CN111640455A - Grassland state identification method based on audio identification - Google Patents

Abstract

The grassland state identification method based on audio identification comprises a motion platform capable of moving autonomously, wherein the bottom of the motion platform is provided with a driving wheel and a supporting wheel, the bottom of the motion platform is provided with a mowing device and a pickup adopting an audio signal, a processor for performing centralized control is arranged in the motion platform and is connected with the pickup, and the grassland state identification method is arranged in the processor and comprises a first filter module for identifying and filtering environmental noise; a second filter module for identifying and filtering the noise of the driving wheel; a third filter module that identifies and filters noise from the lawn mowing device; the grass sound identification module is used for identifying the grass sound emitted when the grass cutting device interacts with the grass stems and the grass blades; and the motion control module is planned according to the grassland state motion and is connected with the driving wheel.

Description

Grassland state identification method based on audio identification

Technical Field

The invention relates to a grassland state identification method based on audio identification, and belongs to the field of pattern identification.

Background

For an intelligent system, a negative feedback link in closed-loop control plays a very important role. The controller outputs a control action for eliminating the deviation after operation according to a deviation signal obtained by comparing the feedback signal with the given signal, so that the controlled object is equal to the given value. However, the existing mowing robot does not realize closed-loop control and is low in efficiency. Because the lawn mowing robot cannot detect the lawn state, the lawn mowing robot can only move on the flat ground without purpose, and does not know where the lawn is, where the lawn is not, and whether the lawn at the current position is cut or not. The careless behavior mode has low efficiency, the mowing robot needs to work on the lawn for a long time and is frequently charged, the battery configuration is large, and the manufacturing cost of the mowing robot is directly increased. If the state of the lawn can be identified, the efficiency of the mowing robot can be greatly improved through reasonable working mode selection.

Disclosure of Invention

Aiming at the problems, the invention provides a grassland state identification scheme based on audio identification by adopting a method for identifying the characteristic sound of interaction of a grass cutting device and grass leaves and grass stems.

The technical scheme adopted by the invention for solving the technical problems is as follows:

the utility model provides a grassland state identification method based on audio identification, includes the motion platform that can independently remove, the bottom of motion platform set up drive wheel and supporting wheel, can realize the autonomous movement, motion platform bottom set up grass cutting device and adopt audio signal's adapter, the motion platform inside set up the treater that carries out centralized control, with the adapter be connected, the treater can gather the audio signal of adapter output, the treater inside set up the grassland state identification method, confirm the grassland state of the region of motion platform, grassland state identification method include:

-a first filter module capable of identifying and filtering out ambient noise;

-a second filter module able to identify and filter the noise of said driving wheel;

-a third filter module capable of identifying and filtering noise of said mowing device;

-a grass recognition module capable of recognizing the sound produced by the grass cutting device when interacting with the grass stems and grass blades;

-a motion control module for performing a motion planning based on grass conditions, said motion control module being connected to said drive wheel, said motion control module driving said motion platform into a slow mode when said grass recognition module output is active; when the output of the grass sound identification module is invalid, the motion control module drives the motion platform to enter a fast mode.

The invention has the following beneficial effects: the characteristic sound of interaction of the mowing device with the grass blades and the grass stalks is identified, the state of the lawn is obtained, and a closed-loop control system is formed, so that the efficiency of the mowing robot is greatly improved.

Drawings

FIG. 1 is a flow chart of a grassland state identification method of the present invention.

Detailed Description

The invention is further described below with reference to the accompanying drawings:

referring to fig. 1, the grassland state identification method based on audio recognition comprises a moving platform capable of moving autonomously, wherein driving wheels and supporting wheels are arranged at the bottom of the moving platform, and the moving platform can move straight, retreat and rotate at any angle.

The bottom of the motion platform is provided with a mowing device and a sound pick-up adopting an audio signal. The mowing device is set as a mowing rope and can also be set as a mowing cutter. The adapter be used for gathering grass cutting device act on the sound that grass leaf and grass stem sent, can gather environmental noise simultaneously, drive wheel and grass cutting device's the noise that the motor sent.

The motion platform inside set up the treater that carries out centralized control, with the adapter be connected, the treater can gather the audio signal of adapter output. The processor is internally provided with a grassland state identification method for confirming the grassland state of the area of the motion platform, and the grassland state identification method comprises the following steps:

-a first filter module capable of identifying and filtering out ambient noise;

the presence of various noises in the environment, such as the automobile driving nearby, the barking of animals, the speaking voice of pedestrians, etc., can be identified and filtered.

-a second filter module able to identify and filter the noise of said driving wheel;

the driving wheel is driven by a motor, the motor can generate noise in the rotating process, and the noise of the motor with fixed power and structure is also fixed frequency and can be identified and filtered.

-a third filter module capable of identifying and filtering noise of said mowing device;

the mowing device is driven by the motor, the motor can generate noise in the rotating process, and the noise of the motor with fixed power and structure is also fixed in frequency and can be identified and filtered.

-a grass recognition module capable of recognizing the sound produced by the grass cutting device when interacting with the grass stems and grass blades;

the grass cutting device is used for maintaining the grassland in a mode of cutting or folding the grass at a fixed height, and in the process, obvious characteristic sounds are generated among the grass cutting device, the grass and can be used for recognizing and outputting the characteristic sound as flourishing grassland. The grass sound recognition module is set as a neural network and forms a software module capable of recognizing the state of the grassland after training.

-a motion control module for performing a motion planning based on grass conditions, said motion control module being connected to said drive wheel, said motion control module driving said motion platform into a slow mode when said grass recognition module output is active; when the output of the grass sound identification module is invalid, the motion control module drives the motion platform to enter a fast mode.

The grass voice recognition module is used as a negative feedback link and forms a feedback control system with the motion control module, so that the working efficiency of the motion platform can be effectively improved.

Claims (1)

1. The grassland state identification method based on audio recognition comprises a motion platform capable of moving autonomously, wherein a driving wheel and a supporting wheel are arranged at the bottom of the motion platform, so that autonomous motion can be realized, and the grassland state identification method is characterized in that: the sport platform bottom set up grass cutting device and adopt audio signal's adapter, the sport platform inside set up the treater that carries out centralized control, with the adapter connect, the treater can gather the audio signal of adapter output, the treater inside set up meadow state identification method, confirm the meadow state in sport platform's the region, meadow state identification method include:

-a first filter module capable of identifying and filtering out ambient noise;

-a second filter module able to identify and filter the noise of said driving wheel;

-a third filter module capable of identifying and filtering noise of said mowing device;

-a grass recognition module capable of recognizing the sound produced by the grass cutting device when interacting with the grass stems and grass blades;

-a motion control module for performing a motion planning based on grass conditions, said motion control module being connected to said drive wheel, said motion control module driving said motion platform into a slow mode when said grass recognition module output is active; when the output of the grass sound identification module is invalid, the motion control module drives the motion platform to enter a fast mode.

Images