CN103699136B - Intelligent home service robot system based on the algorithm that leapfrogs and method of servicing - Google Patents

Abstract

The present invention provides an intelligent home service robot system and service method based on leapfrog algorithm. The robot uses a voice recognition module or a wireless communication module to receive instructions from service objects, determine the target to be grasped and the location of the service object, and use The environment map is used to locate the target and the service object; the robot uses the improved leapfrog algorithm to formulate the path planning of the robot according to the environment map, and calculates all the best position points to be passed from the current position of the robot to the target; the robot controls the motor driver to drive the robot to When reaching the target, the manipulator is driven to move to the target position to grab the target, and automatically runs to the service object to complete the service work. The intelligent home service robot system proposed by the present invention has simple structure, good intelligence and flexibility, can save computing time and storage capacity, improves the accuracy of robot path planning, and improves the work efficiency of the home service robot.

Description

Intelligent home service robot system and service method based on leapfrog algorithm

technical field

The invention relates to a home service robot, belongs to the field of robot path planning in an unknown environment, is an application combining robot technology and swarm intelligence technology, and particularly relates to an intelligent home service robot system and method based on leapfrog algorithm.

Background technique

With the continuous development of service robot technology and home application technology, home service robots have attracted more and more people's attention, and are getting closer to people's daily life, and have broad application prospects.

In the research of mobile robot-related technologies, path planning is an important link and a key topic. It is a hot issue in the field of mobile robot research, and it is also a very challenging problem in mobile robot navigation. At present, there are many methods of path planning, such as artificial potential field method, genetic algorithm, neural network method, ant colony algorithm and so on. However, there are several defects in the traditional artificial potential field method: there is a trap area; paths cannot be found between similar obstacles; swings in narrow passages; when there are obstacles near the target point, the robot cannot reach the target point; the genetic algorithm has Good global search ability, but the disadvantages of slow calculation speed, large storage space, and easy to fall into premature maturity; neural network method has good learning ability, but when there are many obstacles and the environment is dynamic, the network structure is huge and neural The threshold value of the element needs to be changed continuously with the change of time; the convergence speed of the ant colony algorithm is slow, and it is easy to fall into the local optimum.

Leaping Frog Algorithm (SFLA) is a group-based sub-heuristic collaborative search group intelligence algorithm proposed by Eusuff et al. in 2003. This algorithm combines the advantages of memetic algorithm and particle swarm optimization algorithm. It is a new attempt to apply the leapfrog algorithm to path planning. By imitating the behavior of frog predation, the robot can find a path to the goal, but the algorithm is easy to fall into local optimum.

Contents of the invention

The invention provides an intelligent home service robot system and method based on leapfrog algorithm. After acquiring and processing the indoor environment information, the improved leapfrog algorithm is used to plan the path of the robot, so that the robot can plan a path to avoid obstacles to reach the target position after receiving the corresponding instructions, and grab the objects required by the service object. The goal, and sent to the hands of the service object, to complete the task.

The technical scheme of the present invention to realize the above-mentioned purpose of the invention is: an intelligent home service robot system based on leapfrog algorithm, characterized in that: comprising

A walking robot driven by controllable left and right wheels, the robot has a manipulator driven by a motor driver;

The front end of the robot carries an airborne camera for scanning the height information of the target object;

One or more ultrasonic sensors are arranged directly in front of the robot to detect obstacles in the progress of the robot;

The robot receives the instruction of the service object through the wireless communication module or the voice recognition module, and determines the target;

It also includes a camera installed on the indoor roof to take pictures of the indoor environment. The camera sends the pictures taken to the robot through the wireless network; Self position positioning, and then path planning through the improved leapfrog algorithm, according to the calculation of the best position the robot will pass through and the obstacles detected by the robot, the target position and its own position, calculate the angular velocity of the left and right wheels of the robot, and control the robot reach the target point.

The intelligent home service robot service method based on leapfrog algorithm is characterized in that: comprising the following steps:

(1) The robot carries a storage device, an ultrasonic sensor, an airborne camera, a voice recognition module, a wireless communication module, a motor driver and a manipulator;

(2) The camera installed on the indoor roof acquires the indoor environment image, and sends it to the robot for processing, establishes an environmental map, and completes the initial position positioning of the robot;

(3) The robot establishes its own kinematics model, and its motion state variable is (x, y, θ) ^T , where (x, y) is the coordinate of the robot in the plane coordinate system, and θ is the direction angle of the robot;

(4) The robot uses the speech recognition module or wireless communication module to receive instructions from the service object, determines the target to be grasped and the location of the service object, and uses the environmental map to locate the target and the service object;

(5) According to the environment map, the robot uses the improved leapfrog algorithm to formulate the path planning of the robot, and calculates all the best position points to pass from the current position of the robot to the target;

(6) The robot calculates the angular velocity of its left and right wheels according to its own motion model and the next best position to be reached, and controls the motor driver to drive the robot to move forward;

(7) The robot judges whether it has reached the target. If it does not return to step (6), if it reaches the target, it uses the onboard camera to perform three-dimensional scanning, obtains the vertical height of the target from the ground, and drives the manipulator to move to the target position to grab the target;

(8) After the robot grabs the target, it automatically runs to the service object to complete the service work.

The step (2) indoor roof is equipped with a camera to obtain the indoor environment image, and send it to the robot for processing, to set up the environment map, and to complete the initial position positioning of the robot refers to:

(2a) A camera is installed on the indoor roof to capture indoor environment images, and the captured pictures are sent to the robot through the wireless communication module. Good common object shape and floor color style knowledge to build an environmental grid map, if a certain grid does not contain any obstacles, it is a free grid, otherwise it is an obstacle grid; both free space and obstacles can be represented form a set of grid blocks, and mark the obstacle grid set as O;

(2b) Cartesian coordinate method is used to mark the grid: the upper left corner of the grid map is taken as the coordinate origin, the horizontal direction to the right is the positive direction of the X-axis, and the vertical downward direction is the positive direction of the Y-axis. Each grid interval corresponds to A unit length on the coordinate axis, any grid is uniquely identified by Cartesian coordinates (x, y), so that the environment map is represented by a two-dimensional matrix map(p, q):

The concrete steps of described step (3) are:

(3a) Let x and y be the abscissa and ordinate of the robot in the plane coordinate system respectively, θ is the direction angle of the robot, v is the velocity of the center of mass of the robot, and ω is the angular velocity of the robot, then the nonholonomic constraint of the robot motion is obtained as The dynamic function of the robot center of mass motion is:

$\left\{\begin{array}{c}\stackrel{<span\; class="notranslate">\; \&Center\; Dot;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\\ \stackrel{<span\; class="notranslate">\; \&CenterDot;</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\\ \stackrel{<span\; class="notranslate">\; \&CenterDot;</span>}{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

(3b) The discrete-time model of robot motion can be obtained from formula (1), as follows:

$<span\; class="notranslate">\; \{</span>\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; T</span>}\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; \&CenterDot;</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; T</span>}\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>\\ \mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&Delta;</span>}\mathrm{<span\; class="notranslate">\; T</span>}\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span>\end{array}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

According to the motion principle of rigid body kinematics, the motion of the robot is controlled by the angular velocity of the left and right wheels, namely:

$\left[\begin{array}{c}\stackrel{<span\; class="notranslate">\; \&CenterDot;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}\\ \stackrel{<span\; class="notranslate">\; \&Center\; Dot;</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}\\ \stackrel{<span\; class="notranslate">\; \&Center\; Dot;</span>}{\mathrm{<span\; class="notranslate">\; \&theta;</span>}}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}\frac{\mathrm{<span\; class="notranslate">\; r</span>}\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; r</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; r</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}& \frac{\mathrm{<span\; class="notranslate">\; r</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&theta;</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ \frac{\mathrm{<span\; class="notranslate">\; r</span>}}{\mathrm{<span\; class="notranslate">\; d</span>}}& <span\; class="notranslate">\; -</span>\frac{\mathrm{<span\; class="notranslate">\; r</span>}}{\mathrm{<span\; class="notranslate">\; d</span>}}\end{array}\right]\left[\begin{array}{c}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; R</span>}}\\ {\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; L</span>}}\end{array}\right]<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Among them, ω _L and ω _R are the angular velocity of the left and right wheels of the robot, r is the radius of the robot wheel, and d is the distance between the two wheels.

Concrete steps in the described step (5) are:

(4a) Parameter initialization: set the total number of frogs in the population as N, the number of subpopulations as k, and the number of frogs in the subpopulation as n, satisfying N=k*n; the number of local search iterations is L, and the number of global iterations G, the maximum step size allowed by the frog to move is S _max ;

(4b) Generate the initial frog group: randomly generate N frogs as the initial frog group P={X ₁ ,X ₂ ,...X _N }, the jth frog represents the jth solution, use X _j =(x _j1 , x _j2 ,...x _j s) represents, where, 0≤j≤N, s represents the dimension of each solution;

(4c) Calculate the fitness value: define the fitness function as follows:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; o</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Calculate the fitness value of each frog according to the formula (4), where ω ₁ and ω ₂ are constants, ||·|| is to calculate the Euclidean distance between them, O _j represents the obstacle, T represents the target position; when X _i is closer to the target, the value of ||X _i -T|| is smaller, and when X _i is farther away from the obstacle, the value of min||X _i -O _j || will be becomes larger, so that the value of f(X _i ) becomes smaller;

(4d) Divide the frog population: sort the N frogs according to their fitness values from good to bad, record the frog with the best global fitness as X _g ; divide the whole frog population into k groups by random grouping, each group Contains n frogs, satisfying N=kn;

(4e) Local search: record the frog with the best fitness in each subpopulation as X _b ＝(x _b1 ,x _b2 ,…,x _b s), and the frog with the worst fitness as X _ω ＝(x _ω1 ,x _ω2 ,…,x _ωs ), and then use the median strategy to update the frog individuals with the worst fitness in the subpopulation, and repeat the update process until the set number of iterations L is reached, then stop the local search;

(4f) After the local search is completed, the frogs of all groups are remixed, and the current best frog is recorded, and then step (4d) and step (4e) are performed, and this operation is repeated until the set global iteration number G is reached.

The method of random grouping in the step (4d) refers to:

In the entire frog population, the first k frogs are randomly entered into k populations, and each frog can only enter one population, and then, the k+1 to 2k frogs randomly enter k populations, and each frog can only enter one population , the 2k+1 to 3k frogs randomly enter k groups, and each frog can only enter one group, and so on, until all the frogs are allocated.

In the step (4e), the worst individual frog is updated using the median strategy, and the specific steps are as follows:

(6a) Suppose the center point of each group is X _z ＝(x _z1 ,x _z2 ,…,x _zs ), then

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; z</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

S＝rand×(X _z -X _ω ) (6)

newX _ω ＝X _ω +S,-S _max ≤S≤S _max (7)

Among them, rand represents a random number between 0 and 1, S represents the adjustment vector of the individual frog, S _max represents the maximum step size that the frog is allowed to move, and newX _ω represents the worst solution after the update;

(6b) Recalculate the fitness value of the new solution to determine whether it has been improved, if improved, replace X _ω with newX _ω ; if not improved, replace the formula (6) with the optimal solution X _g of the entire population X _z in re-updates the worst solution; if the new solution is still not improved, the worst solution X _ω is randomly updated.

In described step (6), concrete steps are as follows:

(7a) The robot calculates all the best positions to reach the target point according to the leapfrog algorithm, and the robot uses formula (1) and formula (2) to calculate and

(7b) The robot obtains according to the calculation and Use formula (3) to calculate the angular velocity ω _L and ω _R of the left and right wheels of the robot, and control the movement of the robot.

The present invention has the following advantages:

(1), the intelligent home service robot system proposed by the present invention adopts the indoor roof installation camera to collect indoor environmental information, establishes an environmental map, has a simple structure, and can effectively complete home service work;

(2), the present invention adopts leapfrog algorithm to realize indoor service robot path planning, expands the application range of bionic technology, and improves the working reliability and safety of home service robot;

(3), the present invention adopts random grouping method and median strategy to improve leapfrog algorithm, has strengthened the optimization ability of algorithm, has better balanced global search and local search ability of algorithm;

(4), the present invention proposes an intelligent home service robot system and method based on the leapfrog algorithm. This method can imitate the characteristics of frog predation, has good intelligence, flexibility, can save computing time and storage capacity, and greatly improve The accuracy of robot path planning is improved, and the work efficiency of home service robots is improved.

Description of drawings

Fig. 1 is a block diagram of hardware equipment of the present invention;

Fig. 2 is the flow chart of robot path planning based on leapfrog algorithm in the present invention;

Fig. 3 is the leapfrog algorithm flow chart improved among the present invention;

Fig. 4 is a schematic diagram of the working principle of the smart home service robot based on the leapfrog algorithm in the present invention.

detailed description

The present invention will be further described below in conjunction with the accompanying drawings.

A block diagram of a hardware device implementing the present invention is shown in Figure 1, including a robot, several ultrasonic sensors, a storage device, an airborne camera, a speech recognition module, a motor driver, a manipulator, and a wireless communication module. The robot carries equipment such as storage devices, ultrasonic sensors, onboard cameras, voice recognition modules, wireless communication modules, motor drivers and manipulators. Among them, the robot carries two wheels to control the operation of the robot; ultrasonic sensors are evenly distributed in front of the robot to detect obstacles in the progress of the robot; the robot receives service objects through a wireless communication module or a voice recognition module instructions to determine the target. The camera installed on the indoor roof takes pictures of the indoor environment and sends them to the robot through the wireless network. After processing, the robot builds an indoor grid map, and locates the target, service object and the robot itself, and then uses the improved leapfrog algorithm. Path planning, calculate the angular velocity of the left and right wheels of the robot according to the calculation of the best position the robot will pass through, the obstacles detected by the robot, the target position, and its own position, so as to control the robot to reach the target point, and then pass the robot itself The carried airborne camera scans the height information of the target object, drives the manipulator to grab the object, then plans the path to the service object, and delivers the object to the service object to complete the task.

The present invention is an intelligent home service robot system and method based on leapfrog algorithm, and its specific process is shown in Figure 2, including the following steps:

(1) The robot carries equipment such as storage devices, ultrasonic sensors, onboard cameras, voice recognition modules, wireless communication modules, motor drivers, and manipulators;

(2) The camera installed on the indoor roof acquires the indoor environment image, and sends it to the robot for processing, establishes an environmental map, and completes the initial position positioning of the robot;

(3) The robot establishes its own kinematics model, and its motion state variable is (x, y, θ) ^T , where (x, y) is the coordinate of the robot in the plane coordinate system, and θ is the direction angle of the robot;

(4) The robot uses the speech recognition module or wireless communication module to receive instructions from the service object, determines the target to be grasped and the location of the service object, and uses the environmental map to locate the target and the service object;

(5) According to the environment map, the robot uses the improved leapfrog algorithm to formulate the path planning of the robot, and calculates all the best position points to pass from the current position of the robot to the target;

(6) The robot calculates the angular velocity of its left and right wheels according to its own motion model and the next best position to be reached, and controls the motor driver to drive the robot to move forward;

(7) The robot judges whether it has reached the target. If it does not return to step (6), if it reaches the target, it uses the onboard camera to perform three-dimensional scanning, obtains the vertical height of the target from the ground, and drives the manipulator to move to the target position to grab the target;

(8) After the robot grabs the target, it automatically runs to the service object to complete the service work.

The search method of the leapfrog algorithm improved among the present invention, its concrete process is as shown in accompanying drawing 3, comprises the following steps:

(1) Parameter initialization. Suppose the total number of frogs in the population is N, the number of sub-populations is k, and the number of frogs in the sub-population is n, satisfying N=k*n. The number of local search iterations is L, the number of global iterations is G, and the maximum step size allowed by the frog to move is S _max .

(2) Generate the initial frog group. Randomly generate N frogs as the initial frog group P={X ₁ ,X ₂ ,...X _N }, the jth (0≤j≤N) frog represents the jth solution, use X _j =(x _j1 , x _j2 ,...x _js ), and s represents the dimension of each solution.

(3) Calculate the fitness value. Define the fitness function as follows:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; o</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Calculate the fitness value of each frog according to the formula (4), where ω ₁ and ω ₂ are constants, ||·|| is to calculate the Euclidean distance between them, O _j represents the obstacle, T represents the target position; when X _i is closer to the target, the value of ||X _i -T|| is smaller, and when X _i is farther away from the obstacle, the value of min||X _i -O _j || will be becomes larger, so that the value of f(X _i ) becomes smaller.

(4) Divide frog populations. Sort the N frogs according to their fitness values from good to bad, and record the frog with the best global fitness as X _g . The entire frog population is divided into k groups by random grouping, and each group contains n frogs, satisfying N=kn.

(5) Local search. Note that the frog with the best fitness in each subpopulation is X _b = (x _b1 ,x _b2 ,…,x _bs ), and the frog with the worst fitness is X _ω =(x _ω1 ,x _ω2 ,…,x _ωs ) , and then use the median strategy to update the frog individuals with the worst fitness in the subpopulation, repeat the update process, and stop the local search of each subpopulation until the set iteration number L is reached.

(6) After the local search is completed, the frogs of all groups are remixed, and the current best frog is recorded, and then step (4d) and step (4e) are performed, and this operation is repeated until the set global iteration number G is reached.

The method of random grouping in the described step (4) refers to:

In the entire frog population, the first k frogs are randomly entered into k populations, and each frog can only enter one population, and then, the k+1 to 2k frogs randomly enter k populations, and each frog can only enter one population , the 2k+1 to 3k frogs randomly enter k groups, and each frog can only enter one group, and so on, until all the frogs are allocated.

In the step (5), the median strategy is used to update the worst frog individual, and the specific steps are as follows:

(5a) Suppose the center point of each subpopulation is X _z =(x _z1 ,x _z2 ,…,x _zs ), then

${\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; z</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

S＝rand×(X _z -X _ω ) (6)

newX _ω ＝X _ω +S,-S _max ≤S≤S _max (7)

Among them, rand represents a random number between 0 and 1, S represents the adjustment vector of the individual frog, S _max represents the maximum step size that the frog is allowed to move, and newX _ω represents the worst solution after the update.

(5b) Recalculate the fitness value of the new solution to determine whether it has been improved, if improved, replace X _ω with newX _ω ; if not improved, replace the formula (6) with the optimal solution X _g of the entire population X _z in re-updates the worst solution; if the new solution is still not improved, the worst solution X _ω is randomly updated.

The schematic diagram of the intelligent home service robot system and method based on leapfrog algorithm in the present invention is shown in Fig. The best position, the construction of the fitness function in the leapfrog algorithm can ensure that the fitness value of the point where the target is located is the smallest, and the fitness value of the obstacle is the largest, so that according to this model, the trajectory of the robot will be a path that can automatically The best path to avoid obstacles, avoid collisions with obstacles, and quickly reach the required target. The robot arrives at the target, gets the required object, and then returns to the original path, delivering the object to the service object to complete the task.

The intelligent home service robot system proposed by the present invention adopts indoor roof-installed cameras to collect indoor environmental information and establishes environmental maps. The scope of application improves the reliability and safety of home service robots. By improving the leapfrog algorithm, the optimization ability of the algorithm is strengthened, and the global search and local search capabilities of the algorithm are better balanced. The intelligent home service robot system and method based on the leapfrog algorithm proposed by the present invention can imitate the characteristics of frog predation, has good intelligence and flexibility, can save computing time and storage capacity, and greatly improves the accuracy of robot path planning To improve the working efficiency of the home service robot, it can better complete the work of the home service robot.

Claims (8)

1. Intelligent home service robot system based on the algorithm that leapfrogs, it is characterised in that: include

One is driven, by left and right wheel of controlled motion, the robot walked, and robot has the mechanical hand driven by motor driver；

Described robot carries airborne photographic head in front end, for scanning the elevation information of target object；

Described robot dead ahead arranges one or more ultrasonic sensor, the obstacle during advancing for sniffing robot Thing；

Described robot receives the instruction of service object by wireless communication module or sound identification module, determines target；

Also including that one is arranged on indoor roof for shooting the photographic head of indoor environment picture, photographic head will be clapped by Wi-Fi The picture taken the photograph is sent to robot；After picture is processed by robot, set up indoor grille map, and carry out target, service Object and robot self-position location, then carry out path planning by the algorithm that leapfrogs, according to being calculated institute of robot Barrier, target location and the self-position that optimum position to be passed through and robot probe arrive, calculates a left side for robot The angular velocity of right wheel, controls robot and arrives at impact point.

2. Intelligent home service robot method of servicing based on the algorithm that leapfrogs, it is characterised in that: comprise the steps:

(1) robot carries storage device, ultrasonic sensor, airborne photographic head, sound identification module, radio communication mold Block, motor driver and mechanical hand；

(2) indoor environment image is obtained by the photographic head installed by indoor roof, and sends and process to robot, builds Vertical environmental map, completes robot initial position location；

(3) robot sets up displacement model, and its kinestate variable is (x, y, θ)^T, wherein (x is y) that robot is flat Coordinate in areal coordinate system, θ is the deflection that robot advances；

(4) robot utilizes sound identification module or wireless communication module to receive the instruction from service object, determines needs The target captured and the position of service object, and utilize environmental map to carry out target and service object location；

(5) robot is according to environmental map, utilizes the path planning of the algorithmization robot that leapfrogs improved, calculates from machine People current location is to target all optimum positions point to be passed through；

(6) robot is according to displacement model and next optimum position point to be arrived, and calculates its left and right angle speed taken turns Degree, controls motor driver driven machine people and travels forward；

(7) robot judges whether to arrive target, without returning step (6), if arriving target, then utilizes airborne shooting Head carries out stereoscan, obtains target vertical height overhead, drives robot movement to capture target to target location；

(8), after robot captures target, automatically run at service object, complete services.

Intelligent home service robot method of servicing based on the algorithm that leapfrogs the most according to claim 2, it is characterised in that: Described step (2) indoor roof is provided with photographic head and obtains indoor environment image, and sends and process to robot, Set up environmental map, complete robot initial position location and refer to:

(2a) indoor roof is provided with photographic head and carries out indoor environment image taking, and by the picture that photographs through radio communication mold Block is sent to robot, after robot receives picture, first divides image with equivalently-sized grid, and according to thing The conventional body form first stored and the color pattern knowledge on floor set up environment grating map, if in some grid not Containing any barrier, then it is free grid, otherwise is barrier grid；Free space and barrier all can be expressed as grid block Set, is designated as O by barrier grid collection；

(2b) use method of direct coordinate that grid is identified: with the grating map upper left corner as zero, horizontal right direction For X-axis positive direction, vertically downward direction is Y-axis positive direction, a unit length on each grid interval respective coordinates axle, Any one grid all with rectangular coordinate (x, y) uniquely identifies, thus by environmental map with a two-dimemsional number matrix map (p, q) Represent:

Intelligent home service robot method of servicing based on the algorithm that leapfrogs the most according to claim 2, it is characterised in that: Concretely comprising the following steps of described step (3):

(3a) setting x, y is respectively robot horizontal stroke in plane coordinate system, vertical coordinate, and θ is the deflection of robot, and v is robot The speed of barycenter, ω is the angular velocity of robot, then the nonholonomic constraint obtaining robot motion is The kinematic function of robot center of mass motion is:

$\left\{\begin{array}{c}\stackrel{\&CenterDot;}{x}=v\left(t\right)cos\mathrm{\&theta;}\left(t\right)\\ \stackrel{\&CenterDot;}{y}=v\left(t\right)sin\mathrm{\&theta;}\left(t\right)\\ \stackrel{\&CenterDot;}{\mathrm{\&theta;}}=\mathrm{\&omega;}\left(t\right)\end{array}\right.---\left(1\right)$

(3b) discrete time model of robot motion can be obtained by formula (1), as follows:

$\left\{\begin{array}{c}x\left(t_{k+1}\right)=x\left(t_k\right)+\mathrm{\&Delta;}T\&CenterDot;v\left(t_k\right)\&CenterDot;\cos \left(\mathrm{\&theta;}\right(t_k)+\mathrm{\&Delta;}T\mathrm{\&omega;}(t_k)/2)\\ y\left(t_{k+1}\right)=y\left(t_k\right)+\mathrm{\&Delta;}T\&CenterDot;v\left(t_k\right)\&CenterDot;\sin \left(\mathrm{\&theta;}\right(t_k)+\mathrm{\&Delta;}T\mathrm{\&omega;}(t_k)/2)\\ \mathrm{\&theta;}\left(t_{k+1}\right)=\mathrm{\&theta;}\left(t_k\right)+\mathrm{\&Delta;}T\mathrm{\&omega;}\left(t_k\right)\end{array}\right.---\left(2\right)$

Wherein, Δ T is to use sample time interval, and according to the motion principle of rigid body kinematics, robot moves through left and right wheels Angular velocity control, it may be assumed that

$\left[\begin{array}{c}\stackrel{\&CenterDot;}{x}\\ \stackrel{\&CenterDot;}{y}\\ \stackrel{\&CenterDot;}{\mathrm{\&theta;}}\end{array}\right]=\left[\begin{array}{cc}\frac{rcos\mathrm{\&theta;}}{2}& \frac{rsin\mathrm{\&theta;}}{2}\\ \frac{rsin\mathrm{\&theta;}}{2}& \frac{rsin\mathrm{\&theta;}}{2}\\ \frac{r}{d}& -\frac{r}{d}\end{array}\right]\left[\begin{array}{c}{\mathrm{\&omega;}}_R\\ {\mathrm{\&omega;}}_L\end{array}\right]---\left(3\right)$

Wherein ω_LAnd ω_RBeing respectively the left and right angular velocity taken turns of robot, r is the radius of wheels of robot, and d is two wheels The distance length of countershaft.

Intelligent home service robot method of servicing based on the algorithm that leapfrogs the most according to claim 2, it is characterised in that: Concretely comprising the following steps in described step (5):

(4a) parameter initialization: the number setting Rana nigromaculata in population total is k as N, sub-population number, and in sub-population, the number of Rana nigromaculata is n, Meet N=k*n；Local Search iterations is L, and global iterative number of times is G, and the maximum step-length of the allowed displacement of Rana nigromaculata is S_max；

(4b) initial frog group is generated: N Rana nigromaculata of stochastic generation is as initial frog group P={X₁,X₂,...X_N, jth Rana nigromaculata represents Jth solution, uses X_j=(x_j1,x_j2,...x_js) represent, wherein, 0≤j≤N, s represent the dimension of each solution；

(4c) fitness value is calculated: definition fitness function is as follows:

$f\left(X_i\right)={\mathrm{\&omega;}}_1\&CenterDot;e^{-\min \left|\right|X_i-O_j\left|\right|}+{\mathrm{\&omega;}}_2\&CenterDot;\left|\right|X_i-T\left|\right|---\left(4\right)$

The fitness value of every Rana nigromaculata, wherein ω is calculated according to formula (4)₁、ω₂For constant, | | | | it is to calculate between the two Euclidean distance, O_jRepresent is barrier, and T represents target location；Work as X_iTime close to target, | | X_i-T | | value relatively Little, work as X_iFrom barrier farther out time, min | | X_i-O_j| | value will become big, so that f (X_i) value diminish；

(4d) Rana nigromaculata group is divided: N Rana nigromaculata is sorted to bad by good according to fitness value, the Rana nigromaculata that note overall situation fitness is best For X_g；Whole frog group is divided into k group by the method using random packet, and each group comprises n Rana nigromaculata, meets N=kn；

(4e) Local Search: remember that the Rana nigromaculata that in every sub-population, fitness is best is X_b=(x_b1,x_b2,…,x_bs), fitness is worst Rana nigromaculata be X_w=(x_w1,x_w2,…,x_ws), then in antithetical phrase population, the Rana nigromaculata individuality of worst fitness uses intermediate value strategy to carry out Update operation, repeat renewal process, until after reaching the iterations L set, just stopping the Local Search of each sub-population；

(4f) after Local Search completes, then the Rana nigromaculata of all groups is re-mixed, write down the most optimal Rana nigromaculata, then perform Step (4d) and step (4e), repeat this operation until reaching the global iterative number of times G set.

Intelligent home service robot method of servicing based on the algorithm that leapfrogs the most according to claim 5, it is characterised in that: In described step (4d), the method for random packet refers to:

In whole frog group, front k Rana nigromaculata is the most random enters k group, and every Rana nigromaculata can only enter a group, then, kth + 1 to 2k Rana nigromaculata enters k group at random, and every Rana nigromaculata can only enter a group, and 2k+1 is the most random to 3k Rana nigromaculata Entering k group, every Rana nigromaculata can only enter a group, the like, until all of Rana nigromaculata is all assigned.

Intelligent home service robot method of servicing based on the algorithm that leapfrogs the most according to claim 5, it is characterised in that: Described step (4e) uses intermediate value policy update to worst Rana nigromaculata individuality, specifically comprises the following steps that

(6a) central point of each group is set as X_z=(x_z1,x_z2,…,x_zs), then

$x_{zj}=\underset{i=1}{\overset{n}{\&Sigma;}}{x}_{ij}/n---\left(5\right)$

S=rand × (X_z-X_w) (6)

newX_w=X_w+S,-S_max≤S≤S_max (7)

Wherein, rand represents the random number between 0 and 1, and S represents the adjustment vector that Rana nigromaculata is individual, S_maxRepresent that Rana nigromaculata allows mobile Maximum step-length, newX_wRepresent and update later worst solution；

(6b) recalculating the fitness value of obtained new explanation, it is judged that whether it is improved, if improved, then using newX_wTake For X_w；Without improvement, then by whole population optimal solution X_gReplace the X in formula (6)_zAgain worst solution is updated；If new explanation The most do not improve, then to worst solution X_wRandomly update.

Intelligent home service robot method of servicing based on the algorithm that leapfrogs the most according to claim 2, it is characterised in that: In described step (6), specifically comprise the following steps that

(7a) robot is calculated according to the algorithm that leapfrogs and arrives impact point all optimum positions to be passed through, robot according to Current location and next step position, utilize formula (1) and formula (2) to calculate With

(7b) robot is according to calculatedWithFormula (3) is utilized to calculate the angular velocity of left and right wheels of robot ω_LAnd ω_R, control robot motion.