CN105629992A - UUV navigation path planning method under threat Internet - Google Patents

Abstract

The invention relates to the technical field of path planning, and especially relates to a UUV navigation path planning method under a threat Internet. The invention aims at solving a problem that the safety probability, which should be kept, of a UUV cannot be set in advance in an environment of the threat Internet when there are an obstacle region and a threat region at the same time, and a problem that a navigation path is difficult to remain the shortest as much as possible. The method enables a safety probability calculation rule and an ant colony state transfer probability calculation method based on safety guarantee to be used in an ant algorithm. Different from a conventional ant colony algorithm which has a fixed target point in a planning process, the target points in the invention will change sequentially in a planning process according to a traversal sequence. A path obtained by each ant is a complete path from an arrangement point and all designated points to a recovery point. The planning is not planned in a segmented and jointed, but is completed at a time. The method can be used in a technical field of path planning.

Description

A kind of threaten the Route planner of UUV under the Internet

Technical field

The present invention relates to Path Planning Technique field, particularly relate to and a kind of threaten the Route planner of UUV under the Internet.

Background technology

UUV (UnmannedUnderwaterVehicle, UAV navigation) in routeing process, threatening area is different from barrier zone, and barrier is that strictly no thoroughfare, and to threaten district be can select by but can undertake the region of certain risk. Along with the development of communication technology, each threatens district no longer to work alone, but by networking share target information, therefore after UUV is by a certain threat district, when threatening district again by other, its risk born will be greatly increased. So under threatening the Internet, the simple of district's threat level that threaten that the threat that UUV is subject to is no longer its traverse is added up, Tian Kuo etc. are in " under threat netting, unmanned plane path is planned online ", and symbol little Wei etc. has inquired into, in " air defense threat networking modeling with analyze ", the Cost Model threatened under the Internet. Wu Xianghong etc. think when target is found after once in " the autonomous underwater robot path decision technique study based on ant group optimization ", and the threat level in each threat district presses certain way to be increased; And the coverage threatening district is expanded by Jiang Li equality in " under threat netting the research of low-level penetration routeing ", embody the increase of threat.

These documents above-mentioned are all devise threat Cost Model, then this model are introduced in planning algorithm and pass through and threaten district limit air route, thus obtaining safer path. But these methods the maximum risk of manual control can not bear probability (or aircraft should keep safe probability), and UUV threatens district, the threat district occurred more afterwards of detouring often by what relatively early occur, in practical application, such phenomenon is difficult to ensure that air route is more excellent, because under meeting some restrictive condition, the risk that UUV detours the threat relatively early occurred and is likely to bear by the threat occurred more afterwards is less, and path optimizes more.

Summary of the invention

The theoretical content that the present invention relates to

1, safe probability computation rule

The risk that path planning is born by the present invention changes into safe probability safep, and the risk that namely UUV bears is more big, and safe probability is more little. The dangerous probability threatening district increases by certain way index according to the UUV threat district number passed through. The safe probability computation rule in path is as follows:

$safep=\underset{i=0}{\overset{N}{\Π}}(1-w^{i-1}\×thread\_pro\_i)$

Wherein, i is the threat district sequence number passed through on path, and N is the threat district sum passed through on path, and thread_pro_i is the threat probabilities that the i-th passed through on path threatens district, and w is for threatening upgrading weights, w > 1;

2, based on the ant colony state transition probability computational methods of safety guarantee

The present invention uses Visual Graph method to set up the state branchpoint of Formica fusca in ant group algorithm, in order to be met the path of safe probability requirement, devises following state transition probability computation rule:

${p_{ij}}^k=\frac{{\left({\mathrm{\τ}}_{ij}\right)}^{\mathrm{\α}}{\left({\mathrm{\η}}_j\right)}^{\mathrm{\β}}}{\underset{r\∈allow}{\Σ}{\left({\mathrm{\τ}}_{ir}\right)}^{\mathrm{\α}}{{\mathrm{\η}}_r}^{\mathrm{\β}}}$

In above formula,Transfer to the probability of visible dots j in current goal point i place for Formica fusca k; Allow is the Formica fusca k visible dots set allowing to do state transfer in current goal point i place; �� is heuristic function, for the inverse of visible dots to be transferred to current goal point air line distance, ��_jIt is the visible dots j to be transferred inverse to current goal point i air line distance, ��_rBe in allow any point to the inverse of current goal point i air line distance; ��_ijFor pheromone concentration; �� is the significance level of pheromone; �� is the significance level of heuristic function;

Can through threatening area if transferring to ant colony point j from ant colony point i, then need according to already by threatening area quantity and danger classes and the threat district quantity that will pass through between i, j of point and danger classes correction its transfer weights (��_ij)^��(��_j)^��Size obtains revising transfer weights ((��_ij)^��(��_j)^��)^*, detailed process is as follows:

1) utilize safe probability computation rule, according to current already by threatening area quantity num and danger classes calculate current safety probability now_safep;

2) obtain and transfer to the j threatening area quantity that can pass through and danger classes from an i, and arrange according to the sequencing of process; Calculate the safe probability after transferring to j from i

$ifTransfer\_safep=now\_safep\×\underset{l=1}{\overset{m}{\Π}}(1-\frac{w^{num+l}\×threat\[l\]}{10.0})$

Wherein thread [] transfers to, from an i, the danger classes that each threatening area sequential storage of j process is corresponding, and danger classes is from the arbitrary integer between 0 to 10; M is the threat district number transferring to j process from an i; W is for threatening upgrading weights, w > 1;

If 1. ifTransfer_safep < safep_limit, then give (��_ij)^��(��_j)^��Compose a minimum value;

If 2. ifTransfer_safep > safep_limit, then,

((��_ij)^��(��_j)^��)^*=�� �� (��_ij)^��(��_j)^����ifThransfer_safep

Wherein, �� is the random number between 0��1;

Heretofore described Formica fusca can substitute with UUV;

The present invention solves under the environment of existing threat the Internet, when there is barrier simultaneously and threaten district, existence can not arrange the UUV safe probability that should keep in advance, and is difficult to ensure that the problem that air route is the shortest as far as possible, and proposes a kind of to threaten the paths planning method of UUV under the Internet.

A kind of threaten the paths planning method of UUV under the Internet, sequentially include the following steps:

Step one: incoming lay point, recovery point, must through the array list [] of point by traversal order storage; Obtain the safe probability safep_limit that user requires; Ant colony population quantity m_AntNum, maximum iteration time Max_generation, history optimum reservation number of times Max_histBest, pheromone volatility coefficient vol, pheromone upper limit up_limit, pheromone lower limit low_limit are set; Step 2: initialization information prime matrix, history optimum Formica fusca HistoryBestAnt, history optimum Formica fusca has preserved number of times history_best_hold=0, loop iteration number m_generation=0;

Step 3: Formica fusca k=0 is set, forwards step 4 to;

Step 4: if k is > m_AntNum, forward step 8 to, if, k < m_AntNum, obtaining must through the length listnum of a chained list, and setting through a t=0, must forward step 5 to;

Step 5: if through a t < listnum, specific item punctuate subAimPoint must be arranged for through a list [t], step 6 must be forwarded to; If t=listnum, arranging specific item punctuate subAimPoint is recovery point;

Step 6: set now Formica fusca point as g_iIf, g_iFor subAimPoint, and subAimPoint is must through point, and Formica fusca has found this sub-goal, assignment t=t+1, forwards step 5 to; If g_iFor subAimPoint, and subAimPoint is recovery point, and Formica fusca completes whole route searching, assignment k=k+1, forwards step 4 to; Otherwise, then step 7 is forwarded to;

Step 7: if g_iCorresponding visible dots set allow_iFor sky, Formica fusca is dead, and assignment k=k+1 forwards step 4 to; Based on the ant colony state transition probability computational methods of safety guarantee, calculate allow_iThe transfer weights of middle every bit, utilize roulette to select branchpoint g_j��allow_i, and from g_jAllow_jA middle deletion point g_i, forward step 6 to;

Step 8: assignment m_generation=m_generation+1; If in current iteration, the path that Formica fusca is passed by is more excellent than history optimum Formica fusca HistoryBestAnt, updates HistoryBestAnt, history_best_hold=0; Otherwise assignment history_best_hold=history_best_hold+1;

Step 9: if m_generation > Max_generation or history_best_hold > Max_histBest, algorithm terminates; Otherwise update pheromone, forward step 3 to.

The present invention includes following beneficial effect:

1, the present invention is in barrier and threatens district to exist simultaneously, and threaten a kind of ant colony path planning algorithm of design under the environment of the Internet, make the some of the complete user setup of the traversal path cooked up must return to recovery point after point, make shortest path when trying one's best through a traversal order is certain and the safe probability of user setup can be met;

2, the present invention ant colony state transition probability computational methods by safe probability computation rule with based on safety guarantee apply to ant algorithm, when ensureing to arrange the UUV safe probability that should keep in advance, it is ensured that air route is the shortest as far as possible, and the two can be taken into account;

3 have fixing impact point different from traditional ant group algorithm in planning process, its impact point of algorithm that the present invention proposes can change successively according to traversal order in planning process, and the path that every Formica fusca obtains is all from laying the traversed all of fullpath that must return to recovery point after point of point;

4, compared with the planning of traditional ant group algorithm, the planning of the present invention is disposable completing rather than splicing after segmentation planning.

Accompanying drawing explanation

Fig. 1 is one to be threatened district and requires when safe probability is 0.8, physical planning path simulation proof diagram when danger classes is low;

Fig. 2 is one to be threatened district and requires when safe probability is 0.8, physical planning path simulation proof diagram during danger classes height;

Fig. 3 is multiple threat district and requires when safe probability is 0.8, through the physical planning path simulation proof diagram in single threat district;

Fig. 4 is multiple threat district, barrier and the physical planning path simulation proof diagram that in time putting and require that safe probability is 0.8, must pass single threat district;

Fig. 5 is multiple threat district, barrier and the physical planning path simulation proof diagram that in time putting and require that safe probability is 0.6, must pass multiple threat districts.

Detailed description of the invention

Understandable for enabling the above-mentioned purpose of the present invention, feature and advantage to become apparent from, below in conjunction with detailed description of the invention, the present invention is further detailed explanation.

A kind of described in detailed description of the invention one, present embodiment threatens the paths planning method of UUV under the Internet, sequentially includes the following steps:

Step one: incoming lay point, recovery point, must through the array list [] of point by traversal order storage; Obtain the safe probability safep_limit that user requires; Ant colony population quantity m_AntNum, maximum iteration time Max_generation, history optimum reservation number of times Max_histBest, pheromone volatility coefficient vol, pheromone upper limit up_limit, pheromone lower limit low_limit are set; Step 2: initialization information prime matrix, history optimum Formica fusca HistoryBestAnt, history optimum Formica fusca has preserved number of times history_best_hold=0, loop iteration number m_generation=0;

Step 3: Formica fusca k=0 is set, forwards step 4 to;

Step 4: if k is > m_AntNum, forward step 8 to, if, k < m_AntNum, obtaining must through the length listnum of a chained list, and setting through a t=0, must forward step 5 to;

Step 5: if through a t < listnum, specific item punctuate subAimPoint must be arranged for through a list [t], step 6 must be forwarded to; If t=listnum, arranging specific item punctuate subAimPoint is recovery point;

Step 6: set now Formica fusca point as g_iIf, g_iFor subAimPoint, and subAimPoint is must through point, and Formica fusca has found this sub-goal, assignment t=t+1, forwards step 5 to; If g_iFor subAimPoint, and subAimPoint is recovery point, and Formica fusca completes whole route searching, assignment k=k+1, forwards step 4 to; Otherwise, then step 7 is forwarded to;

Step 7: if g_iCorresponding visible dots set allow_iFor sky, Formica fusca is dead, and assignment k=k+1 forwards step 4 to; Based on the ant colony state transition probability computational methods of safety guarantee, calculate allow_iThe transfer weights of middle every bit, utilize roulette to select branchpoint g_j��allow_i, and from g_jAllow_jA middle deletion point g_i, forward step 6 to;

Step 8: assignment m_generation=m_generation+1; If in current iteration, the path that Formica fusca is passed by is more excellent than history optimum Formica fusca HistoryBestAnt, updates HistoryBestAnt, history_best_hold=0; Otherwise assignment history_best_hold=history_best_hold+1;

Step 9: if m_generation > Max_generation or history_best_hold > Max_histBest, algorithm terminates; Otherwise update pheromone, forward step 3 to.

Present embodiment includes following beneficial effect:

1, the present invention is in barrier and threatens district to exist simultaneously, and threaten a kind of ant colony path planning algorithm of design under the environment of the Internet, make the some of the complete user setup of the traversal path cooked up must return to recovery point after point, make shortest path when trying one's best through a traversal order is certain and the safe probability of user setup can be met;

2, the present invention ant colony state transition probability computational methods by safe probability computation rule with based on safety guarantee apply to ant algorithm, when ensureing to arrange the UUV safe probability that should keep in advance, it is ensured that air route is the shortest as far as possible, and the two can be taken into account;

3 have fixing impact point different from traditional ant group algorithm in planning process, its impact point of algorithm that the present invention proposes can change successively according to traversal order in planning process, and the path that every Formica fusca obtains is all from laying the traversed all of fullpath that must return to recovery point after point of point;

4, compared with the planning of traditional ant group algorithm, the planning of the present invention is disposable completing rather than splicing after segmentation planning.

Detailed description of the invention two, present embodiment are that a kind of described in detailed description of the invention one threatens further illustrating of the Route planner of UUV under the Internet, and the computation rule related in step 7 has safe probability computation rule, and its particular content is as follows:

The risk born by path planning changes into safe probability safep, and the risk that namely UUV bears is more big, and safe probability is more little; The dangerous probability threatening district increases by certain way index according to the UUV threat district number passed through, and the safe probability computation rule in path is as follows:

$safep=\underset{i=0}{\overset{N}{\&Pi;}}(1-w^{i-1}\&times;thread\_pro\_i)$

Wherein, i is the threat district sequence number passed through on path, and N is the threat district sum passed through on path, and thread_pro_i is the threat probabilities that the i-th passed through on path threatens district, and w is for threatening upgrading weights, w > 1.

Detailed description of the invention three, present embodiment are that a kind of described in detailed description of the invention one threatens further illustrating of the Route planner of UUV under the Internet, and the computation rule related in step 7 has state transition probability computational methods, and its particular content is as follows:

State transition probability computation rule:

${p_{ij}}^k=\frac{{\left({\mathrm{\&tau;}}_{ij}\right)}^{\mathrm{\&alpha;}}{\left({\mathrm{\&eta;}}_j\right)}^{\mathrm{\&beta;}}}{\underset{r\&Element;allow}{\&Sigma;}{\left({\mathrm{\&tau;}}_{ir}\right)}^{\mathrm{\&alpha;}}{{\mathrm{\&eta;}}_r}^{\mathrm{\&beta;}}}$

In above formula,Transfer to the probability of visible dots j in current goal point i place for Formica fusca k; Allow is the Formica fusca k visible dots set allowing to do state transfer in current goal point i place; �� is heuristic function, for the inverse of visible dots to be transferred to current goal point air line distance, ��_jIt is the visible dots j to be transferred inverse to current goal point i air line distance, ��_rBe in allow any point to the inverse of current goal point i air line distance; ��_ijFor pheromone concentration; �� is the significance level of pheromone; �� is the significance level of heuristic function;

Ant colony point i transfers to ant colony point j can pass through danger side of body region, then need according to already by threatening area quantity and danger classes and the threat district quantity that will pass through between i, j of point and danger classes correction its transfer weights (��_ij)^��(��_j)^��Size obtains revising transfer weights ((��_ij)^��(��_j)^��)^*��

Detailed description of the invention four, present embodiment are that a kind of described in detailed description of the invention three threatens further illustrating of the Route planner of UUV under the Internet, correction transfer the weights ((�� described in step 7_ij)^��(��_j)^��)^*Acquisition detailed process as follows:

1) utilize safe probability computation rule, according to current already by threatening area quantity num and danger classes calculate current safety probability now_safep;

2) obtain and transfer to the j threatening area quantity that can pass through and danger classes from an i, and arrange according to the sequencing of process; Calculate the safe probability after transferring to j from i

$ifTransfer\_safep=now\_safep\&times;\underset{l=1}{\overset{m}{\&Pi;}}(1-\frac{w^{num+1}\&times;threat\&lsqb;l\&rsqb;}{10.0})$

Wherein thread [] transfers to, from an i, the danger classes that each threatening area sequential storage of j process is corresponding, and danger classes is from the arbitrary integer between 0 to 10; M is the threat district number transferring to j process from an i; W is for threatening upgrading weights, w > 1;

If 1. ifTransfer_safep < safep_limit, then give (��_ij)^��(��_j)^��Compose a minimum value;

If 2. ifTransfer_safep > safep_limit, then,

((��_ij)^��(��_j)^��)^*=�� �� (��_ij)^��(��_j)^����ifThransfer_safep

Wherein, �� is the random number between 0��1.

Following emulation experiment is carried out for checking beneficial effects of the present invention

The present invention threaten upgrading weights be taken as w=1.2, analogous diagram is shown in Fig. 1 to Fig. 5, black solid is barrier, the block that grid is filled indicates " spy " word for threatening district, and indicating two number, last digit represents threat district ID, and a rear numeral is danger classes, for from 0 to 10 integer, correspond to dangerous probability is 0.0 to 1.0; Must indicate through point " must " word, and indicate must through No. ID; As seen from Figure 1, the threat level threatening district is 2 grades, and threat probabilities is 0.2, it is 0.8 that the safe probability computation rule proposed by the present invention obtains it by the safe probability behind threat district, meeting user's requirement, therefore algorithms selection directly passes through threat district, it is to avoid detour increase path. As seen from Figure 2,0.7 being unsatisfactory for user's requirement by threatening district safe probability to be, so no matter how far, path all can select to walk around threat district. Fig. 3 to Fig. 5, sets forth exist simultaneously substantial amounts of barrier, must through point, threaten district environmental simulation, can be seen that the path obtained all comparatively optimizes, physical planning path safe probability respectively 0.9,0.8 and 0.684, all can meet the safe probability requirement of user setup.

Claims (4)

1. one kind threatens the Route planner of UUV under the Internet, it is characterised in that it sequentially includes the following steps:

Step one: incoming lay point, recovery point, must through the array list [] of point by traversal order storage; Obtain the safe probability safep_limit that user requires; Ant colony population quantity m_AntNum, maximum iteration time Max_generation, history optimum reservation number of times Max_histBest, pheromone volatility coefficient vol, pheromone upper limit up_limit, pheromone lower limit low_limit are set;

Step 2: initialization information prime matrix, history optimum Formica fusca HistoryBestAnt, history optimum Formica fusca has preserved number of times history_best_hold=0, loop iteration number m_generation=0;

Step 3: Formica fusca k=0 is set, forwards step 4 to;

Step 4: if k > m_AntNum, forward step 8 to, if, k < m_AntNum, obtaining must through the length listnum of a chained list, and setting through a t=0, must forward step 5 to;

Step 5: if through a t < listnum, specific item punctuate subAimPoint must be arranged for through a list [t], step 6 must be forwarded to; If t=listnum, arranging specific item punctuate subAimPoint is recovery point;

Step 6: set now Formica fusca point as gi, if g_iFor subAimPoint, and subAimPoint is must through point, and Formica fusca has found this sub-goal, assignment t=t+1, forwards step 5 to; If g_iFor subAimPoint, and subAimPoint is recovery point, and Formica fusca completes whole route searching, assignment k=k+1, forwards step 4 to; Otherwise, then step 7 is forwarded to;

Step 7: if g_iCorresponding visible dots set allow_iFor sky, Formica fusca is dead, and assignment k=k+1 forwards step 4 to; Based on the ant colony state transition probability computational methods of safety guarantee, calculate allow_iThe transfer weights of middle every bit, utilize roulette to select branchpoint g_j��allow_i, and from g_jAllow_jA middle deletion point g_i, forward step 6 to;

Step 8: assignment m_generation=m_generation+1; If in current iteration, the path that Formica fusca is passed by is more excellent than history optimum Formica fusca HistoryBestAnt, updates HistoryBestAnt, history_best_hold=0; Otherwise assignment history_best_hold=history_best_hold+1;

Step 9: if m_generation > Max_generation or history_best_hold > Max_histBest, algorithm terminates; Otherwise update pheromone, forward step 3 to.

2. as claimed in claim 1 a kind of threaten the Route planner of UUV under the Internet, it is characterised in that the computation rule related in step 7 has safe probability computation rule, and its particular content is as follows:

The risk born by path planning changes into safe probability safep, and the risk that namely UUV bears is more big, and safe probability is more little; The dangerous probability threatening district increases by certain way index according to the UUV threat district number passed through, and the safe probability computation rule in path is as follows:

$safep=\underset{i=0}{\overset{N}{\&Pi;}}(1-w^{i-1}\&times;thread\_pro\_i)$

Wherein, i is the threat district sequence number passed through on path, and N is the threat district sum passed through on path, and thread_pro_i is the threat probabilities that the i-th passed through on path threatens district, and w is for threatening upgrading weights, w > 1.

3. as claimed in claim 1 or 2 a kind of threaten the Route planner of UUV under the Internet, it is characterised in that the computation rule related in step 7 has state transition probability computational methods, and its particular content is as follows:

State transition probability computation rule:

${p_{ij}}^k=\frac{{\left({\mathrm{\&tau;}}_{ij}\right)}^{\mathrm{\&alpha;}}{\left({\mathrm{\&eta;}}_j\right)}^{\mathrm{\&beta;}}}{\underset{r\&Element;allow}{\&Sigma;}{\left({\mathrm{\&tau;}}_{ir}\right)}^{\mathrm{\&alpha;}}{{\mathrm{\&eta;}}_r}^{\mathrm{\&beta;}}}$

In above formula,Transfer to the probability of visible dots j in current goal point i place for Formica fusca k; Allow is the Formica fusca k visible dots set allowing to do state transfer in current goal point i place; �� is heuristic function, for the inverse of visible dots to be transferred to current goal point air line distance, ��_jIt is the visible dots j to be transferred inverse to current goal point i air line distance, ��_rBe in allow any point to the inverse of current goal point i air line distance; ��_ijFor pheromone concentration; �� is the significance level of pheromone; �� is the significance level of heuristic function;

Ant colony point i transfers to ant colony point j can pass through danger side of body region, then need according to already by threatening area quantity and danger classes and the threat district quantity that will pass through between i, j of point and danger classes correction its transfer weights (��_ij)^��(��_j)^��Size obtains revising transfer weights ((��_ij)^��(��_j)^��)^*��

4. as claimed in claim 3 a kind of threaten the Route planner of UUV under the Internet, it is characterised in that correction transfer the weights ((�� described in step 7_ij)^��(��_j)^��)^*Acquisition detailed process as follows:

1) utilize safe probability computation rule, according to current already by threatening area quantity num and danger classes calculate current safety probability now_safep;

2) obtain and transfer to the j deathtrap quantity that can pass through and danger classes from an i, and arrange according to the sequencing of process; Calculate the safe probability after transferring to j from i

$ifTransfer\_safep=now\_safep\&times;\underset{l=1}{\overset{m}{\&Pi;}}(1-\frac{w^{num+l}\&times;threat\&lsqb;l\&rsqb;}{10.0})$

Wherein thread [] transfers to, from an i, the danger classes that each deathtrap sequential storage of j process is corresponding, and danger classes is from the arbitrary integer between 0 to 10; M is the threat district number transferring to j process from an i; W is for threatening upgrading weights, w > 1;

If 1. ifTransfer_safep < safep_limit, then give (��_ij)^��(��_j)^��Compose a minimum value;

If 2. ifTransfer_safep > safep_limit, then,

((��_ij)^��(��_j)^��)^*=�� �� (�� i_j)^��(��_j)^����ifThransfer_safep

Wherein, �� is the random number between 0��1.