CN-121979022-A - Robot sampling planning method and system based on environment extremum position estimation

CN121979022ACN 121979022 ACN121979022 ACN 121979022ACN-121979022-A

Abstract

The invention provides a robot sampling planning method and a system based on environment extremum position estimation, which relate to the technical field of robot control and comprise the steps of constructing a model for an environment to be monitored through a Gaussian Markov random field, and mapping a sampable point in the environment into a random variable in the environment model; the method comprises the steps of calculating mutual information values of each non-sampling point and other non-sampling points based on an environment model, obtaining an environment extremum position estimation result through a Gaussian process fitting and a Kriging estimation method according to observed values of the sampled points, obtaining a next active sampling point through a probability decision method integrating mutual information and a space distance according to the mutual information values of the non-sampling points and the extremum position estimation result, and improving accuracy and rapidity of robot sampling planning based on a Gaussian Markov random field and based on sampling information gain.

Inventors

LI TENG
ZHU JIAXIN
WU TIANHAO
ZHAO RUKUN

Assignees

山东大学

Dates

Publication Date: 20260505
Application Date: 20251225

Claims (10)

1. The robot sampling planning method based on the environment extremum position estimation is characterized by comprising the following steps of: Building a model for an environment to be monitored through a Gaussian Markov random field, and mapping a sampable point in the environment into a random variable in the environment model; calculating mutual information values of each non-sampling point and other non-sampling points based on an environment model; according to the observed value of the sampled point, obtaining an environmental extremum position estimation result by a Gaussian process fitting and Kriging estimation method; and obtaining the next active sampling point by a probability decision method integrating the mutual information and the space distance according to the mutual information value and the extremum position estimation result of the non-sampling point.
2. A method of planning a sampling of a robot based on an environmental extremum position estimate as defined in claim 1, wherein the environmental model is formulated as: Wherein, the As a model of the environment, Representing sampable points The corresponding random variable is used to determine the random number, As a function of the mean value of the function, In the form of a matrix of precision, Is a gaussian markov random field.
3. The robot sampling plan method based on the environmental extremum position estimation of claim 1, wherein the mutual information value is formulated as: Wherein, the As an un-sampled point Is used to determine the mutual information value of the (c), As a set of the sampled points, For all the set of sampable points, To remove The rest of the non-sampled point sets, det, represent determinant calculations, Is an accuracy matrix.
4. The robot sampling plan method based on the environmental extremum position estimation of claim 1, wherein the environmental extremum position estimation result is expressed as: Wherein, the In order to estimate the location of the resulting environmental extremum, Is a sampling point set Is provided with a plurality of grooves, wherein the grooves are arranged at any position of the groove, Respectively mean function, covariance vector and covariance matrix, Representing the value of the sampled value, A base matrix representing a polynomial function, Is obtained by a generalized least square method in the Kriging method.
5. The robot sampling plan method based on the environmental extremum position estimation of claim 1, wherein the probability decision method is formulated as: Wherein, the As an un-sampled point The probability of being selected as an alternative sampling point, Representing sampling points And the estimated position of the environmental extremum The Euclidean distance between the two electrodes, Is a probability dependent parameter; to replace the probability threshold of the sampling point, Is the sampling point And selecting the alternative sampling point with the largest MI value from the alternative sampling points as the next active sampling point.
6. A method of planning a sampling of a robot based on an environmental extremum position estimate as recited in claim 5, wherein said probability dependent parameters And dynamically adjusting in the sampling process, namely setting preset smaller values at initial time to ensure that sampling points are uniformly distributed, gradually increasing along with sampling, and gradually concentrating the sampling points towards the estimated extreme value position.
7. A robotic sampling planning system based on an environmental extremum position estimate, comprising: An environment model construction module configured to construct a model for an environment to be monitored by a gaussian markov random field, and map sampable points in the environment to random variables in the environment model; the mutual information value calculation module is configured to calculate the mutual information value of each non-sampling point and the rest non-sampling points based on the environment model; The extremum position estimating module is configured to obtain an environment extremum position estimating result through a Gaussian process fitting and kriging estimating method according to the observed value of the sampled point; The optimal position decision module is configured to obtain the next active sampling point by a probability decision method of fusing the mutual information and the space distance according to the mutual information value and the extremum position estimation result of the non-sampling point.
8. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements a robot sampling planning method based on an environmental extremum position estimation according to any of claims 1-6.
9. A non-transitory computer readable storage medium storing computer instructions which, when executed by a processor, implement a robot sampling planning method based on an environmental extremum position estimate according to any of claims 1-6.
10. An electronic device comprising a processor, a memory and a computer program, wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute a robot sampling planning method for performing an environmental extremum based position estimation according to any of claims 1-6.

Description

Robot sampling planning method and system based on environment extremum position estimation Technical Field The invention relates to the technical field of robot control, in particular to a robot sampling planning method and system based on environment extremum position estimation. Background The intelligent mobile robot can sense and monitor the leakage condition of chemicals in real time through environment monitoring, if leakage occurs, early warning can be sent out at the first time and an environment extreme value caused by the leakage of the chemicals can be searched, the process can be regarded as the problem of locating the extreme point of relevant parameters in an environment field, the robot needs to actively sense the environment and search for locating the extreme point of the environment, the extreme point locating problem of the intelligent mobile robot in a complex dynamic environment is researched, the capability of the robot for troubleshooting the danger is improved, and the intelligent mobile robot has important significance in pushing the application of the intelligent mobile robot in environment inspection. In the process of positioning the extreme point of the environment, the intelligent mobile robot actively samples the environmental data of the surrounding environment based on the sensor, and guides the robot to sample through the extreme point estimation algorithm, the current robot sampling planning method has poor effect under the condition that the extreme point is at the boundary, and the quantity of sampling points required for positioning the extreme point is large, so that the efficiency is low. Disclosure of Invention In order to solve the problems, the invention provides a robot sampling planning method and a system based on environment extremum position estimation, wherein the environment where an intelligent mobile robot is positioned is described as a Gaussian Markov random field (Gaussian Markov Random Field, GMRF), mutual information (Mutual Information, MI) is used as the measurement of information quantity, and the next active sampling point is determined by combining the spatial distance between an un-sampled point and the estimated extremum position, thereby taking the accuracy and the rapidity into consideration. According to some embodiments, the present invention employs the following technical solutions: a robot sampling planning method based on environment extremum position estimation comprises the following steps: Building a model for an environment to be monitored through a Gaussian Markov random field, and mapping a sampable point in the environment into a random variable in the environment model; calculating mutual information values of each non-sampling point and other non-sampling points based on an environment model; And according to the observed value of the sampled point, obtaining an environmental extremum position estimation result by a Gaussian process fitting and Kriging estimation method. And obtaining the next active sampling point by a probability decision method integrating the mutual information and the space distance according to the mutual information value and the extremum position estimation result of the non-sampling point. According to some embodiments, the present invention employs the following technical solutions: A robotic sampling planning system based on an environmental extremum position estimate, comprising: An environment model construction module configured to construct a model for an environment to be monitored by a gaussian markov random field, and map sampable points in the environment to random variables in the environment model; the mutual information value calculation module is configured to calculate the mutual information value of each non-sampling point and the rest non-sampling points based on the environment model; And the extremum position estimation module is configured to obtain an environment extremum position estimation result through a Gaussian process fitting and kriging estimation method according to the observed value of the sampled point. The optimal position decision module is configured to obtain the next active sampling point by a probability decision method of fusing the mutual information and the space distance according to the mutual information value and the extremum position estimation result of the non-sampling point. According to some embodiments, the present invention employs the following technical solutions: A computer program product comprising a computer program which when executed by a processor implements the described method of robotic sample planning based on an environmental extremum position estimate. According to some embodiments, the present invention employs the following technical solutions: A non-transitory computer readable storage medium for storing computer instructions which, when executed by a processor, implement the method of robotic sampling planning based on environment