CN1734238A - Two-stage multi-path optimization method for centrally controlled vehicle navigation system - Google Patents

Abstract

This invention discloses a carrying guidance system in the route optimum domain, which comprises the following steps: establishing a biparametric path unit standard dynamic file; considering the path rational constrained and joint failure constrained; using the method of heuristics weight to establish the optional path set off line; improving A* heuristics function to increase the efficiency of optional path set which satisfies the customer rational constrained and joint failure constrained; Off-line optional path set coded storing and backtrack on line; screening of on-line dynamic path, supplementary searching and multiple path issue.

Description

Two-stage multi-path optimization method for centrally controlled vehicle navigation system

technical field

The invention relates to the field of route optimization of vehicle navigation systems. Based on unimpeded reliability analysis, an effective algorithm for two-stage dynamic route optimization for centrally controlled vehicle navigation system blocking risk avoidance and cluster risk avoidance is designed and implemented.

Background technique

Vehicle Navigation System (VIS), as one of the applications of ITS, not only provides users with better route information services, but also helps reduce traffic jams, shorten travel time and save energy. Therefore, it has been widely used in recent years. Optimal route optimization is an important key technology in vehicle automatic navigation system. Our research in this area is still in its infancy. According to the source of information based on route optimization, vehicle automatic navigation systems can be divided into dynamic navigation and static navigation. Dynamic navigation optimizes routes based on dynamic real-time information, and static navigation optimizes routes based on historical static information. In addition, according to the different execution units of path calculation, the vehicle automatic navigation system can be divided into central control type and decentralized control type vehicle navigation system. The central control path calculation is completed by the computer in the information center, which has strong computing power and is conducive to the realization of the overall optimization goal of the induction system. Decentralized control path calculation is completed by the on-board computer, and the calculation capacity is limited. The present invention performs route optimization design for a centrally controlled dynamic vehicle navigation system.

At present, vehicle navigation systems mostly perform single-path calculation and distribution with the goal of user optimization. With the gradual increase and popularization of navigation vehicles in the road network, this user-optimized single-path induction is likely to cause induced vehicles to cluster on the same path during peak hours, resulting in congestion due to induction. This is not only detrimental to the function of the road network system, but also the benefits of navigation users will not be guaranteed in the end. However, the commonly used route optimization algorithms that take into account both user optimization and system optimization, such as the fix point method, require a large number of iterations and take too long to be used in real-time navigation. Multi-path induced information release can effectively avoid induced clustering phenomenon, and users can choose a satisfactory path according to their own preferences, so as to achieve the coordination between user optimization and system optimization. Therefore, multi-path guidance will become the mainstream technology of dynamic vehicle navigation in the future.

However, in the current dynamic multi-path optimization, there are several problems in the following aspects: 1) The acquisition of high-precision real-time information. At present, the real-time information forecasting technology at home and abroad is not yet mature, and the forecasting accuracy needs to be improved. 2) The real-time problem of path calculation. Using traditional path algorithms such as Dijkstra method and A* algorithm for path calculation, the calculation time increases exponentially or non-linearly with the increase of road network scale. Although the computing power of the central control computer is powerful, with the popularization of navigation vehicles in the road network, when a large number of vehicles send navigation requests at the same time during peak hours, the calculation delay will still be caused by excessive calculation. The multi-path calculation makes it more difficult to realize the real-time calculation of the path. 3) Constrained multipath computing problem. Although multi-path planning can effectively avoid the clustering phenomenon, if the detour is too far, the user will lose trust in the navigation information and will not obey the guidance of the center, thereby reducing the system efficiency of the guidance of the center. Multipath computing therefore involves a constrained path optimization problem. Studying its efficient algorithm is the key to realize dynamic real-time navigation.

In view of the above problems, how to provide an efficient and accurate multi-path optimization algorithm and implementation technology that is efficient and accurate, meets user preferences, and takes into account the optimization objectives of the road network system is an urgent need to solve for the central control vehicle navigation system. The problem.

Contents of the invention

Based on the above analysis, the present invention designs a two-stage center-controlled vehicle navigation mechanism and algorithm for setting up and storing offline (offline) quasi-dynamic alternative route sets, and online (online) real-time path screening and multi-path release. (see accompanying drawing 1) with the help of reliability analysis, by refining historical information, set up the quasi-dynamic road attribute file with average transit time and unimpeded reliability as double parameter; Based on this quasi-dynamic road attribute file, utilize heuristic weighting The method invented a method for establishing a partially overlapping alternative path set, which satisfies user preference constraints and common failure constraints; according to online dynamic information, multiple feasible paths can be screened out from these alternative paths in real time to induce information release . The invention improves the rationality and online validity of the alternative paths based on historical information because the possibility of road blockage and the possibility of common failure (blockage) of alternative paths are considered. At the same time, the workload of online path search is reduced by off-line alternative road set calculation. The characteristic of this algorithm is that it does not need to rely too much on dynamic information, has a fast real-time response speed, and has a linear relationship between the calculation time and the scale of the road network and the number of induced vehicles. The invention can not only reduce the risk of travel delay, but also reduce the risk of clustering, and is beneficial to realize the coordination of system optimization and user optimization.

Technical thought feature of the present invention is:

1. The establishment of the quasi-dynamic attribute file of the road unit with two parameters.

2. Consider path rationality constraints and common failure constraints, and use the heuristic weighting method to establish a set of alternative paths offline.

3. Improve the heuristic function of A* to improve the efficiency of establishing alternative road sets.

4. Offline alternative road set coded storage and online backtracking.

5. Online dynamic path screening, supplementary search and multi-path publishing.

The following is a detailed description of the technical idea features and specific solutions of the present invention.

1. The first stage: offline stage

(1) Calibrate the road unit attribute file with two parameters

Divide a day into t typical periods, t>=3; the following operations are performed for any one of the typical periods, and the operation methods for other periods are the same;

The historical information is processed and refined into two parameter forms to serve the navigation algorithm: one is the average transit time of road units in each typical time period, and the other is the unimpeded reliability of road units in each typical time period.

Road units generally include road segments and intersections. The existing technology is used to transform the road network into an arc weight network, that is, each turn at the intersection is represented by a virtual road section, and the delay time of each turn is the passing time of the virtual road section. In the road network where the intersection is not seriously congested, the delay at the intersection may not be considered, and only the passage time of the road section may be considered. Thus, road network road units may only refer to road segment units.

For each road unit, the average passing time of each typical time period and the two parameters of unimpeded reliability are calibrated, and the average passing time of the road unit is established, using the following method:

For a road unit with traffic flow history data of more than one month, the average passing time of road unit i under normal conditions (no accidents, disasters, etc.) in a certain period of time is calculated by statistical methods, and this average passing time is the road unit i The weight W _i of i; for a road segment without traffic flow history data of more than one month, the weight W _i of road unit i can be divided by the length of the road segment divided by the design speed of the road segment.

There are two processes to determine the unimpeded reliability of road units:

a. Make a binary state assumption for the road unit: the state of the road unit is divided into two types: smooth and blocked; the boundary to distinguish the smooth and blocked road section is the speed threshold of the road section unit set in advance, and if it is greater than this value, it is regarded as unblocked , less than this value is regarded as blocking; and the boundary to distinguish between smooth and blocked intersections is the intersection delay threshold set in advance; less than this value is regarded as smooth, and greater than this value is regarded as blocked, the above threshold can be According to the "Ministry of Public Security's Opinions on the Implementation of the National Urban Road Traffic Management "Smooth Project"" to formulate smooth traffic standards;

b. Determine the unimpeded reliability of the road unit. The unimpeded reliability of the road unit can be defined as the probability that the road unit is unblocked or not blocked within a specified time period; its determination method is as follows:

Collect traffic flow data of more than 3 months for a road unit, and use the following formula to calculate the approximate value of the unimpeded reliability r _i of road unit i:

The process of double-parameter calibration of road units is shown in Figure 2.

(2) Search the set of alternative paths

Theoretically, it is possible to enumerate all the optional paths between any pair of points, compare their transit time, path reliability, and compare the road set reliability of various possible path combinations, so as to determine the best solution. However, this enumeration method undoubtedly has a huge amount of calculation.

In fact, according to the theory of system reliability, avoiding alternative paths passing through units with less reliability and avoiding overlapping units among alternative paths, especially overlapping units with less reliability, can effectively improve the reliability of alternative paths. and the reliability of the set of alternative paths. Because once an overlapping unit is blocked, all paths that share this unit will be blocked. Although the reliability of the candidate road set is relatively high when all the candidate routes have no overlapping road units, it is difficult to find enough reasonably non-overlapping road units between a origin-destination pair due to the limitation of detour time. path, for this purpose, it is necessary to establish a set of alternative paths where some units overlap.

In order to solve the problem that the enumeration method establishes an excessive amount of calculation for the alternative road set, the present invention proposes a constrained multi-path heuristic algorithm, the algorithm idea is based on the following fact: in the shortest path algorithm, if the road section i has a larger , then it will have a greater possibility not to be included in the shortest path P _{n, 0} of a certain origin-destination pair n. If W _i = ∞, then road segment _i will never appear in Pn _{, 0} in.

In this algorithm, the shortest path with average transit time under normal conditions will be used as the first alternative path, and then the road units appearing on the first alternative path and other high-risk units in the road network will be weighted, and the unit reliability The lower the value, the greater the weighting. The shortest path is recalculated after weighting, so that the road units that have appeared in the first alternative path, especially those with low reliability (high risk of blockage) will be more likely to no longer appear in the second alternative path superior. In the algorithm, the weighting process is mainly to avoid overlapping of alternative paths on high blocking risk units as much as possible, so as to reduce the possibility of common failure. If the obtained second path satisfies the reasonable path constraints, it is reserved as an alternative path, otherwise, the weighted range is reduced and the alternative path is recalculated. Repeat the similar weighting and reasonable path constraint inspection process, and continuously check the reliability and number of candidate road sets until the reliability or number of candidate road sets meets the online optional requirements.

The algorithm is as follows:

1) Let Q be the total number of origin-destination pairs in the entire road network. For all origin-destination pairs, initialize its candidate road set S _n = φ, n represents the nth origin-destination pair in Q, n=1, 2, 3...Q;

Taking the passing time of each road unit under normal conditions as the right of way, use the traditional Dijkstra algorithm to calculate the shortest path P _{n, 1} for all point pairs, and then calculate its weight sum as L _{n, 1} , which is the passage of the path P _{n, 1} Time; store P _{n, 1} in the candidate route set S _n of the nth origin-destination pair as the first candidate route;

Initialize the number of alternative paths K _n =1 for all origin-destination pairs;

Let n=1, that is, calculate the first origin-destination pair;

2) For the nth start-destination pair, the number of initialization iterations m=0;

3) For road units with low reliability on the road network (units with r _i <0.5-0.9 are recommended to be road units with low reliability, the lower limit is taken for cities that are more crowded in daily life, and the upper limit is taken for cities with relatively smooth daily life) and S _n Add weight Δw _i to each road unit on the existing path in , so that the new round of road unit weight W _i ' is:

W _i '＝W _i +ΔW _i ＝W _i + α ^m (1-r _i ) ^q W ₀ (2)

In the formula, W _i is the unit average transit time right of way, when m=0, q=0, otherwise q=1; α is the relaxation factor, 0<α<1, the larger α is, the road unit is included in the new path The smaller the possibility in the path, on the contrary, the smaller the α, the greater the possibility that the unit will be included in the new path; W ₀ is a large positive number, it is recommended to take W ₀ =1.5L _n,1 ~ 3L _n,1 ;

4) Set the number of alternative paths in the new round K _n ′=K _n +1, the number of iterations in the new round m′=m+1, m=m′, take W _i ' as the right of way, and use the existing A* algorithm Calculate the shortest path P _{n, Kn} , K _n =K' _n , restore the road weight to W _i , recalculate L _{n, Kn} , L _{n, Kn} is the sum of the road weights W _i on P _{n, Kn} ;

5) If P _{n, Kn} violate the reasonable path constraints L _{n, Kn} <βL _{n, 1} , then K _n ′=K _n -1, go to 3); β is the allowable coefficient, generally related to the user's own wishes, The larger the β, the longer the user's detour on the new path. On the contrary, the smaller the β, the shorter the user's detour. It is recommended to take 1.0-1.5; otherwise, store P _{n and Kn} in S _n ;

6) If N'<K _n <N, N is the maximum number of alternative paths, and N' is the smallest number of alternative paths, then return 3);

7) If all the origin-destination pairs of the entire road network are calculated, then end, otherwise return to step 2) calculate the n+1th origin-destination pair, until all Q origin-destination pairs are calculated and end;

The algorithm first adds a large positive value to the weight of the shortest path and other high-risk road units in the road network under normal conditions, and then calculates the new shortest path. When the length of the new shortest path exceeds the detour constraint, it gradually The added weight is reduced, so that under the circumvention constraints, a reliable path with less overlap with the higher delay risk (low reliability) units in the obtained alternative paths can be effectively searched. At the same time, try to avoid other high-risk units in the road network. In the above steps, the function α ^m (1-r _i ) ^q W ₀ can guarantee to increase the weight of the units that you want to avoid on the one hand, thereby reducing the possibility of these units appearing in the new alternative path, on the other hand, It can ensure that as the number of iterations increases, the amount of weight increase decreases, thereby ensuring that some units that are originally expected to be avoided are passed by alternative paths to meet the detour constraints. This algorithm belongs to the heuristic algorithm, although it may not get the optimal solution, but it can get the acceptable result faster.

(3) Encode and store alternative way sets

Although online dynamic screening based on alternative paths has good real-time responsiveness, storing the set of alternative paths for all point pairs in the road network undoubtedly requires a huge storage space. Therefore, the present invention uses existing path encoding technology to store alternative paths . Among them, for the shortest path, the present invention assumes that the starting point is s, the end point is t, and the intermediate nodes are m ₁ , m ₂ ,...m _n for the shortest path between each start-destination point pair, and the following code is used to record the road:

<end point number t, next node number m ₁ next to the starting point, shortest path length>

For other alternative paths that are not the shortest path, this encoding form cannot be used, and the existing technology can be used to store the full information of the path. However, due to the large daily traffic volume, the origin-destination pairs are more likely to have clustering phenomenon and lead to peak calculation delays. For this reason, alternative routes can be calculated and stored only for important origin-destination pairs with large daily traffic volume. In reality, multi-path information is released only for important origin-destination pairs with large daily traffic volume. For other origin-destination pairs, only single-path information publication is required.

2. The second stage: online stage

(1) Online backtracking of alternative paths

Use the existing path backtracking method to backtrack the path online. For the shortest path in the alternative path, according to the user's needs, first find the start-end pair, that is, first find the start point s, and then find the end point t in all the code tables of the start point. , and then use this code to find the next node m ₁ next to the starting point, and then use m ₁ as a new starting point, find the code of the end point t in the coding table of m ₁ , and then find the second next node next to the new starting point m ₁ Node m ₂ , take m ₂ as the new starting point, and so on until the next node is the end point, and record these nodes in turn to get the shortest path; for the non-shortest path in the alternative path, all stored Node numbers are traced back in order from start point to end point.

(2) Online dynamic path screening, supplementary search and multi-path publishing

When online, route screening can be performed according to dynamic traffic information, that is, routes that are blocked and delayed in the alternative routes are deleted, and then the remaining alternative routes are randomly released to each navigation vehicle. When there are no alternative routes available online or there are not enough When the number of available alternative routes is used to avoid clustering, the alternative route set algorithm can be re-used according to real-time traffic information to re-search the route and publish it online, and the algorithm ends.

Description of drawings

Figure 1 Algorithm Overall Flowchart

Figure 2 Flow chart of calibrating dual-parameter road attribute files

Figure 3 Algorithm flow chart for establishing alternative paths

Figure 4 Alternative paths without detour constraints

Figure 5 Alternative paths with detour constraints but without considering unit reliability

Figure 6 Alternative paths with detour constraints and unit reliability considerations

Figure 7 uses the Euclidean distance divided by the maximum speed of the road network as the path search range of the estimated value of the A* heuristic function

Figure 8 Using the shortest travel time under normal conditions as the path search range of the estimated value of the A* heuristic function

Figure 9 The relationship between search time and the number of nodes

Detailed ways

This part only implements the first stage of the algorithm. The implementation method is to carry out experiments under virtual network and conditions. First, a small network with 36 nodes and 60 road sections is tested under three conditions. Reliable path search with detour constraints at the starting point of travel, using the experimental results to show the rationality of the invention, and then using the experimental results of a large road network with 2800 nodes to show the search efficiency of the algorithm. The nodes are represented by circles, and the node numbers are marked inside the circles.

The small networks are shown in Figure 4 to Figure 6, which respectively indicate that 1) detour constraints are not considered; 2) detour constraints are considered, but unit reliability is not considered; 3) detour constraints and unit reliability are considered at the same time Condition. The implementation steps are as follows:

Step 1: Calibrate the road dual-parameter attribute file

Since the implementation process is carried out under the virtual road network, the method of random assignment is used to assign values to the double parameters of the road unit in the calibration of the road double parameter attribute file;

The specific method is: randomly assign speeds to road units within the speed range of 30-60, then calculate the length of road units according to the coordinates, and divide the length by the randomly assigned speed to obtain the average transit time of road units; finally, the reliability of road units The reliability of the road unit is randomly assigned within the range of 0.7-0.99 to analyze the effectiveness of the algorithm;

Step 2, search the set of alternative paths

Use the above-mentioned algorithm for establishing an alternative path set to search for alternative paths in the road network, where the relaxation factor α takes a value of 0.5, the allowable coefficient β takes a value of 1.2, and a large positive number W ₀ =1.5L _n,1 , the largest alternative path The number N and the minimum number of alternative paths N' are respectively 5 and 3. Here, r _i <0.9 is regarded as a high-risk road unit. See the above algorithm description for the specific implementation method;

Step 3: Coded storage of each routing set

Code and store the alternative paths according to the above code storage method, and mark them in the figure, see attached drawings 4-6;

After the algorithm is implemented, the result description:

The effect after the first stage of the final algorithm is implemented is shown in Figure 4-8, and the average speed (km/h) and reliability under normal conditions are marked next to the relevant road sections.

The starting point and the ending point are represented by black rectangles, the starting point number is 24, and the ending point number is 35. The first candidate path (the shortest path with average transit time) is shown in black, the second candidate path is shown in green, and the third candidate path is shown in red.

Under the assumption that the road sections are independent after failure, the reliability of the path is the product of the reliability of all the road sections in the path, and this product can be used to evaluate the reliability of a path. Table 1 shows the attribute information of the candidate road sets in the three cases.

It can be seen from the accompanying drawings that although there are more overlapping units in the candidate road set in Figure 6 than in Figure 4 and Figure 5, the reliability of the candidate road set is higher than that in Figure 4 and Figure 5, because it is in the construction of the candidate road set. Units with high blocking risk are effectively avoided during set time.

In order to test the effectiveness of the alternative path calculation algorithm after improving the heuristic function of the A* algorithm, an experiment was carried out on a road network with 2800 nodes. Extract point pairs to calculate the second alternative route. Attached Figures 7 and 8 respectively show the path search range by dividing the Euclidean distance by the maximum speed of the road network and using the shortest passing time under normal conditions as the estimated value of the A* heuristic function. The black circles in the figure represent During the search process, the number of expansion points of the former is 612, and the number of expansion points of the latter is 81. It can be seen that the search efficiency is greatly improved by using the shortest transit time under normal conditions as the estimated value of the A* heuristic function.

According to the random calculation and analysis of 1000 times on the notebook with 64M memory and 800M CPU for different scale grid road networks, the second alternative path can be obtained through 1-2 iterations on average, and the third alternative path can be obtained through an average of 1-2 iterations. 3 to 4 iterations can be obtained. The total offline calculation time is shortened by an average of 66% compared with the total calculation time of the constrained A* algorithm based on Euclidean distance estimation. The online calculation time only has a linear growth relationship with the number of vehicles requesting navigation per unit time and the scale of the road network. Figure 9 is a comparison of the online calculation time of the two-stage dynamic multi-path optimization strategy and the efficiency of the fully online dynamic A* algorithm path optimization under different scales of road networks.

Table 1 Alternative road set attribute information in three cases

Claims (2)

1, a kind of center control type onboard navigation system two-step multi-path optimization method is characterized in that, may further comprise the steps:

Phase one: off-line phase

(1) demarcates biparametric roadway element property file

T typical period of time, t＞=3 will be divided in one day; Below operation carry out at any one typical period of time wherein, the method for operating of other period is identical;

Historical information processed with refine into two seed ginseng number form formulas and serve navigation algorithm: the one, the average transit time of the roadway element of each typical period of time, the one, the unblocked reliability of each typical period of time roadway element;

Roadway element generally comprises highway section and crossing; Utilize prior art that road network is converted into arc power network, soon the crossing respectively turns to virtual segment and represents that respectively turning to the delay time at stop is the transit time of virtual segment; The crowded not serious road network in the crossing, also the highway section transit time is only considered in the not delay of considering intersection; The segment unit thereby the road network roadway element can only show the way;

Each roadway element is carried out average transit time of each typical period of time and the demarcation of unblocked reliability biparametric, wherein sets up the average transit time of roadway element, adopt following method:

For roadway element with traffic flow historical data more than month, calculate roadway element i average transit time in a certain period under unimpeded condition with statistical method, this average transit time is exactly the weight w of roadway element i _iFor the highway section of the traffic flow historical data of neither one more than the moon, can road section length divided by the weight w of highway section design rate as roadway element i _i

Determine that wherein the roadway element unblocked reliability will be through two processes:

A. roadway element is done the binary condition hypothesis: the state that is about to roadway element is divided into two kinds: unimpeded and obstruction; Distinguish highway section boundary unimpeded and that block and be the unit of setting in advance, the highway section threshold speed that travels, then be considered as unimpededly greater than this value, then be considered as obstruction less than this value; It then is the intersection delay threshold value of setting in advance that the differentiation crossing respectively turns to boundary unimpeded and that block; Then be considered as unimpededly less than this value, then be considered as blocking greater than this value; Can be with upper threshold value according to Ministry of Communications's smooth traffic project standard formulation;

B. determine the unblocked reliability of roadway element, the unblocked reliability of roadway element can be defined as the probability that in official hour section roadway element is unimpeded or do not block; It determines that method is as follows:

To the traffic flow data of roadway element collection more than 3 months, adopt following formula to calculate roadway element i unblocked reliability r _iApproximate value:

(2) set up the alternative path collection

1) make Q be total origin and destination logarithm of whole road network, right to all origin and destination, its alternative road collection S of initialization _n=φ, n represent that n origin and destination among the Q are right, n=1,2,3 ... Q;

With the transit time under each roadway element normal condition is right of way, to have a few to calculating shortest path P with traditional dijkstra's algorithm _{N, 1}, and then calculate its weight and be L _{N, 1}, be path P _{N, 1}Transit time; With P _{N, 1}Deposit n the alternative road collection S that origin and destination are right in _n, as article one alternative path;

The right alternative path bar in all origin and destination of initialization is counted K _n=1;

Make n=1, promptly to first origin and destination to calculating;

2) right to n the beginning and the end, initialization number of iterations m=0;

3) the lower roadway element of online fiduciary level that satisfies the need, r is got in suggestion _i＜0.5-0.9, and S _nIn each roadway element on the existing path increase weight Δ w _i, make new round roadway element weight w _i' be:

w _i’＝w _i+Δw _i＝w _i+α ^m(1-r _i) ^qW ₀ (2)

In the formula, w _iFor cell-average transit time right of way, when m=0, q=0, otherwise q=1; α is a relaxation factor, 0＜α＜1, and the possibility that big more this roadway element of α is included in the new route is more little, and the possibility that opposite more little this unit of α is included in the new route is big more; W ₀Be big positive number, W is got in suggestion ₀=1.5L _{N, 1}～3L _{N, 1}

4) make new round alternative path count K _n'=K _n+ 1, new round iterations m '=m+1, m=m ' is with w _i' be right of way, with existing A ^*Algorithm computation shortest path P _{N, Kn}, K _n=K ' _n, right of way is reverted to w _i, recomputate L _{N, Kn}, L _{N, Kn}Be P _{N, Kn}On right of way w _iSum;

5) if P _{N, Kn}Violated reasonable path constraint condition L _{N, Kn}＜β L _{N, 1}K then _n'=K _n-1, forward 3 to); β is for allowing coefficient, and general relevant with user's self wish, it is long more that the big more new route user of β detours, and the opposite more little user of β detours short more, and 1.0-1.5 is got in suggestion; Otherwise with P _{N, Kn}Deposit S in _n

6) if N '＜K _n＜N, N are maximum alternative path number, and N ' is minimum alternative path bar number, then returns 3);

7), otherwise return step 2 if the origin and destination of whole road network are intact then finish to whole calculating) to calculate n+1 origin and destination right, up to all Q origin and destination to all calculating the bundle that finishes;

(3) the alternative road of off-line code storage collection

The shortest path of concentrating for alternative road wherein, the present invention to each origin and destination to shortest path, suppose that starting point is s, terminal point is t, intermediate node is m ₁, m ₂... m _n, adopt this road of following encoded recording:

＜terminal point numbering t, the next node numbering m of next-door neighbour's starting point ₁, shortest path is long 〉

For other alternative paths of non-shortest path, then can not use this coding form, can adopt prior art to carry out path perfect information storage;

Subordinate phase: online stage

(2.1) alternative path is online recalls

Utilize the existing route retrogressive method that the path is carried out onlinely recalling, for the shortest path in the alternative path, according to user's demand, find terminus right earlier, promptly find starting point s earlier, in the coding schedule of all these starting points, find terminal point t then, find the next node m of next-door neighbour's starting point subsequently by this coding ₁, and then with m ₁As new starting point, at m ₁Coding schedule in find the coding of terminal point t, find next-door neighbour's ground zero m then ₁Second next node m ₂, again with m ₂Be new starting point, and the like, be that terminal point stops up to next node, these nodes are write down successively just obtained shortest path; For the non-shortest path in the alternative path, then recall by order from origin-to-destination by all node serial numbers that store;

(2.2) online dynamic route screening, additional search and multipath issue

Can carry out the path screening according to dynamic information when online, promptly leave out the path of incuring loss through delay takes place in the alternative path to block, and then will remain alternative path each navigation vehicle will be carried out random walk issue, when online when not having available alternative path or not having the available alternative path of sufficient amount to be used to avoid the clustering phenomenon, can utilize again according to Real-time Traffic Information and set up the online issue of search path again of alternative path set algorithm, algorithm finishes.

2, center according to claim 1 control type onboard navigation system two-step multi-path optimization method is characterized in that:

In the phase one: demarcate the unblocked reliability of determining roadway element in the biparametric roadway element property file in the off-line phase, when the traffic flow data that do not have more than 3 months, adopt the unblocked reliability of following functional form approximate estimation roadway element:

In the formula,

r _i---the unblocked reliability of roadway element i;

v _i/ c _i---be defined as the saturation degree of roadway element i, wherein v _iBe flow, can obtain c by traffic flow data in a short time by roadway element i _iBe the traffic capacity of unit i, different with grade and different according to road type, be a definite value, can check in by document;

Beta, gamma---return undetermined coefficient; C---constant term.

Images