CN104298239B - A kind of indoor mobile robot strengthens map study paths planning method - Google Patents

Abstract

The invention discloses a kind of indoor mobile robot and strengthen map study paths planning method, the steps include: that (1) obtains ambient condition information, set up obstacle pdf model；(2) greedy algorithm and enhancing cartography learning method is utilized to carry out path planning；(3) indoor mobile robot Path selection and adaptive speed adjustable strategies.Use strengthen map study path planning can be according to the intrinsic nonholonomic constraint of the present situation of indoor mobile robot and robot, plan current best path in real time, the most adaptive speed adjustable strategies can move the obstacle climbing ability of robot with both in-door, impact point convergence capabilities and planning efficiency, enable indoor mobile robot safely and effectively to arrive appointment position.

Description

A kind of indoor mobile robot strengthens map study paths planning method

Technical field

The present invention relates to the independent navigation field of ground wheeled robot, strengthen cartography particularly to a kind of indoor mobile robot Practise paths planning method.

Background technology

Along with the development of robotics and deepening continuously of artificial intelligence study, intelligent robot plays the part of more to come in human lives The most important role.As the one of common life robot, indoor mobile robot is used for as the substitute of service personal Indoor moving exhibition, home services, the complex dynamic environment such as lounge guiding.In this kind of environment, environmental information non-structural Changing, static dynamic barrier is staggered to be existed, and environmental information changes substantially, these factors ability to work to indoor mobile robot Propose challenge and requirement greatly.Completing service role preferably, indoor mobile robot needs have detecting obstacles thing, Division identification barrier, plans feasible path, the ability of stability contorting action in real time.Along with sensor technology, computer technology With the development of the network communications technology, real-time route planning becomes indoor mobile robot research as the brain of intelligent robot The most important thing.

Indoor mobile robot task in the environment can be understood as from where coming？Where go？How to go？What is done？ And the environmental information that path planning can plan as a whole sensor acquisition makes the path decision best suiting robot the present situation, the most how Go.Conventional path planning can distinguish static programming and dynamic programming, and static programming refers to that robot has possessed global context letter Breath, robot obtains path planning optimum under global context by calculated off line, and active path planning is many for the most unknown Path selection in environment, robot understands limited and environment to environmental information it may happen that change, has height with truth Similarity.For indoor mobile robot, it is desirable to its most optimal road being capable of in indoor moving dynamic environment Footpath selects and dynamic barrier is hidden.

As controlling and the object of planning, indoor mobile robot is that a class has time-varying, close coupling and the multi input of incomplete property Multi output nonlinear system.Complicating due to environment more and need to consider more multifactor, it is sufficiently complex that its programmed decision-making becomes.? In existing technology, conventional path planning, such as fuzzy programming, genetic algorithm, ant group algorithm, neutral net etc., the most not The requirement of dynamic environment and real-time can be met simultaneously.It addition, the incomplete property that wheeled mobile robot exists also governs indoor The Path selection of mobile robot.Therefore, research has the path planning algorithm of learning capacity becomes present stage Real-time and Dynamic path One main trend of project study.And design a kind of simple and reliable, real-time good, facilitate implementation, can to deal with multiclass dynamic The indoor mobile robot planing method of circumstances not known is to ensure that the key technology and realistic problem that services the most effectively carries out.

Summary of the invention

The present invention is directed to above-mentioned prior art be difficult in present paths planning method meet Dynamic Unknown Environment planning with in real time simultaneously The requirement of planning, uses the paths planning method strengthening map study, along with the movement of indoor mobile robot, constantly deepens reason The ambient condition information that Xie Xin obtains, iterative computation randomly selects the cost function in path, and it is optimum that study calculates current time Path, it is ensured that the good avoidance performance under the static-obstacle of indoor mobile robot, and meet dynamic disorder when occurring Real-time planning function, reach independent navigation and the paths planning method with the indoor mobile robot of relatively high-intelligentization.

A kind of indoor mobile robot strengthens map study paths planning method, including following step:

Step 1: set up the probabilistic model that search coverage is affected by barrier；

First, by sonar sensor that indoor mobile robot is self-contained, it is thus achieved that the surrounding letter of indoor mobile robot Breath；Secondly, using indoor mobile robot the region of process as search coverage, set up according to described ambient condition information The probabilistic model that search coverage is affected by barrier, and according to the ambient condition information real-time update of sonar sensor Real-time Collection The probabilistic model that search coverage is affected by barrier；

The probabilistic model that described search coverage is affected by barrier is as follows:

$F(X,Y)=1-\underset{i=1}{\overset{M}{\mathrm{\Π}}}\[1-f_i(X,Y)\]$

Wherein, indoor moving indoor mobile robot work space information collection is combined intoDescribed spatial information includes all target positions Put and all Obstacle Positions；In current investigative range, work space information collection is combined intoThe work space information detected Collection is combined into

Relative distance information between barrier and robot current location that indoor mobile robot is arrived by sonar radar detection, Utilize the self-contained mileage gauge of indoor mobile robot and inertial navigation system, relative distance information is converted to initial position The positional information fastened for the base coordinate of initial point and range information.

{ (X, Y) } is search coverage,?Map on be detected with M barrier, f_i(X, Y) is the influence function to indoor mobile robot Path selection of i-th barrier, uses normal state to divide Cloth is expressed as follows:

$f_i(X,Y)=\frac{1}{\sqrt{2\mathrm{\π}}{\mathrm{\σ}}_i}{e}^{(-\frac{{D_i}^2}{2{\mathrm{\σ}}_i})}$

Wherein, σ_iFor the coverage coefficient of i-th barrier, span is [0,1]；D_iFor i-th barrier to detecting The distance matrix of all positions in region, matrix size and map are in the same size for N × N, and each element in distance matrix is barrier Hinder thing to map the physical distance of each position；

Step 2: based on greedy algorithm and enhancing study iterative strategy, in maximum iteration time k set_maxIn, iteration updates to be worked as Front position p_now(t)With target location p_goalBetween path cost function, to reach the path corresponding to path cost function of convergence As the optimal path of current time, detailed process is as follows:

Step 2.1: make k=1, k represent iterations；

Step 2.2: judge that current iteration number of times has exceeded the maximum iteration time of setting the most, if exceeding, then returns step 2.1； Otherwise, step 2.3 is entered；

Return step 2.1 and refer to from the beginning of first time iteration, after rebuilding initial path, be again introduced into iterated search optimal path；

Step 2.3: from the known map that the ambient condition information obtained is corresponding, randomly choose a position p_kInsert indoor moving Robot place current location p_now(t)With target location p_goal, wherein p_kMeet condition{p_rFor it Path point set under front iteration optimal path cost function；Calculate the path cost function that kth time iteration obtains as follows

Above-mentioned formula is to utilize to strengthen the path cost function computing formula that study mechanism obtains；

Wherein,The path cost function that-1 iteration of kth obtains, and For Current location p on known map_now(t)With selected location p_kBetween connection weights,For use greedy algorithm obtain in kth Secondary iteration complete after obtain from position p_k-1To target location p_goalOptimal path cost function:

$G_p^k=\left\{\begin{array}{cc}{R}_{p_k,p_{goal}}+G_{p_{goal}}& k=1\\ \min \{R_{p_k,p_{goal}}+G_{p_{goal}},R_{p_k,p_{k-1}}+G_p^{k-1}\}& k\≠1\end{array}\right.$

Wherein,For target location p_goalCost function, and For on known map in kth time institute Select location point p_kWith-1 selected location p of kth_k-1Between connection weights；

Any two positions p in known map_aAnd p_bBetween connection weightsStraight with this by the air line distance between two positions The barrier passed on line affects probability and constitutes:

Wherein,For position p_aTo position p_bPhysical location distance between 2；For position p_aTo position p_bRoad Footpath point set, ds is path integral unit；Max () is that maximum finds a function, and i.e. tries to achieve position p_aTo position p_bPath collection The maximum of probability that in conjunction, a certain path is affected by barrier；

Step 2.4: the path cost function obtained after judging iterationReach convergence, if convergence, then exited Iterative process, enters step 3 using path corresponding for the path cost function restrained as optimal path；Otherwise, by iteration time Number k adds 1, returns step 2.2；

Step 3: the optimal path obtained using step 2 is as the current preselected path of indoor mobile robot, according to preliminary election routing Footpath and the position of indoor mobile robot and velocity judge deflection angle φ of indoor mobile robot_tWhether meet wheeled robot Can not the nonholonomic constraint of lateral sliding, determine whether indoor mobile robot moves according to preselected path direction:

If being unsatisfactory for, then return step 2, rebuild all paths of current time；Owing to, in step 2.3, inserting indoor Mobile robot place current location p_now(t)With target location p_goalBetween position p_kFor randomly choosing, wherein p_kMeet condition{p_rIt is the path point set under iteration optimal path cost function before,For indoor moving machine People's work space information set.Therefore, after returning step 2, randomly choose the on position obtained, can be with choosing before The on position crossed is different, can obtain new path；

If meeting, moving with the optimal path direction that the indoor mobile robot translational speed set is cooked up along current time, entering Enter the path planning of subsequent time, t=t+1, return step 1, until indoor mobile robot moves to target location, complete Path planning；

Deflection angle φ of described indoor mobile robot_tWhether meet wheeled robot can not the nonholonomic constraint of lateral sliding refer to: φ_t∈[0,60°]∪[120°,180°]；

Wherein,p_tAnd p_t-1It is respectively indoor mobile robot current Position and a upper moment position,For first in t, the optimal path that indoor mobile robot is obtained by planning Path point, i.e. indoor mobile robot are at the pre-arriving at location in t+1 moment.

p_t-1By p_t-1=p_t-vt obtains, and v is the indoor mobile robot translational speed set.

Described step 2 is inserted indoor mobile robot place current location p_now(t)With target location p_goalBetween position p_k's Basis for selecting is as follows:

1) in target location occurs in the current investigative range of indoor mobile robotThen using target location as The position chosen during the 1st iteration: p₁=p_goal；From the beginning of the 2nd iteration i.e. k=2, p_kCondition need to be met

2) when target location does not appears in the current investigative range of indoor mobile robot, then position p in each iterative process_k For the position randomly selected from known map.

In described step 1, sonar sensor entrained by indoor mobile robot self refers to that used chassis Pioneer-2DX takes self The sonar contact radar of band.

Any two positions p in described known map_aAnd p_bBetween connection weightsCalculate as follows:

$R_{p_a,p_b}=d_{p_a,p_b}+k_m\×\max \left(\underset{C_{p_a,p_b}}{\∫}F\right(X,Y\left)ds\right);$

Wherein, k_mAffecting amplification coefficient for barrier and be conducive to improving the avoidance effect of robot, span is [50,1000].

The indoor mobile robot movement speed v set in described step 3 carries out self-adaptative adjustment as follows:

Wherein, p_barrierFor Obstacle Position, v₀The basal rate setting value moved for indoor mobile robot,Visit for current Survey operated within range spatial information set, including the Obstacle Position in current investigative range, aiming spot, and removable district Territory map position information.

Beneficial effect

Compared with prior art, it is an advantage of the current invention that:

1, the present invention uses the probabilistic Modeling of map barrier, the barrier detected by the indoor mobile robot side by probability Formula builds affects model, makes barrier that surrounding to constitute continuous print impact, and this impact is just for indoor moving machine The region that people had detected；

2, the present invention is compared with the most traditional mobile robot path planning, by indoor service is added map learning capacity, Allow the robot to ever-increasing cartographic information is analyzed study, obtain under current location best by greedy algorithm study Path planning.

3, invention provides for nonholonomic constraint when indoor mobile robot is planned and adaptive speed shift strategy, not only solve The real system application problem of path planning algorithm, and taken into account planning efficiency and the mobile security of indoor mobile robot.

Accompanying drawing explanation

Fig. 1 is the flow chart of the method for the invention；

Fig. 2 is indoor mobile robot hardware system structure schematic diagram；

Fig. 3 is indoor mobile robot nonholonomic constraint schematic diagram；

Fig. 4 is to use the method for the invention at the path planning design sketch dynamically and under static environment, and wherein, figure (a) strengthens Map study path planning hides design sketch during dynamic disorder；When figure (b) enhancing map study path planning hides static-obstacle Design sketch；Figure (c) strengthens design sketch when map study path planning is finally completed.

Detailed description of the invention

Below with reference to accompanying drawing, the present invention will be further described with being embodied as case.

As it is shown in figure 1, the present invention is a kind of indoor mobile robot strengthens map study paths planning method, including following Step:

Step 1: set up the probabilistic model that search coverage is affected by barrier；

First, by sonar sensor that indoor mobile robot is self-contained, it is thus achieved that the surrounding letter of indoor mobile robot Breath；Secondly, using indoor mobile robot the region of process as search coverage, set up according to described ambient condition information The probabilistic model that search coverage is affected by barrier, and according to the ambient condition information real-time update of sonar sensor Real-time Collection The probabilistic model that search coverage is affected by barrier；

In described step 1, sonar sensor entrained by indoor mobile robot self refers to that used chassis Pioneer-2DX takes self The sonar contact radar in eight directions, front of band, as shown in Figure 2.The obstacle information detected according to sonar sensor, structure Build that to detect the barrier impact probability model in region as follows:

$F(X,Y)=1-\underset{i=1}{\overset{M}{\mathrm{\Π}}}\[1-f_i(X,Y)\]---\left(1\right)$

Wherein, indoor mobile robot work space information collection is combined intoDescribed spatial information includes all target locations and owns Obstacle Position；In current investigative range, work space information collection is combined intoThe work space information collection detected is combined into

[the relative distance letter between barrier and robot current location that indoor mobile robot is arrived by sonar radar detection Breath, utilizes the self-contained mileage gauge of indoor mobile robot and inertial navigation system, is converted to relative distance information with initially Position is the positional information fastened of the base coordinate of initial point and range information.】

{ (X, Y) } is search coverage,?Map on be detected with M barrier, f_i(X, Y) is the influence function to indoor mobile robot Path selection of i-th barrier, uses normal state to divide Cloth is expressed as follows:

$f_i(X,Y)=\frac{1}{\sqrt{2\mathrm{\π}}{\mathrm{\σ}}_i}{e}^{(-\frac{{D_i}^2}{2{\mathrm{\σ}}_i})}---\left(2\right)$

Wherein, σ_iFor the coverage coefficient of i-th barrier, span is [0,1]；D_iFor i-th barrier to visiting The distance matrix of all positions in survey region, matrix size is in the same size for N × N with map；

Step 2: based on greedy algorithm and enhancing study iterative strategy, in maximum iteration time k set_maxIn, iteration updates to be worked as Front position p_now(t)With target location p_goalBetween path cost function, to reach the path corresponding to path cost function of convergence As the optimal path of current time, detailed process is as follows:

Step 2.1: make k=1, k represent iterations；

Step 2.2: judge that current iteration number of times has exceeded the maximum iteration time of setting the most, if exceeding, then returns step 2.1,； Otherwise, step 2.3 is entered；

Return step 2.1 and refer to from the beginning of first time iteration, after rebuilding initial path, be again introduced into iterated search optimal path；

Step 2.3: from the known map that the ambient condition information obtained is corresponding, randomly choose a position p_kInsert indoor moving Robot place current location p_now(t)With target location p_goal, calculate the path cost function that kth time iteration obtains as follows

$G_{p_{now\left(t\right)}}^{k+1}=\min \{G_{p_{now\left(t\right)}}^k,R_{p_{now\left(t\right)},p_k}+G_p^k\}---\left(3\right)$

Wherein,The path cost function that-1 iteration of kth obtains, and For Current location p on known map_now(t)With selected location p_kBetween connection weights,For use greedy algorithm obtain in kth Secondary iteration complete after obtain from position p_k-1To target location p_goalOptimal path cost function:

$G_p^k=\left\{\begin{array}{cc}{R}_{p_k,p_{goal}}+G_{p_{goal}}& k=1\\ \min \{R_{p_k,p_{goal}}+G_{p_{goal}},R_{p_k,p_{k-1}}+G_p^{k-1}\}& k\≠1\end{array}\right.---\left(4\right)$

Wherein,For target location p_goalCost function, and For on known map in kth time institute Select location point p_kWith-1 selected location p of kth_k-1Between connection weights；

Any two positions p in known map_aAnd p_bBetween connection weightsStraight with this by the air line distance between two positions The barrier passed on line affects probability and constitutes:

$R_{p_a,p_b}=d_{p_a,p_b}+k_m\×\max \left(\underset{C_{p_a,p_b}}{\∫}F\right(X,Y\left)ds\right)---\left(5\right)$

Wherein,For position p_aTo position p_bPhysical location distance between 2；For position p_aTo position p_bRoad Footpath point set, ds is path integral unit；Max () is that maximum finds a function；

Step 2.4: the path cost function obtained after judging iterationReach convergence, if convergence, then exited Iterative process, enters step 3 using path corresponding for the path cost function restrained as optimal path；Otherwise, by iteration time Number k adds 1, returns step 2.2；

Step 3: the optimal path obtained using step 2 is as the current preselected path of indoor mobile robot, according to preliminary election routing Footpath and the position of indoor mobile robot and velocity judge deflection angle φ of indoor mobile robot_tWhether meet wheeled robot Can not the nonholonomic constraint of lateral sliding, determine whether indoor mobile robot moves according to preselected path direction:

If being unsatisfactory for, then return step 2, rebuild all paths of current time；Owing to, in step 2.3, inserting indoor Mobile robot place current location p_now(t)With target location p_goalBetween position p_kFor randomly choosing, wherein p_kMeet condition{p_rIt is the path point set under iteration optimal path cost function before,For indoor moving machine People's work space information set.Therefore, after returning step 2, randomly choose the on position obtained, can be with choosing before The on position crossed is different, can obtain new path；

If meeting, moving with the optimal path direction that adaptive speed is cooked up along current time, entering the path rule of subsequent time Draw, t=t+1, return step 1, until indoor mobile robot moves to target location, complete path planning；

Generally, due to wheeled mobile robot cannot occur shifted laterally, it may be assumed that wheeled mobile robot is in motor process Middle without skidding, only make pure rolling.I.e. thinking that robot does not has component motion on its transverse axis, this nonholonomic constraint can be expressed as:

$\stackrel{\·}{x}\sin \mathrm{\θ}-\stackrel{\·}{y}cos\mathrm{\θ}=0$

Deflection angle φ of described indoor mobile robot_tWhether meet wheeled robot can not the nonholonomic constraint of lateral sliding refer to: φ_t∈ [0,60 °] ∪ [120 °, 180 °], as shown in Figure 3；

Wherein,p_tAnd p_t-1It is respectively indoor mobile robot current Position and a upper moment position,For first in t, the optimal path that indoor mobile robot is obtained by planning Path point, i.e. indoor mobile robot are at the pre-arriving at location in t+1 moment.

Indoor mobile robot moves with adaptive speed v:

Algorithm performance is analyzed

1, convergence

In study iterative strategy, by formula (3), (4) understandIf iterations is sufficiently large,Will be final Converge to a stable point.

Lemma 1 is for any point p in map_i,i∈N²With impact point p_goalBetween the cost function in pathBe one with Iterations k increases and the sequence of monotone decreasing.

From formula (4), the map location p chosen for the first time in step 2₁Will be when first time iteration as current location and target On position between position, forms initial path.From the beginning of second time iteration, new randomly selects positionAfter insertion, path cost function will choose one of minimum in two selections below:

(1) starting point newly inserted position p_kThe path line of impact point.

(2) starting point newly inserted position p_kIteration optimal cost function path { p before_r| the path line of r ∈ k}.

Theorem 1 assumes that the connection weights R perseverance of any two position is just, and the meeting of first time iteration produces current location and target The initial path cost function of positionUtilize formula (3) and formula (4) to update path cost function and there is following several character.

1, optional position and the path cost sequence of function of target locationCan finally restrain.

2, the cost function of target locationFor on map optional position and path cost letter between target location NumberTo keep stable after certain iterations.

3, from impact point p_goalStarting, one finds starting point surely in limited step.

Prove:

Character 1: from formula (5), the distance between any two diverse locationsi,j∈N²For just, accumulative on path Impact probability ∫_CF (X, Y) ds is also positive number, and therefore, the connection weights R of any two diverse locations is positive number, cost function G Also it is positive number.According to lemma 1, for any one p_now(t)And the cost function sequence in path between target locationAlong with The increase of iterations will converge on minimum border.

Character 2: initialized the cost function understanding impact point by cost functionPerseverance is 0, and do as one likes matter 1 understands,Will Final convergence, if iterations is more than the number of position on map, then the path between current location and the position of impact point is chosen Travel through full map, i.e. get optimal solution, continue iteration then cost function and keep minimum constant.

Character 3: according to learning strategy, choose current point and path between impact point after, this path comprises choose with Machine position sequence { p_r| r ∈ k} is inadvisable in next iteration.Therefore, all positions after certain iterations, in map Put and all can be selected, including starting point.

2, complexity analysis

In order to represent the complexity of algorithm, defined variable T represents the time (step number) of path finding, and concrete analysis of complexity is such as Shown in lower:

(1) if definition two-dimensional map size is N × N, then the connection weights between all positions of full map are calculatedComplexity For N⁴, wherein, i, j ∈ N²,i≠j。

(2) in learning strategy, robot is by constantly looking for the path cost function of minimum of each momentUsed here as The complexity of greedy search algorithm beWhereinAverage time for path finding.

(3) choosing when robot completes the path of t, after entering the t+1 moment, the connection weights complexity updating map is N⁴。

Therefore, indoor mobile robot is represented by from the Path complexity of origin-to-destination:

Hereinafter with a concrete application example, the operation of the present invention is described in detail, the enhancing map learning path of the present invention Planning algorithm is mainly used in the independent navigation of indoor mobile robot and avoidance planning, here mainly for static and dynamic two Impact point in class obstacle environment arrives at and embodies its performance.Specifically it is provided that

Experimental situation is set to grating map, map size N=50m, has dynamic and static two class obstacles concurrently, dynamically hinder in map Hindering and be initially located at (5,10), with speed 1m/s back and forth movement between (5,10) Yu (15,10), static-obstacle thing is L-shaped obstacle, Wherein three summits of static-obstacle thing 1 are respectively (18,20), and (25,20), (25,10), three summits of static-obstacle thing 2 are divided Not Wei (30,40), (30,35), (38,35).Indoor mobile robot can only be obtained by eight sonar sensors that its chassis is carried The map obstacle information in mobile eight directions, front.Barrier affects amplification coefficient k_m=500, barrier coverage σ_i=1, I=1,2 .., M, investigative range R of indoor mobile robot_detect=10m, basis movement speed v₀=1m/s, indoor moving machine People's starting point is (0,0), and impact point is (48,48).

As shown in Figure 4, give indoor moving indoor mobile robot and have the Real-time and Dynamic path dynamically and in static environment concurrently Planned trajectory.Wherein, figure (a) strengthens effect when map study path planning hides dynamic disorder；Figure (b) strengthens map Effect during static-obstacle is hidden in learning path planning；Figure (c) strengthens effect when map study path planning is finally completed；“*” Representing robot mobile route, "○" represents barrier, and the encapsulated coil around barrier represents barrier probability density, the closer to Barrier, probability density is the biggest, i.e. the blockade line number of turns is the most, and " " represents target location.It can be seen that providing mesh After cursor position, although the obstacle information in environment is obtained incomplete by indoor moving indoor mobile robot, and there is dynamic disorder, It relies on enhancing map learning algorithm to be capable of good target and arrives at and barrier avoiding function, has preferable planning efficiency and avoidance Function.

Claims (5)

1. an indoor mobile robot strengthens map study paths planning method, it is characterised in that include following step:

Step 1: set up the probabilistic model that search coverage is affected by barrier；

First, by sonar sensor that indoor mobile robot is self-contained, it is thus achieved that the surrounding letter of indoor mobile robot Breath；Secondly, using indoor mobile robot the region of process as search coverage, set up according to described ambient condition information The probabilistic model that search coverage is affected by barrier, and according to the ambient condition information real-time update of sonar sensor Real-time Collection The probabilistic model that search coverage is affected by barrier；

The probabilistic model that described search coverage is affected by barrier is as follows:

$F(X,Y)=1-\underset{i=1}{\overset{M}{\Π}}\[1-f_i(X,Y)\]$

Wherein, indoor mobile robot work space information collection is combined intoDescribed spatial information includes all target locations and owns Obstacle Position；In current investigative range, work space information collection is combined intoThe work space information collection detected is combined into

{ (X, Y) } is search coverage,?Map on be detected with M barrier, f_i(X, Y) is the influence function to indoor mobile robot Path selection of i-th barrier, uses normal state to divide Cloth is expressed as follows:

$f_i(X,Y)=\frac{1}{\sqrt{2\mathrm{\π}}{\mathrm{\σ}}_i}{e}^{(-\frac{{D_i}^2}{2{\mathrm{\σ}}_i})}$

Wherein, σ_iFor the coverage coefficient of i-th barrier, span is [0,1]；D_iFor i-th barrier to detecting The distance matrix of all positions in region, matrix size is in the same size for N × N with map；

Step 2: based on greedy algorithm and enhancing study iterative strategy, in maximum iteration time k set_maxIn, iteration updates to be worked as Front position p_now(t)With target location p_goalBetween path cost function, to reach the path corresponding to path cost function of convergence As the optimal path of current time, detailed process is as follows:

Step 2.1: make k=1, k represent iterations；

Step 2.2: judge that current iteration number of times has exceeded the maximum iteration time of setting the most, if exceeding, then returns step 2.1, Otherwise, step 2.3 is entered；

Step 2.3: from the known map that the ambient condition information obtained is corresponding, randomly choose a position p_kInsert indoor moving Robot place current location p_now(t)With target location p_goal, wherein p_kMeet condition{p_rFor it Path point set under front iteration optimal path cost function；Calculate the path cost function that kth time iteration obtains as follows

Wherein,The path cost function that-1 iteration of kth obtains, and For Current location p on known map_now(t)With selected location p_kBetween connection weights,For use greedy algorithm obtain in kth Secondary iteration complete after obtain from position p_kTo target location p_goalOptimal path cost function:

$G_p^k=\left\{\begin{array}{cc}{R}_{p_k,p_{goal}}+G_{p_{goal}}& k=1\\ min\{R_{p_k,p_{goal}}+G_{p_{goal}},R_{p_k,p_{k-1}}+G_p^{k-1}\}& k\≠1\end{array}\right.$

Wherein,For target location p_goalCost function, and For on known map in kth time institute Select location point p_kWith-1 selected location p of kth_k-1Between connection weights；

Any two positions p in known map_aAnd p_bBetween connection weightsStraight with this by the air line distance between two positions The barrier passed on line affects probability and constitutes:

Wherein,For position p_aTo position p_bPhysical location distance between 2；For position p_aTo position p_bRoad Footpath point set, ds is path integral unit；Max () is that maximum finds a function；

Step 2.4: the path cost function obtained after judging iterationReach convergence, if convergence, then exited Iterative process, enters step 3 using path corresponding for the path cost function restrained as optimal path；Otherwise, by iteration time Number k adds 1, returns step 2.2；

Step 3: the optimal path obtained using step 2 is as the current preselected path of indoor mobile robot, according to preliminary election routing Footpath and the position of indoor mobile robot and velocity judge deflection angle φ of indoor mobile robot_tWhether meet wheeled robot Can not the nonholonomic constraint of lateral sliding, determine whether indoor mobile robot moves according to preselected path direction:

If being unsatisfactory for, then return step 2；If meeting, cook up along current time with the indoor mobile robot translational speed set Optimal path direction move, enter the path planning of subsequent time, t=t+1, return step 1, until indoor moving machine People moves to target location, completes path planning；

Deflection angle φ of described indoor mobile robot_tWhether meet wheeled robot can not the nonholonomic constraint of lateral sliding refer to: φ_t∈[0,60°]∪[120°,180°]；

Wherein,p_tAnd p_t-1It is respectively indoor mobile robot current Position and a upper moment position,For first in t, the optimal path that indoor mobile robot is obtained by planning Path point, i.e. indoor mobile robot are at the pre-arriving at location in t+1 moment；

p_t-1By p_t-1=p_t-vt obtains, and v is the indoor mobile robot translational speed set.

Indoor mobile robot the most according to claim 1 strengthens map study paths planning method, it is characterised in that described Step 2 is inserted indoor mobile robot place current location p_now(t)With target location p_goalBetween position p_kBasis for selecting As follows:

1) in target location occurs in the current investigative range of indoor mobile robotThen using target location as The position chosen during the 1st iteration: p₁=p_goal；From the beginning of the 2nd iteration i.e. k=2, p_kCondition need to be met

2) when target location does not appears in the current investigative range of indoor mobile robot, then position p in each iterative process_k For the position randomly selected from known map.

Indoor mobile robot the most according to claim 2 strengthens map study paths planning method, it is characterised in that described In step 1, sonar sensor entrained by indoor mobile robot self refers to used sound self-contained for chassis Pioneer-2DX Detection radar.

Indoor mobile robot the most according to claim 3 strengthens map study paths planning method, it is characterised in that described Any two positions p in known map_aAnd p_bBetween connection weightsCalculate as follows:

$R_{p_a,p_b}=d_{p_a,p_b}+k_m\×max\left(\underset{C_{p_a,p_b}}{\∫}F(X,Y)ds\right);$

Wherein, k_mAffecting amplification coefficient for barrier and be conducive to improving the avoidance effect of robot, span is [50,1000].

5. strengthening map study paths planning method according to the indoor mobile robot described in any one of claim 1-4, its feature exists In, the indoor mobile robot movement speed v set in described step 3 carries out self-adaptative adjustment as follows:

Wherein, p_barrierFor Obstacle Position, v₀The basal rate setting value moved for indoor mobile robot,Visit for current Survey operated within range spatial information set, including the Obstacle Position in current investigative range, aiming spot, and removable district Territory map position information.