CN109164812B

CN109164812B - Mobile robot multi-behavior fusion enzyme numerical film control method in unknown environment

Info

Publication number: CN109164812B
Application number: CN201811235990.1A
Authority: CN
Inventors: 张葛祥; 黄振; 王学渊; 荣海娜
Original assignee: Southwest Jiaotong University
Current assignee: Southwest Jiaotong University
Priority date: 2018-10-23
Filing date: 2018-10-23
Publication date: 2020-04-07
Anticipated expiration: 2038-10-23
Also published as: CN109164812A

Abstract

The invention discloses a multi-behavior fusion enzyme numerical membrane control method for a mobile robot in an unknown environment, which calculates the straight-line distance and angle between the target and the robot, judges the environment where the robot is located, and judges whether the straight-line distance between the robot and the target is the smallest in history value, judging whether the target and the obstacle or wall are on the same side of the robot, making multi-behavior selection, executing the selected behavior, and judging whether the mobile robot reaches the target, etc. The invention combines the enzyme numerical membrane system and the multi-behavior fusion algorithm, and introduces the distance judgment between the robot and the target and the judgment of whether the target, obstacle or wall is on the same side of the robot, which solves the problem of autonomous walking of mobile robots in an unknown environment to a certain extent. Locking issues and bypassing issues. The enzyme numerical membrane controller is used to integrate various behavioral controllers, so that the robot can adapt to the complex environment.

Description

Mobile robot multi-behavior fusion enzyme numerical film control method in unknown environment

Technical Field

The invention relates to the technical field of intelligent robot control, in particular to a multi-behavior fusion enzyme numerical film control method for a mobile robot in an unknown environment.

Background

The membrane calculation (or membrane system) is the youngest branch of natural calculation, is a calculation model abstracted from the information processing cooperation mode of cell groups such as functions and structures of cells, organs and tissues and has excellent characteristics such as parallelism, uncertainty and distribution. The research shows that: theoretically, the membrane computing model has the same computing power as the Turing machine, and even has the possibility of exceeding the limitation of the Turing machine. The enzyme numerical membrane system is one of membrane systems, has the excellent characteristics of distribution, parallelism, easy programming and modularization, and is suitable for robot control.

Disclosure of Invention

The invention aims to provide a method for controlling a multi-behavior fusion enzyme numerical film of a mobile robot in an unknown environment. The mobile robot acquires environment information by using a sensor, classifies different environments, and selects and executes one of a plurality of behaviors according to different environments and the distance and the angle between a target and the robot until the target point is reached.

The technical scheme for realizing the purpose of the invention is as follows:

a method for controlling a multi-behavior fusion enzyme numerical value film of a mobile robot in an unknown environment is characterized by comprising the following steps

Step 1: calculating the linear distance CurDest between the target and the robot and the Angle between the target and the forward direction of the robot;

step 2: according to the information acquired by the sensor carried by the robot, the environment of the robot is judged, wherein the environment comprises a left wall, a right wall, a corridor, a left wall corner, a right wall corner, a front wall, an upper left side, an upper right side, two sides, dead corners and no barriers, and C is used for respectively_{i＝1,2,...,11}0, wherein i is 1,2, 11 corresponds to the 11 environments respectively; the environment is divided into 4 types: wall type E_waCorresponding to the environment of the left wall, the right wall, the corridor, the left wall corner and the right wall corner; obstacle type E_obCorresponding to the environment of the front wall, the upper left side and the upper right side; dead zone type E_deCorresponding to the environments of two sides and dead corners; obstacle-free type E_noCorresponding to the environment without obstacles;

and step 3: judging whether the straight-line distance CurDist between the robot and the target is the historical minimum MinDist: when MinDist > CurDes, the history is the minimum value, the variable IfMin is 1, and MinDist is made to be CurDes; otherwise, IfMin is 0; and 4, step 4: judging whether the target and the barrier or the wall surface are on the same side of the robot:

judging whether the target is on the left side or the right side of the robot according to the Angle, and when the Angle is greater than 0, the target is on the right side of the robot, and when the Angle is less than 0, the target is on the left side of the robot;

such as C ₁0 or C ₄0 or C₇If the value is 0, judging that the obstacle or the wall surface is positioned on the left side of the robot, and representing the obstacle or the wall surface by using a variable IfLeft as 1; such as C ₂0 or C ₅0 or C₈If the value is 0, judging that the obstacle or the wall surface is positioned on the right side of the robot, and representing the obstacle or the wall surface by using a variable IfLeft as 0;

when Angle is greater than 0 and IfLeft is 1, the target and the obstacle or the wall surface are not on the same side of the robot, and the variable ifSame is 0;

when Angle is greater than 0 and IfLeft is 0, the target and the obstacle or the wall surface are on the same side of the robot, and IfSame is 1; when Angle is <0 and IfLeft is 1, IfSame is 1; when Angle is less than 0 and IfLeft is 0, IfSame is 0;

and 5: after multi-behavior selection is carried out, behavior variables are output to an execution system:

when the robot is in C_iWhen the environment is 0, i is 9, 10, the type of the dead zone is determined, and the dead zone is selected

When the robot is in C_iWhen the environment is 0 and i is 11, judging as a barrier-free type: selecting Com if IfMin is 1_gr1, if ifMin is 0

When the robot is in C_iWhen the environment is 0, i is 6,7 and 8, the obstacle type is judged:

and if IfMin is equal to 0, selecting a corresponding obstacle avoidance behavior: i-6 selection

i-7 selection

i-8 selection

If ifMin is equal to 1, further judging whether the target and the obstacle are on the same side of the robot, and when Ifsame is equal to 1, selecting a corresponding obstacle avoidance behavior: i-6 selection

i-7 selection

i-8 selection

Selecting Com when IfSame is 0_gr＝1；

When the robot is in C_iWhen i is equal to 1,2, 3,4, 5, the environment is judgedThe wall surface type:

if IfMin is 0, selecting the corresponding wall-following behavior: i is selected from 1 or 4

i is 2 or 5 selected

i-3 selection

If ifMin is equal to 1, further judging whether the target and the wall surface are on the same side of the robot, and if Ifsame is equal to 1, selecting a corresponding wall-following behavior: i is selected from 1 or 4

i is 2 or 5 selected

i-3 selection

Selecting Com when IfSame is 0_gr＝1；

The above behavior variables are respectively: front obstacle avoiding barrier

Obstacle avoidance device for left obstacle

Obstacle avoidance device for right obstacle

Left side wall face following wall

Right side wall surface following wall

Crossing channel

Tendency target Com_grIn-situ U-turn Com_deAnd self-rotation

Step 6: the execution system executes the behavior output in the step 5;

and 7: judging whether the mobile robot reaches the target, namely judging whether CurDest is equal to 0: if the value is equal to 0, the robot reaches the target, and the control is ended; if not, returning to the step 1 for continuation.

Further, in the step 2, the method for determining the environment of the robot based on the information acquired by the sensor mounted on the robot includes: the information acquired by the sensor is 8 distances d around the robot_xX is 1,2 … 8; wherein d is₁、d₂Distance, d, obtained for the sensor at the right front side of the robot₃Distance, d, obtained for the sensor on the right side of the robot₄Distance, d, obtained for the sensor on the right rear side of the robot₅Distance, d, obtained for the sensor at the left rear side of the robot₆Distance, d, obtained for the sensor on the left side of the robot₇、d₈The distance obtained for the sensor located on the left front side of the robot; when d is_x<t and t are distance threshold values, the sensor detects the obstacle and sends the corresponding distance d_xBinarization, wherein 1 represents that an obstacle is detected and 0 represents that the obstacle is not detected;

and then determining the environment of the robot according to the table.

The invention provides a multi-behavior fusion enzyme numerical value membrane control method for the field of robot control and membrane system control in an unknown environment, which combines an enzyme numerical value membrane system and a multi-behavior fusion algorithm, introduces the judgment of the distance between a robot and a target and the judgment of whether the target and an obstacle (or a wall surface) are on the same side of the robot, and solves the problems of deadlock and winding of autonomous walking of a mobile robot in the unknown environment to a certain extent. And an enzyme numerical value membrane controller is adopted to fuse various behavior controllers, so that the robot can adapt to complex environments.

Drawings

FIG. 1 is a multi-behavior fusion flow diagram;

FIG. 2 is a diagram of eleven robot environment modes;

FIG. 3 is a robot distance sensor profile;

FIG. 4 is a diagram of a multi-activity fusion enzyme numerical membrane of the present invention;

FIG. 5 is a graph of experimental results comparing the present invention and fuzzy logic control in a G-type and convoluted obstacle environment.

Detailed Description

The following further describes the embodiments of the present invention with reference to the drawings.

A multi-behavior fusion enzyme numerical membrane control flow chart is shown in FIG. 1.

The specific related technologies adopted by the invention are as follows:

1) placing a mobile robot in an unknown environment with the unknown environment and a robot target as input, given a target point (x) to be reached by the robot_g,y_g). According to the distance formula

Calculating the distance CurDest between the target and the mobile robot and calculating the Angle of the target relative to the forward direction of the robot according to an Angle calculation formula (taking the right side of the forward direction of the robot as the positive direction).

2) Control of mobile robot

a) Acquisition and determination of the environment in which a mobile robot is located

In order to make a mobile robot cope with a complicated environment, the present invention divides the environment type in which the mobile robot is located into (as shown in fig. 2 below): the wall comprises a left wall, a right wall, a corridor, a left wall corner, a right wall corner, a front wall, an upper left side, an upper right side, two sides, dead corners and no obstacles. With C_i＝1,...,11Is represented by (C)_i＝1,...,110 denotes in the corresponding ambient mode: left wall, right wall, corridor, left wall corner, right wall corner, front wall, upper left side, upper right side, two sides, dead corner, no obstacle).

The mobile robot judges the environment of the robot according to the information acquired by the mounted sensor. Specifically, for example, a mobile robot is equipped with 8 distance sensors (see fig. 3), and the 8 sensors are distributed in a ring shape on the mobile robot. 8 sensors can acquire 8 distance information (d) around the robot₁、d₂…d₈) (wherein d is₁、d₂Distance information obtained for a sensor located on the right front side of the robot, d₃Distance information obtained for sensors on the right side of the robot, d₄Distance information obtained for the sensor on the right rear side of the robot, d₅Distance information obtained for the sensors on the left rear side of the robot, d₆Distance information obtained for the sensor on the left side of the robot, d₇、d₈Distance information obtained for a sensor located on the front left side of the robot). Setting a threshold value t when d_x<When t (x is 1,2 … 8), the sensor is considered to detect an obstacle, and the corresponding distance value d is set_tBinarized, 1 represents detected obstacle and 0 represents not detected. After the processing, the binary sensor distance value of the current environment where the robot is located can be obtained. Eleven environments correspond to eleven binary sensor distance values (see table 1). And comparing the current binary sensor distance value with binary sensor distance values corresponding to eleven environment types, judging the current environment, and outputting the current environment to the multi-behavior fusion enzyme numerical value membrane system. Enzyme numerical membrane systems classify the input environment into 4 broad categories: the wall type (left wall, right wall, corridor, left corner and right corner environment correspond to the type), the barrier type (front wall, upper left side and upper right side environment correspond to the type), the dead zone type (two sides and dead angle environment correspond to the type), and the barrier-free type (barrier-free environment corresponds to the type). Respectively using variable E_wa、E_ob、E_de、E_noAnd (4) showing.

TABLE 1 Environment type and binary sensor value correspondence table

b) Judging whether the distance between the robot and the target is the historical minimum value or not

The historical minimum value between the robot and the target is stored by using a variable MinDist (initialized by using CurdIst), and whether the distance between the robot and the target CurdIst is the minimum value or not is judged by comparing the CurdIst and the MinDist. Min dist > curdest, is the minimum history value (indicated by variable IfMin ═ 1), and min dist ═ curdest.

c) Determine whether the target, the obstacle (or the wall surface) are on the same side of the robot

The input Angle can determine whether the target is on the left or right side of the robot. Angle>0, target on the right side of the robot; angle<0, target on the left side of the robot. The position of the obstacle (or wall) can be determined by the input environment C_i＝1,...,11And (4) determining. C ₇0 represents the upper left environment, that is, it is determined that the obstacle is located on the left side of the robot (represented by variable IfLeft 1); c₈An upper right environment is indicated by 0, i.e. an obstacle is located on the right side of the robot (indicated by the variable IfLeft being 0). C ₁0 denotes left wall environment or C₄The left corner is represented by 0, and the wall surface is judged to be positioned on the left side of the robot (represented by a variable IfLeft which is 1); c ₂0 represents the right wall or C₅The right corner is indicated by 0, and it is determined that the obstacle is located on the right side of the robot (indicated by the variable IfLeft being 0).

Based on the above determination, it can be determined whether the target and the obstacle (or the wall surface) are on the same side of the robot. When Angle >0 and IfLeft ═ 1, the target, the obstacle (or the wall) are not on the same side of the robot (represented by the variable IfSame ═ 0); angle >0, if IfLeft is 0, IfSame is 1; when Angle is less than 0 and IfLeft is equal to 1, IfSame is equal to 1; when Angle is less than 0 and IfLeft is 0, IfSame is 0. Summarized as the following formula

d) Multiple behavior selection

The invention defines five behaviors, namely obstacle avoidance (obstacle avoidance of front, left and right obstacles), wall following (wall following and passing through channels of left and right side walls), target tendency, in-situ turning and autorotation. Respectively using obstacle avoidance action variables

Wall-dependent behavior variables

Tendency target behavior variable Com_grIn-situ turn-around behavior variable Com_deSelf-rotation behavior variables

And (4) showing.

When a specific condition is satisfied, a corresponding behavior is selected (i.e., the corresponding behavior variable value is assigned 1), and a behavior variable is output.

When the robot is in C_iWhen the environment is 0(i is 9 or 10), the environment type is judged to be the dead zone type, the robot is blocked due to the advancing direction, and the in-situ turning behavior is selected

When the robot is in C₁₁When the environment is 0, the obstacle-free type is judged. If IfMin is equal to 1, then the trend target behavior Com is selected_gr1 is ═ 1; if IfMin is equal to 0, selecting the rotation behavior

When the robot is in C_iWhen the environment is 0(i is 6,7, 8), it is determined as the obstacle type. If IfMin is equal to 0, selecting corresponding obstacle avoidance behavior

If IfMin is equal to 1, judging whether the target and the obstacle are on the same side of the robot, and if IfSame is equal to 1, selecting a corresponding obstacle avoidance behavior

When IfSame is equal to 0, the tendency toward target behavior Com is selected_gr＝1。

When the robot is in C_iWhen the environment is 0(i is 1,2, 5), it is determined as the wall type. If IfMin is equal to 0, selecting corresponding wall-following behavior

If IfMin is equal to 1, judging whether the target and the wall surface are on the same side of the robot, and if IfSame is equal to 1, selecting the corresponding wall-following behavior

The multi-behavior fusion enzyme numerical membrane system outputs behavior variables to an execution system.

3) Perform corresponding actions

The behavior of the enzyme numerical membrane system output is performed.

4) Determining whether the mobile robot reaches a target

And judging whether the distance value CurDest is equal to 0. If the value is equal to 0, the mobile robot reaches the target, and the control is ended. If not, the mobile robot continues to acquire the environment state and determines the execution behavior until the target is reached.

The invention is simulated on a PC, and the CPU is 2.8HZ, 4GB RAM, software platforms MATLAB2012, Windows7OS and Webots. Taking the mobile robot Epuck as an example, the Epuck robot has 8 infrared sensors and is driven by two wheels in a differential mode.

Referring to fig. 4 and 5, the invention adopts the following steps:

step 1, taking unknown environment and robot target as input

The mobile robot Epuck is placed in the unknown environment shown in figure 5, and the Epuck obtains the distance D between the current Epuck and the target according to a distance and angle calculation formula_curAnd Angle, input to the multi-behavior fusion enzyme numerical membrane system of fig. 4. Angle passes through variable A_gr1、A_gr2Preservation, A_gr1＝Angle，A_gr2＝-Angle。

Step 2. control of the mobile robot

a) Acquisition and determination of the type of environment in which a mobile robot is located

The environmental modes are divided into 11 types: left wall, right wall, corridor, left wall corner, right wall corner, front wall, left upper side, right upper side, two sides, dead corner, no obstacle, using C_i＝1,...,11Is represented by (C)_i＝1,...,110 is in the corresponding environment mode: left wall, right wall, corridor, left wall corner, right wall corner, front wall, upper left side, upper right side, two sides, dead corner, no obstacle). The Epuck acquires 8 pieces of distance information (sensor value d) around the robot through 8 infrared sensors₁、d₂…d₈). The 8 sensors are distributed on the mobile robot in a ring shape (as in fig. 3), (where d₁、d₂Distance information obtained for a sensor located on the right front side of the robot, d₃Distance information obtained for sensors on the right side of the robot, d₄Distance information obtained for the sensor on the right rear side of the robot, d₅Distance information obtained for the sensors on the left rear side of the robot, d₆Distance information obtained for the sensor on the left side of the robot, d₇、d₈Distance information obtained for a sensor located on the front left side of the robot). Threshold value 70 when d_x>70(x is 1,2 … 8), the sensor is considered to detect the obstacle, and d is set to_x1, d_x<70，d_xA 0 indicates that no obstacle was detected. (the value of the Epuck sensor decreases as the distance of the obstacle increases). The 8 sensor processed values constitute a set of binary sensor values, such as: when the disease is not obstructed [ 00000000]Comparing the currently obtained binary set with the defined 11 environment binary set (see table 1), and determining the environment C of the robot_i＝1,...,11And output to the enzyme numerical membrane system of figure 4. Enzyme numerical membrane systems classify the input environment into 4 broad categories: wall type (left wall, right wall, corridor, left corner and right corner environment correspond to the type), barrier type (front wall, upper left side and upper right side environment correspond to the type), dead zone type (two sides and dead angle environment correspond to the type), and barrier-freeType (no obstacle environment corresponds to this class). Respectively using variable E_wa、E_ob、E_de、E_noAnd (4) showing. This step is accomplished by the Judge Environment model of FIG. 4, which contains a total of 11 rules Pr_i＝1,2..11And Case. Rules 1-5 (i.e., Pr)_i＝1,2..5Case) divides the environment in which the robot is located into wall types (E)_wa1), rule 6,7,8 is divided into obstacle types (E)_obRule 9, 10 is divided into dead zone types (dead zone type directly corresponds to in-place turnaround behavior, so there is a direct allocation Com implemented_de1), rule 11 is divided into barrier-free types (E)_no＝1)。

Here, the rule execution is described as rule 9, and the rule execution manner is the same hereinafter. Rule 9 is Pr₉,Case:C₉+2(E_c→)1|Com_de+1|E_T。Pr₉Case is a regular name, separated from expressions by colon, C₉+2 is the value generation rule, 1| Com_de+1|E_TValue assignment rule, enzyme variable E in parentheses_c(when an enzyme variable is greater than a variable in a value generation rule, the rule will only activate execution; where E is required_c＞C₉). When E is_c＞C₉When rule 9 is active, the value generation rule generates a value C₉+2, after the value generation rule is executed, the variable in the generation rule is cleared to 0 (i.e., C)₉Will be cleared to 0)), the resulting value is assigned by an assignment rule, C₉+2 divided into 2 portions, one for each variable Com_deOne portion of the solution is given to E_TI.e. Com_de＝(C₉+2)/2，E_T＝(C₉+2)/2。

This step is accomplished in the fig. 4 subparagraph distanceifminimal. Rule 1 determines whether the value is the minimum historical value, and rule 2 outputs and stores the minimum historical value.

Rule

1,2 execution, D _min1 is expressed as the historical minimum, Output D_cur。

And judging whether the target and the obstacle are on the same side of the robot or not is completed by rules 2-7,10 and 11 in the submembrane Judge RobotStateObstacle in the figure 4.

Rules

4,5 determine the position of the target and the robot. A. the_gr1<0 (i.e., Angle)<0) When, the enzyme variable E_a[0]>A_gr1The set of rules 4 is activated, the rule is activated,

indicating that the target is on the left side of the robot. A. the_gr2<0 (i.e., Angle)>0) When, the enzyme variable E_a[0]>A_gr2The set of rules 5 is activated,

indicating that the target is to the right of the robot.

Rule Pr_2、3,obstacle_nUsed for judging the position relation between the obstacle and the robot. When the environment mode C₇When the value is 0 (upper left side), the obstacle is located at the left side of the robot, and the rule Pr is activated₂,obstacle_n: variable O_left[-1]＝1，O_right[-1]0; environmental mode C₈When the value is 0 (upper right side), the obstacle is located at the right side of the robot, and the rule Pr is activated₃,obstacle_n: variable O_right[-1]＝1，O_left[-1]＝0。

Rule Pr_6.7.10.11,obstacle_nBy passing

O_left、O_rightThe relationship between the obstacle, the target and the robot is obtained. When O is present_left1 (obstacle to the left of the robot),

(target on robot right) indicating that obstacle is on robot left target right, rule Pr is activated₆,obstacle_n: enzyme variable EOG _lr2; when O is present_right＝1，

When, indicating that the right target is on the left, rule Pr is activated₇,obstacle_n: enzyme variable EOG _rl2; when O is present_left＝1，

When the system is in use, the system indicates that the obstacles and the targets are all on the left, and activates the rule Pr₁₀,obstacle_n: enzyme variables Eog _left2; when O is present_right＝1，

When the system is in use, the system indicates that the obstacles and the targets are all right, and activates the rule Pr₁₁,obstacle_n: enzyme variables Eog_right＝2。

And judging whether the target and the wall surface are on the same side of the robot or not by rules 2-7,10 and 11 in JudgGeRobotStatewall. The principle is the same, and therefore, the description is omitted.

d) Selection of multiple behaviors

The invention defines 5 behaviors, namely obstacle avoidance (obstacle avoidance of front, left and right obstacles), wall following (wall following and passing through channels of left and right side walls), target tendency, in-situ turning and rotation behaviors. Respectively using obstacle avoidance action variables

Wall-dependent behavior variables

Shown (as in fig. 4). A behavior variable equal to 1 indicates that such behavior variable is selected.

This is accomplished in part by rules 9 or 10 in the Judge Environment model of FIG. 4. Variable E_TAnd (4) stopping the evolution calculation of the membrane system and outputting

When the robot is in C₁₁When the environment is 0, the type is judged to be barrier-free (E)_no1). At this time, if D _min1, choose to trend toward target behavior Com _gr1 is ═ 1; if D is_minWhen the rotation behavior is 0, the rotation behavior is selected

This part is completed by the daughter membrane SelectGoalReachingCase of FIG. 4. If D is_min Rule 1 cannot be executed, and further rule 2 cannot be executed, rule 3 has no enzyme variables and is executed every time it evolves,

further rule 4 execution, Com_gr＝1，E _T1 denotes the evolutionary calculation of the stop membrane system. If D is_min Rule 1, rule 2 performs, 0,

E_T＝1。

when the robot is in C_iWhen the environment is 0(i is 6,7, 8), it is determined as the obstacle type (E)_ob1). FIG. 4 daughter membranes SelectObstacleAvoidecense and Judge RobotStateObstacle judge what behavior to choose. If D is_min Rule 1 execution in SelectObstacleAvoideceCase 0

Rules

2 or 3 or 4 are then implemented to select the corresponding obstacle avoidance behavior

After rule 5 is executed

Making the rules in the judggerootstateobstacle unexecutable. If D is_min1, rule 1 cannot be executed in selectobstacleedamyces case, further,

rules

2,3, and 4 cannot be executed, and after rule 5 is executed

If the target and the obstacle are on the same side of the robot (Eog)_left2 or Eog_right2), rule 12 or 13 is activated,

or

If the target and the obstacle are not on the same side of the robot (EOG)_lr2 or EOG_rl2), rule 8 or 9 is active,

E_T＝1。

when the robot is in C_iWhen the environment is 0(i is 1,2, 5), the environment type is determined to be the wall type (E)_wa1). FIG. 4 the daughter membranes SelectWallFollowCase and JudggeRobotStatewall judge which behaviour is selected. The principle is the same as the obstacle avoidance situation, and therefore, the description is omitted.

Step 3, executing corresponding behaviors

The behavior of the membrane system output is performed.

Step 4, judging whether the mobile robot reaches the target or not

Judging the distance value D_curWhether or not it is equal to 0. If the value is equal to 0, the mobile robot reaches the target, and the control is ended. If not, the mobile robot continues to acquire the environment state and determines the execution behavior until the target is reached.

From the experimental results of fig. 5, it can be seen that the multi-behavior fusion membrane control (MBCMC) walking path is shorter (fig. 5 left) and the dead zone (fig. 5 right) that the fuzzy logic control cannot escape can be escaped.

Claims

1. a mobile robot multi-behavior fusion enzyme numerical membrane control method under an unknown environment, is characterized in that, comprising

Step 1: Calculate the straight-line distance CurDist of the target and the robot and the angle of the target relative to the forward direction of the robot;

Step 2: According to the information obtained by the sensors carried by the robot, determine the environment where the robot is located, including left wall, right wall, corridor, left wall corner, right wall corner, front wall, upper left side, upper right side, both sides, dead corner and no Obstacles are represented by C _{i=1, 2,..., 11} =0 respectively, where i=1, 2,..., 11 correspond to the above 11 environments respectively; the environments are divided into 4 categories: Wall type E _wa , corresponding to the environment of left wall, right wall, corridor, left wall and right corner; obstacle type E _ob , corresponding to the environment of front wall, upper left and upper right; dead zone type E _de , corresponding to both sides and Dead-end environment; no obstacle type E _no , corresponding to the environment without obstacles;

Step 3: Determine whether the straight-line distance between the robot and the target, CurDist, is the historical minimum value MinDist: when MinDist>CurDist, the historical minimum value is represented by the variable IfMin=1, and MinDist=CurDist; otherwise, IfMin=0;

Step 4: Determine whether the target and the obstacle or wall are on the same side of the robot:

Judging whether the target is on the left or right side of the robot according to the angle, when Angle>0 the target is on the right side of the robot, when Angle<0 the target is on the left side of the robot;

If C ₁ =0 or C ₄ =0 or C ₇ =0, judge that the obstacle or wall is located on the left side of the robot, which is represented by the variable IfLeft=1;

If C ₂ =0 or C ₅ =0 or C ₈ =0, judge that the obstacle or wall is located on the right side of the robot, which is represented by the variable IfLeft=0; when Angle>0, IfLeft=1, the target and the obstacle or wall are Not on the same side of the robot, represented by the variable IfSame=0; when Angle>0, IfLeft=0, the target and the obstacle or wall are on the same side of the robot, IfSame=1; When Angle<0, IfLeft=1, IfSame=1 ;Angle<0, IfLeft=0, IfSame=0;

Step 5: After multi-behavior selection, output behavior variables to the execution system:

When the robot is in the environment of C _i = 0, i = 9, 10, it is judged to be a dead zone type, and select

When the robot is in the environment of C _i = 0, i = 11, it is judged as an obstacle-free type: if IfMin=1, select Com _gr =1, if IfMin=0, select

When the robot is in the environment of C _i = 0, i = 6, 7, 8, it is judged as an obstacle type:

If IfMin=0, select the corresponding obstacle avoidance behavior: i=6 select

i=7 choices

i=8 options

If IfMin=1, further judge whether the target and the obstacle are on the same side of the robot.

The corresponding obstacle avoidance behavior: i=6 option

i=7 choices

i=8 options

When IfSame=0 select Com _gr =1;

When the robot is in the environment of C _i = 0, i = 1, 2, 3, 4, and 5, it is judged as a wall type:

If IfMin=0 select the corresponding wall behavior: i=1 or 4 select

i=2 or 5 choice

i=3 options

If IfMin=1, further judge whether the target and the wall are on the same side of the robot. When IfSame=1, select the corresponding wall-following behavior: i=1 or 4 to select

i=2 or 5 choice

i=3 options

When IfSame=0 select Com _gr =1;

The above behavior variables are: obstacle avoidance ahead

Obstacle avoidance on the left

Obstacle avoidance on the right

left side wall

right side wall

crossing the channel

Trending Target Com _gr , Spot U-turn Com _de and Rotation

Step 6: The execution system executes the behavior outputted in Step 5;

Step 7: Determine whether the mobile robot has reached the target, that is, determine whether CurDist is equal to 0: if it is equal to 0, it means that the robot has reached the target, and the control is ended; if it is not 0, return to step 1 to continue.

2. The multi-behavior fusion enzyme numerical membrane control method for a mobile robot in an unknown environment as claimed in claim 1, wherein in the step 2, according to the information obtained by the sensor carried by the robot, the environment where the robot is located is judged The method is: the information obtained by the sensor is 8 distances d _x around the robot, x=1,2...8; where d ₁ and d ₂ are the distances obtained by the sensor on the right front side of the robot, and d ₃ is the right side of the robot. The distance obtained by the sensor on the left side of the robot, d ₄ is the distance obtained by the sensor on the right rear side of the robot, d ₅ is the distance obtained by the sensor on the left rear side of the robot, d ₆ is the distance obtained by the sensor on the left side of the robot, and d ₇ and d ₈ are The distance obtained by the sensor located on the left front side of the robot; when d _x < t, t is the distance threshold, the sensor detects an obstacle, and the corresponding distance d _x is binarized, 1 means an obstacle is detected, 0 means no detection obstacle;

Then determine the environment where the robot is located according to the above table.