if do not travel through all schemes, then turned back to for the 1st step, if traveled through then stop computing, export best subarea splitting scheme, the control subarea is divided and is finished.

Specifically can divide flow process with reference to control subarea as shown in Figure 2, in accompanying drawing 2, basic constraint condition is refered in particular to the degree of association smaller or equal to separating threshold value T _NISAdjacent crossing must unconditionally be separated in different controls subarea, the degrees of association more than or equal to merging threshold value T _NICAdjacent crossing must unconditionally be incorporated in same control subarea; It is the initial association degree in the 1st subarea $({\mathrm{Cor}}_{\left({A_1}^0\right)}={\mathrm{Cor}}_{\left({A_i}^0\right)}),$ Be the subarea combination degree of association value in single cross prong control subarea, its size is a positive pure decimal; S represents the set of optional crossing, S ₁It is the optional crossing set in the 1st subarea; A represents subarea contained crossing set, A ₁ ⁰The 0th layer (initial propagations layer) being the 1st subarea is with interior contained crossing set, Δ A ₁ ⁰Be the 0th layer of contained crossing set in the 1st subarea; I ₁ ⁰It is the initial crossing of primary election in the 1st subarea; The sub-incidence matrix of subregion is the characteristic parameter of descriptor zoning offshoot program, with the relevance size between the crossing in twos under each seed zone splitting scheme of concentrated expression.

The sub-incidence matrix CM of subregion ₀, CM ₁...,

Correspond respectively to character separate Key get 0,1 ..., 2 ^m-1 o'clock subarea dividing condition, wherein CM ₀=(0) _{N * n}, ${\mathrm{CM}}_{2^m-1}={\mathrm{CM}}_T,$ CM _TBe regional total correlation matrix, be shown below:

${\mathrm{CM}}_T={\left(\begin{array}{ccccc}{I}_{\left(1\right)}& I_{\left(2\right)}& ...& I_{(n-1)}& I_{\left(n\right)}\\ 0& {\mathrm{Cor}}_{\left(\mathrm{1,2}\right)}& ...& {\mathrm{Cor}}_{(1,n-1)}& {\mathrm{Cor}}_{(1,n)}\\ {\mathrm{Cor}}_{\left(\mathrm{1,2}\right)}& 0& ...& {\mathrm{Cor}}_{(2,n-1)}& {\mathrm{Cor}}_{(2,n)}\\ .& .& .& .& .\\ .& .& .& .& .\\ .& .& .& .& .\\ {\mathrm{Cor}}_{(1,n-1)}& {\mathrm{Cor}}_{(2,n-1)}& ...& 0& {\mathrm{Cor}}_{(n-1,n)}\\ {\mathrm{Cor}}_{(1,n)}& {\mathrm{Cor}}_{(2,n)}& ...& {\mathrm{Cor}}_{(n-1,n)}& 0\end{array}\right)}_{n\&times;n}\begin{array}{c}{I}_{\left(1\right)}\\ I_{\left(2\right)}\\ .\\ .\\ .\\ I_{(n-1)}\\ I_{\left(n\right)}\end{array}$

Layer broadcast algorithm divided in the subarea will carry out the validity judgement to each subarea splitting scheme by the method that successively spreads, and calculate corresponding subarea sum N and regional total correlation degree V.

It is as follows that layer broadcast algorithm algorithm steps is divided in the subarea:

1) selects a crossing the 1st layer as the 1st control subarea;

A) select and the 1st the 2nd layer of controlling the 1st layer of crossing that is associated, subarea, calculate the combination degree of association of the 1st control current contained crossing, subarea, judge that whether the combination degree of association is smaller or equal to multi-intersection separation threshold value T as the 1st control subarea _MIS, if set up, then the splitting scheme in this control subarea does not pass through, and enters the calculating of next control subarea splitting scheme, if be false, then change next step over to and carries out the 3rd layer of diffusion in the 1st subarea;

B) select and the 1st the 3rd layer of controlling the 2nd layer of crossing that is associated, subarea, calculate the combination degree of association of the 1st control current contained crossing, subarea, judge that whether the combination degree of association is smaller or equal to multi-intersection separation threshold value T as the 1st control subarea _MISIf set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme, if be false, then carry out the 4th layer of diffusion in the 1st subarea, by that analogy, behind the crossing that do not have to continue to spread, the 1st control subarea, changing step 2 over to) diffusion of carrying out next control subarea calculates;

2) in the set of remaining crossing, select a crossing the 1st layer as the 2nd control subarea;

A) select and the 2nd the 2nd layer of controlling the 1st layer of crossing that is associated, subarea, calculate the combination degree of association of the 2nd control current contained crossing, subarea, judge that whether the combination degree of association is smaller or equal to multi-intersection separation threshold value T as the 2nd control subarea _MIS, if set up, then the splitting scheme in this control subarea does not pass through, and enters the calculating of next control subarea splitting scheme, if be false, then change next step over to and carries out the 3rd layer of diffusion in the 2nd subarea;

B) select and the 2nd the 3rd layer of controlling the 2nd layer of crossing that is associated, subarea, calculate the combination degree of association of the 2nd control current contained crossing, subarea, judge that whether the combination degree of association is smaller or equal to multi-intersection separation threshold value T as the 2nd control subarea _MISIf set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme, if be false, then carry out the 4th layer of diffusion in the 2nd subarea, by that analogy, behind the crossing that do not have to continue to spread, the 2nd control subarea, change step 3) over to and carry out the diffusion in next control subarea and calculate;

3) in the set of remaining crossing, select a crossing the 1st layer, spread as the 3rd control subarea, by that analogy, up to all control subareas are traveled through.

Flow process as shown in Figure 3, wherein, i be the subarea sequence number (i=1,2 ...), j be subarea diffusion layer sequence number (j=0,1,2 ...).

5, the real-time crossing degree of association index that obtains according to previous step generates new traffic control subarea splitting scheme.New scheme feeds back to traffic coordinating control center, crossing in this zone is reconfigured, thereby reach the purpose that realizes coordinating control between the big crossing of the degree of association.

Below more in conjunction with the accompanying drawings 4 and embodiment the invention will be further described, but the scope of protection of present invention is not limited to the scope of embodiment statement.

With certain control area is example, and survey region is totally 12 crossings, and its crossing topological structure as shown in Figure 4.12 signalized intersections (I wherein ₁, I ₂..., I ₁₂) highway section (R links to each other with 17 ₁, R ₂..., R ₁₇) control area formed is the survey region of embodiment.

The first step: acquisition of relevant data for associativity analysis

Adjacent crossing spacing:, obtain this static action index through field survey.Its adjacent crossing spacing is as follows:

Highway section R ₁Have: L _{1 (1 → 2)}=L _{1 (2 → 1)}=387 meters, highway section R ₂Have: L _{1 (2 → 3)}=L _{1 (3 → 2)}=410 meters, highway section R ₃Have: L _{1 (3 → 4)}=L _{1 (4 → 3)}=410 meters, highway section R ₄Have: L _{1 (1 → 5)}=L _{1 (5 → 1)}=378 meters, highway section R ₅Have: L _{1 (2 → 6)}=L _{1 (6 → 2)}=460 meters, highway section R ₆Have: L _{1 (3 → 7)}=L _{1 (7 → 3)}=450 meters, highway section R ₇Have: L _{1 (4 → 8)}=L _{1 (8 → 4)}=470 meters, highway section R ₈Have: L _{1 (5 → 6)}=L _{1 (6 → 5)}=535 meters, highway section R ₉Have: L _{1 (6 → 7)}=L _{1 (7 → 6)}=440 meters, highway section R ₁₀Have: L _{1 (7 → 8)}=L _{1 (8 → 7)}=440 meters, highway section R ₁₁Have: L _{1 (5 → 9)}=L _{1 (9 → 5)}=440 meters, highway section R ₁₂Have: L _{1 (6 → 10)}=L _{1 (10 → 6)}=516 meters, highway section R ₁₃Have: L _{1 (7 → 11)}=L _{1 (11 → 7)}=428 meters, highway section R ₁₄Have: L _{1 (8 → 12)}=L _{1 (12 → 8)}=484 meters, highway section R ₁₅Have: L _{1 (9 → 10)}=L _{1 (10 → 9)}=390 meters, highway section R ₁₆Have: L _{1 (10 → 11)}=L _{1 (11 → 10)}=380 meters, highway section R ₁₇Have: L _{1 (11 → 12)}=L _{1 (12 → 11)}=465 meters.All roads are two-way Four-Lane Road in addition, also are n _{1 (i → j)}=2, wherein i, j are the subscript of any two adjacent crossings, this zone.

The road section traffic volume amount: this dynamic action index can obtain real-time data by magnetic test coil or Video Detection, further predicts the maximum volume of traffic that obtains between the adjacent crossing again.The highway section vehicle prediction increment N of related wagon flow _{1 (i → j)}Forecasting process is described below, for sailing rate q into by the crossing, upstream of i → j directional correlation wagon flow _{In (i → j)}=0.2veh/s, the phase place split λ of related wagon flow lets pass for common signal cycle C=100 second, crossing, upstream _{I (i → j)}=0.4, let pass the phase place split λ of related wagon flow the crossing, downstream _{J (i → j)}=0.35 and crossing, upstream let pass the phase differential O between the related wagon flow phase place of related wagon flow phase place and crossing, downstream that lets pass _{F (i → j)}=50 seconds, O is arranged _{F (i → j)}=50 〉=C * λ _{I (i → j)}=40 and O _{F (i → j)}+ C * λ _{J (i → j)}=85≤C=100, N at this moment _{I (i → j)}=q _{In (i → j)}* C * λ _{I (i → j)}=0.2 * 100 * 0.4=8veh.Be 32veh as if the volume of traffic that detects between adjacent crossing this moment, then next period crossing I _iWith crossing I _jThe maximum volume of traffic be N _{Max (i → j)}=32+8=40veh.

In the present embodiment, do not influencing under the general situation, adopting unified is N with the maximum volume of traffic between the adjacent crossing in the setting-up time _{Max (i → j)}=40 road section traffic volume amount data that conduct is further calculated.

The intersection signal timing parameter: signal period C is as the dynamic action factor that directly influences the degree of association, and its, Cycle Length of each crossing was as shown in table 1 a certain finish time period:

Table 1

Second step: the adjacent crossing degree of association is calculated

For the correlation parameter in the survey region, average vehicle length L _v=6m/veh, ratio amplification coefficient K _N=1.2; The related penalty coefficient of the pairing road section traffic volume amount of i → track, j direction highway section total length is $K_{L_{1(i\&RightArrow;j)}}=\frac{4000}{{L_{1(i\&RightArrow;j)}}^{3/2}}+1;$ The periodic associated weight coefficient K of adjacent intersection signal _C=0.2.

According to adjacent crossing degree of association computing formula, calculate in conjunction with the rapid real time data that collects of previous step:

For from crossing, upstream I ₁To crossing, downstream I ₂Direction (being called for short 1 → 2 direction), crossing I ₁With crossing I ₂Between the degree of association

${\mathrm{Cor}}_{(1\&RightArrow;2)}={\mathrm{Cor}}_{S(1\&RightArrow;2)}+{\mathrm{Cor}}_{I\left(\mathrm{1,2}\right)}$

$=\frac{(N_{E(1\&RightArrow;2)}+N_{I(1\&RightArrow;2)})\&times;L_v\&times;K_N}{n_{1(1\&RightArrow;2)}\&times;L_{1(1\&RightArrow;2)}}\&times;K_{L_{1(1\&RightArrow;2)}}$

$-\min \{\frac{C_{\max }\&divide;\mathrm{int}(C_{\max }\&divide;C_{\min })-C_{\min }}{C_{\min }},\frac{\mathrm{int}(C_{\max }\&divide;C_{\min }+1)\&times;C_{\min }-C_{\max }}{C_{\max }}\}\&times;K_C$

$=\frac{40\&times;6\&times;1.2}{2\&times;387}\&times;(\frac{4000}{{387}^{3/2}}+1)$

$-\min \{\frac{80\&divide;\mathrm{int}(80\&divide;70)-70}{70},\frac{\mathrm{int}(80\&divide;70+1)\&times;70-80}{80}\}\&times;0.2$

$=0.5676-0.1429\&times;0.2$

$=0.539$

Similarly can calculate Cor _{(2 → 1)}=Cor _{S (2 → 1)}+ Cor _{I (1,2)}=0.539,

Thereby ${\mathrm{Cor}}_{\left(\mathrm{1,2}\right)}={\mathrm{Cor}}_{(1\&LeftRightArrow;2)}=\max \{{\mathrm{Cor}}_{(1\&RightArrow;2)},{\mathrm{Cor}}_{(2\&RightArrow;1)}\}=0.539.$

Use the same method, can obtain the degree of association between all adjacent two crossings, this zone, as shown in table 2:

Table 2

The 3rd step: the multi-intersection combination degree of association is calculated

According to multi-intersection degree of association computing formula,, utilize the subarea to divide layer broadcast algorithm and obtain degree of association index under the combination of various crossings in conjunction with the resulting real-time adjacent crossing of second step degree of association.Threshold value T is separated in wherein adjacent crossing _NIS=0.2, adjacent crossing merges threshold value T _NIC=0.5, multi-intersection separates threshold value T _MIS=0.12, k _N=2.

For one by 12 signalized intersections (I ₁, I ₂..., I ₁₂) highway section (R links to each other with 17 ₁, R ₂..., R ₁₇) control area formed, the disaggregation space of subarea splitting scheme is 2 ¹⁷Key (R separated in individual 17 binary-coded characters ₁R ₂R ₁₇), R wherein _kGetting the 1 expression k bar highway section that links to each other is taken as association, gets 0 expression k bar highway section that links to each other and be taken as dereferenced.

In the zoning offshoot program searching process, carry out validity judgement and performance Index Calculation, guarantee the global optimum of subarea splitting scheme utilizing subarea division layer broadcast algorithm that each character in the disaggregation space is separated.

Alternatives one: with crossing I ₁, I ₂, I ₃, I ₅Assign to a control subarea, crossing I ₆, I ₇Assign to a control subarea, crossing I ₉, I ₁₀, I ₁₁Assign to a control subarea, crossing I ₄, I ₈, I ₁₂Assign to a control subarea.Utilize the subarea to divide layer broadcast algorithm and analyze, this scheme corresponding characters is separated Key value 11010010100001110, and it is as follows to its pairing control subarea splitting scheme analytic process that son is divided layer broadcast algorithm:

1. choose crossing I ₁The initial crossing I of primary election as the 1st control subarea ₁ ⁰, A ₁ ⁰=Δ A ₁ ⁰=I ₁ ⁰, V=0, ${\mathrm{Cor}}_{\left({A_1}^0\right)}=0.5,$ S ₁＝{I ₁，…，I ₁₇}；

2. carry out the 1st layer of diffusion in the 1st control subarea, S ₁ ¹={ I ₂..., I ₁₇, Δ A ₁ ¹={ I ₂, I ₅(R ₁=1, R ₄=1), A ₁ ¹={ I ₁, I ₂, I ₅, ${\mathrm{Cor}}_{\left({A_1}^1\right)}=0.2925>0.12;$

3. carry out the 2nd layer of diffusion in the 1st control subarea, S ₁ ²={ I ₃, I ₄, I ₆, I ₇..., I ₁₇, Δ A ₁ ²={ I ₃(R ₂=1), A ₁ ²={ I ₁, I ₂, I ₃, I ₅, ${\mathrm{Cor}}_{\left({A_1}^2\right)}=0.1857>0.12;$

4. carry out the 3rd layer of diffusion in the 1st control subarea, S ₁ ³={ I ₄, I ₆, I ₇..., I ₁₇, Δ A ₁ ³=Φ, A ₁ ³=A ₁ ², N=1, V=0.1857, S ₂={ I ₄, I ₆, I ₇..., I ₁₇, ${\mathrm{Cor}}_{\left({A_2}^0\right)}=0.5,$ A ₂ ⁰＝ΔA ₂ ⁰＝I ₂ ⁰＝I ₄；

Begin the layer diffusion computing in the 2nd control subarea, obtain A ₂ ³=A ₂ ²={ I ₄, I ₈, I ₁₂, ${\mathrm{Cor}}_{\left({A_2}^2\right)}=0.1245>0.12,$ N＝2，V＝0.3102，S ₃＝{I ₆，I ₇，I ₉，I ₁₀，I ₁₁}， ${\mathrm{Cor}}_{\left({A_3}^0\right)}=0.5,$ A ₃ ⁰＝ΔA ₃ ⁰＝I ₃ ⁰＝I ₆；

Begin the layer diffusion computing in the 3rd control subarea, obtain A ₃ ⁴=A ₃ ³={ I ₆, I ₇, ${\mathrm{Cor}}_{\left({A_3}^2\right)}=0.4691>0.12,$ N＝3，V＝0.7793，S ₄＝{I ₉，I ₁₀，I ₁₁}， ${\mathrm{Cor}}_{\left({A_4}^0\right)}=0.5,$ A ₃ ⁰＝ΔA ₃ ⁰＝I ₃ ⁰＝I ₉；

Begin the layer diffusion computing in the 4th control subarea, obtain A ₄ ³=A ₄ ²={ I ₉, I ₁₀, I ₁₁, ${\mathrm{Cor}}_{\left({A_4}^2\right)}=0.2286>0.12,$ N=4, V=1.0079, S ₅=Φ, book zoning offshoot program passes through, total performance index PI=-(4 ²)+1.0079=-14.9921.

Therefore, control subarea one (crossing I ₁, I ₂, I ₃, I ₅) the total correlation degree be: 0.1857;

Control subarea two (crossing I ₄, I ₈, I ₁₂) the total correlation degree be: 0.1245;

Control subarea three (crossing I ₆, I ₇) the total correlation degree be: 0.4691;

Control subarea four (crossing I ₉, I ₁₀, I ₁₁) the total correlation degree be: 0.2286;

Dividing the total correlation degree corresponding to the control subarea of this scheme is: 1.0078.

The performance index PI=-14.9921 that book zoning offshoot program is total.

Alternatives two: with crossing I ₁, I ₂, I ₃, I ₄Assign to a control subarea, crossing I ₅, I ₆, I ₇, I ₈, I ₁₂Assign to a control subarea, crossing I ₉, I ₁₀, I ₁₁Assign to a control subarea.Utilize the subarea to divide a layer broadcast algorithm and analyze, this scheme corresponding characters is separated Key value 11100001110001110, similarly, divides layer broadcast algorithm by the subarea and can obtain total correlation degree under this scheme:

Control subarea one (crossing I ₁, I ₂, I ₃, I ₄) the total correlation degree be: 0.1680;

Control subarea two (crossing I ₅, I ₆, I ₇, I ₈, I ₁₂) the total correlation degree be: 0.1465;

Control subarea three (crossing I ₉, I ₁₀, I ₁₁) the total correlation degree be: 0.2286;

Dividing the total correlation degree corresponding to the control subarea of this scheme is: 0.5431.

The performance index PI=-(3 that book zoning offshoot program is total ²)+0.5431=-8.4569.

The performance index numerical value that can compare alternatives two from above two alternativess is bigger, shows that just this control subarea splitting scheme is more excellent.

The 4th step: coordinate the control subarea and dynamically divide

According to the various assembled schemes of the rapid traversal of previous step, obtain the performance index value of various controls subarea splitting scheme, the scheme of its performance index maximum is: 12 crossings in the zone are divided into 3 controls subarea, wherein crossing I ₁, I ₂, I ₃, I ₄, I ₅Be divided into a control subarea, its total correlation degree

Be 0.1294; Crossing I ₆, I ₇, I ₈, I ₁₂Be divided into a control subarea, its total correlation degree

Be 0.2178; Crossing I ₉, I ₁₀, I ₁₁Be divided into a control subarea, its total correlation degree

Be 0.2286.Scheme total correlation degree Cor is 0.5758, and its performance index are PI=-(3 ²)+0.5758=-8.4242.

The 5th the step: according to the splitting scheme that previous step obtains, new control subarea splitting scheme as shown in Figure 5, wherein the crossing be divided into altogether three control subareas, crossing I ₁, I ₂, I ₃, I ₄, I ₅Be divided into a control subarea, crossing I ₆, I ₇, I ₈, I ₁₂Be divided into a control subarea, crossing I ₉, I ₁₀, I ₁₁Be divided into a control subarea.Traffic coordinating control center will carry out the control subarea division of a new round according to this controlling schemes.

Claims (9)

1, the method in a kind of dynamic partitioning traffic control subarea is characterized in that comprising the steps:

(1) acquisition of relevant data for associativity analysis: utilize distance measurement tools to gather adjacent crossing spacing, adjacent crossing spacing will be as the static action factor that influences relevance between the crossing; Gather the related wagon flow volume of traffic between adjacent crossing and predict the most relevance wagon flow vehicle increment of next period, as the maximum volume of traffic between the adjacent crossing, the maximum volume of traffic between the adjacent crossing will be as the dynamic action factor that influences relevance between the crossing with the related wagon flow volume of traffic between adjacent crossing and most relevance wagon flow vehicle increment sum; By communication line each intersection signal periodic transfer is arrived traffic coordinating control center;

(2) determining of the adjacent crossing degree of association: the adjacent crossing degree of association is made up of the periodic associated degree two parts of the road section traffic volume amount degree of association and crossing, wherein, the road section traffic volume amount degree of association reflects the maximum volume of traffic and the pairing influence of holding the adjacent crossing of the comparison relevance of the volume of traffic of adjacent crossing spacing between the adjacent crossing; The periodic associated degree in crossing reflects the influence of the relative deviation of signal period separately of adjacent crossing to adjacent crossing relevance;

(3) determining of the multi-intersection combination degree of association: the multi-intersection combination degree of association is made up of total road section traffic volume amount degree of association and total periodic associated degree two parts in crossing, wherein, total road section traffic volume amount degree of association is the concentrated reflection of each adjacent road section traffic volume amount relevance, and the periodic associated degree in total crossing then depends on the maximum relative deviation of each intersection signal between the cycle;

(4) coordinating the control subarea dynamically divides: adopt the subarea to divide layer broadcast algorithm and travel through all possible control subarea splitting scheme, than the control subarea splitting scheme of selecting the performance index maximum; The division principle in control subarea will be excellent with control subarea sum N less, and number equates under the situation in the control subarea, and is greatly excellent with regional total correlation degree V;

(5), carry out new traffic control subarea and divide according to the maximum performance index scheme that obtains in real time.

2, the method in dynamic partitioning traffic control according to claim 1 subarea is characterized in that image data comprises in the step (1): adjacent crossing I _iTo crossing I _jSpacing L _{I (i → j)}, crossing I _iTo crossing I _jNumber of track-lines n _{I (i → j)}, next signal period crossing I _iTo crossing I _jMaximum volume of traffic N _{Max (i → j)}, crossing I _iWith crossing I _jSignal period.

3, the method in dynamic partitioning traffic control according to claim 1 subarea is characterized in that gathering adjacent crossing I by coil detection technique or video detection technology in the step (1) _iTo crossing I _jBetween related wagon flow volume of traffic N _{E (i → j)}, obtain adjacent crossing I by prediction _iTo crossing I _jBetween most relevance wagon flow vehicle increment N _{I (i → j)}, then adjacent crossing I _iTo crossing I _jBetween maximum volume of traffic N _{Max (i → j)}=N _{E (i → j)}+ N _{I (i → j)}Prediction most relevance wagon flow vehicle increment N _{I (i → j)}Be to open the bright moment as starting point, the most relevance wagon flow vehicle increment that prediction occurs on the highway section in the cycle at next signal, N with the let pass green light of related wagon flow phase place of crossing, upstream _{I (i → j)}Crossing, upstream by i → j directional correlation wagon flow is sailed rate q into _{In (i → j)}, the crossing, downstream rolls rate q away from _{Out (i → j)}With the let pass phase place split λ of related wagon flow of common signal cycle C, the crossing, upstream of upstream and downstream crossing _{I (i → j)}, the let pass phase place split λ of related wagon flow of crossing, downstream _{J (i → j)}And crossing, upstream let pass the phase differential O between the related wagon flow phase place of related wagon flow phase place and crossing, downstream that lets pass _{F (i → j)}Common decision.

4, the method in dynamic partitioning traffic control according to claim 3 subarea is characterized in that working as O _{F (i → j)}〉=C * λ _{I (i → j)}, and O _{F (i → j)}+ C * λ _{J (i → j)}During≤C, N _{I (i → j)}=q _{In (i → j)}* C * λ _{I (i → j)}

5, the method in dynamic partitioning traffic control according to claim 1 subarea, it is characterized in that the road section length between the adjacent crossing is more little in the step (2), number of track-lines more less, the volume of traffic is big more, the then described road section traffic volume amount degree of association is big more; Signal period between the adjacent crossing differs more little, and the periodic associated degree in then described crossing is big more.

6, the method in dynamic partitioning traffic control according to claim 1 subarea is characterized in that in the step (2) adjacent crossing I _iWith crossing I _jDegree of association Cor _{(i, j)}=Cor _{S (i, j)}+ Cor _{I (i, j)}, crossing I wherein _iWith crossing I _jRoad section traffic volume amount degree of association Cor _{S (i, j)}=max{Cor _{S (i → j)}, Cor _{S (j → i)}, crossing I _iTo crossing I _jThe road section traffic volume amount degree of association ${\mathrm{Cor}}_{S(i\&RightArrow;j)}=\frac{(N_{E(i\&RightArrow;j)}+N_{I(i\&RightArrow;j)})}{N_{a(i\&RightArrow;j)}}\&times;K_{L_{1(i\&RightArrow;j)}}=\frac{(N_{E(i\&RightArrow;j)}+N_{I(i\&RightArrow;j)})\&times;L_v\&times;K_N}{n_{1(i\&RightArrow;j)}\&times;L_{1(i\&RightArrow;j)}}\&times;K_{L_{1(i\&RightArrow;j)}},$ Can hold volume of traffic N _{A (i → j)}=(n _{1 (i → j)}* L _{1 (i → j)})/L _v, L wherein _vBe average vehicle length, K _NBe the ratio amplification coefficient,

Be crossing I _iTo crossing I _jThe related penalty coefficient of the pairing road section traffic volume amount of track, direction highway section total length; ${\mathrm{Cor}}_{I(i,j)}=-\min \{\frac{C_{\max }\&divide;\mathrm{int}(C_{\max }\&divide;C_{\min }){-C}_{\min }}{C_{\min }},\frac{\mathrm{int}(C_{\max }\&divide;C_{\min }+1)\&times;C_{\min }-C_{\max }}{C_{\max }}\}\&times;K_C$ Be adjacent crossing I _iWith crossing I _jThe periodic associated degree in crossing, K _CBe the periodic associated weight coefficient of adjacent intersection signal, int is for rounding mathematical operation, and min is for getting little mathematical operation, C _MaxFor at crossing I _iWith crossing I _jSignal period among get big, C _MinFor at crossing I _iWith crossing I _jSignal period among get little.

7, the method in dynamic partitioning traffic control according to claim 1 subarea is characterized in that in the step (3) crossing I ₁, I ₂... I _mThe combination degree of association be ${\mathrm{Cor}}_{(I_1,I_2,...I_m)}={\mathrm{Cor}}_{S(I_1,I_2,...I_m)}+{\mathrm{Cor}}_{I(I_1,I_2,...I_m)},$ I wherein ₁, I ₂... I _mBe each crossing in the control subarea with m crossing, crossing I _iWith crossing I _j∈ { I ₁, I ₂... I _m; Be crossing I ₁, I ₂... I _mTotal road section traffic volume amount degree of association, and have ${\mathrm{Cor}}_{S(I_1,I_2,...I_m)}=\underset{k=1}{\overset{n}{\mathrm{\&Pi;}}}f\left({\mathrm{Cor}}_{S_k}\right),$ Wherein n is a related crossing logarithm, and ∏ calculates for connecting the multiplier student movement,

Be k to the road section traffic volume amount degree of association between the related crossing, $f\left({\mathrm{Cor}}_{S_k}\right)={\left(\min \right\{{\mathrm{Cor}}_{S_k},\mathrm{sgn}\left({\mathrm{Cor}}_{S_k}\right)\left\}\right)}^{\frac{1}{k}}$ Be road section traffic volume amount degree of association composite function, sgn is a sign function,

Will by $\{{\mathrm{Cor}}_{S_1},{\mathrm{Cor}}_{S_2},...,{\mathrm{Cor}}_{S_n}\}=\mathrm{sort}\{{\mathrm{Cor}}_{S(I_i,I_j)},...{\mathrm{Cor}}_{S(I_s,I_t)}\}$ Determine related crossing I _iWith crossing I _j, related crossing I _sWith crossing I _t∈ { I ₁, I ₂... I _m, sort is the ascending sort function, expression rearranges to the road section traffic volume amount degree of association between the related crossing n by from small to large order, compose successively again with

Be crossing I ₁, I ₂... I _mThe periodic associated degree in total crossing, and have ${\mathrm{Cor}}_{I(I_1,I_2,...I_m)}=\min \left\{{\mathrm{Cor}}_{I(I_x,I_y)}\right|I_x,I_y\&Element;\{I_1,I_2,...,I_m\}\},$

Be crossing I _xWith crossing I _yBetween periodic associated degree, crossing I _xWith crossing I _yBe crossing I ₁, I ₂... I _mIn any two.

8, the method in dynamic partitioning traffic control according to claim 1 subarea is characterized in that in the step (4), the performance index of control subarea splitting scheme $\mathrm{PI}=-N^{k_N}+V,$ Wherein $V=\underset{1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{Cor}}_{(I_1,I_2,...I_m)},$ k _NThe weight coefficient of expression control subarea sum.

9, the method in dynamic partitioning traffic control according to claim 1 subarea is characterized in that in the step (4), and it is as follows that layer broadcast algorithm step is divided in the subarea:

1) select a crossing the 1st layer as the 1st control subarea:

1.1) select and the 1st the 2nd layer of controlling the 1st layer of crossing that is associated, subarea as the 1st control subarea, calculate the combination degree of association of the 1st control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme,, then change next step over to and carry out the 3rd layer of diffusion in the 1st subarea if be false;

1.2) select and the 1st the 3rd layer of controlling the 2nd layer of crossing that is associated, subarea as the 1st control subarea, calculate the combination degree of association of the 1st control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme, if be false, then carry out the 4th layer of diffusion in the 1st subarea, by that analogy, behind the crossing that do not have to continue to spread, the 1st control subarea, changing step 2 over to) diffusion of carrying out next control subarea calculates;

2) in the set of remaining crossing, select a crossing the 1st layer as the 2nd control subarea;

2.1) select and the 2nd the 2nd layer of controlling the 1st layer of crossing that is associated, subarea as the 2nd control subarea, calculate the combination degree of association of the 2nd control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme,, then change next step over to and carry out the 3rd layer of diffusion in the 2nd subarea if be false;

2.2) select and the 2nd the 3rd layer of controlling the 2nd layer of crossing that is associated, subarea as the 2nd control subarea, calculate the combination degree of association of the 2nd control current contained crossing, subarea, judge whether the combination degree of association separates threshold value smaller or equal to multi-intersection, if set up, then the splitting scheme in this control subarea does not pass through, enter the calculating of next control subarea splitting scheme, if be false, then carry out the 4th layer of diffusion in the 2nd subarea, by that analogy, behind the crossing that do not have to continue to spread, the 2nd control subarea, change step 3) over to and carry out the diffusion in next control subarea and calculate;

3) in the set of remaining crossing, select a crossing the 1st layer, spread as the 3rd control subarea, by that analogy, up to all control subareas are traveled through.

Images