Method for dynamic adjustment of work path of mobile robot and mobile robot

Abstract

The application discloses a kind of mobile robot's operation path dynamic adjustment method and mobile robot, its method includes: based on the acceleration sensor set on mobile robot gathers acceleration vibration signal in running process;Acceleration vibration signal is preprocessed to generate wavelet packet energy spectrum entropy vector;The wavelet packet energy spectrum entropy vector is as the input quantity of neural network, obtains the motion trend information of mobile robot next time;Based on the motion trend information adjustment the motion trajectory of mobile robot.The application is based on neural network model to predict the motion trend of mobile robot by collecting acceleration vibration signal to realize the adjustment of the motion trajectory of mobile robot, to improve the operation efficiency of mobile robot.

Description

Technical Field

[0001] This invention relates to the field of mobile robot technology, and in particular to a method for dynamically adjusting the working path of a mobile robot and the mobile robot itself. Background Technology

[0002] As mobile robot technology matures, it is increasingly being applied to various industrial production processes or hazardous environments to replace manual labor, thereby reducing operating costs. Current technologies offer high efficiency for tasks with fixed locations or fixed movement trends. However, mobile robots are subject to various road condition disturbances during movement, requiring path adjustments. While existing technologies can collect vibration data in real-time using energy quotient spectrometry, this data reflects changes in road conditions. However, combining this with machine learning cannot accurately predict the work path, resulting in low operational efficiency. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a method for dynamically adjusting the working path of a mobile robot and a mobile robot. It adjusts the motion trajectory of the mobile robot by collecting acceleration vibration signals and predicting the motion trend of the mobile robot based on a neural network model, thereby improving the working efficiency of the mobile robot.

[0004] To address the above problems, this invention proposes a method for dynamically adjusting the work path of a mobile robot, the method comprising:

[0005] Acceleration vibration signals during operation are collected using accelerometers installed on the mobile robot.

[0006] Preprocessing of the acceleration vibration signal generates a wavelet packet energy spectrum entropy vector;

[0007] The wavelet packet energy spectrum entropy vector is used as the input of the neural network to obtain the motion trend information of the mobile robot at the next moment.

[0008] The movement trajectory of the mobile robot is adjusted based on the motion trend information.

[0009] The acceleration vibration signals collected during operation by the acceleration sensor installed on the mobile robot include:

[0010] Simulated acceleration vibration signals are collected using accelerometers on mobile robots;

[0011] The simulated acceleration vibration signal is filtered to remove high-frequency vibration signals;

[0012] The analog acceleration vibration signal after filtering out high-frequency vibration signals is processed by analog-to-digital conversion to generate a digital acceleration vibration signal.

[0013] The preprocessing of the acceleration vibration signal to generate the wavelet packet energy spectrum entropy vector includes:

[0014] The acceleration vibration signal is decomposed into wavelet packets using the wavelet packet decomposition principle.

[0015] The wavelet packet coefficients are reconstructed, and the proportion of energy contained in each frequency band component to the total signal energy is calculated.

[0016] The energy distribution of each wavelet packet coefficient of the acceleration vibration signal is obtained by calculating the energy spectral entropy of each frequency band, and the wavelet packet energy spectral entropy is used as a characteristic parameter of the acceleration vibration signal.

[0017] The step of using the wavelet packet energy spectrum entropy vector as input to the neural network to obtain the motion trend information of the mobile robot at the next moment includes:

[0018] Obtain sample entropy data for mobile robot tasks used for model training;

[0019] A portion of the sample entropy data is input into the constructed neural network for predicting motion trends, and another portion of the sample entropy data is used to perform verification processing to train the motion trend model of the mobile robot.

[0020] The wavelet packet energy spectrum entropy vector is input into the motion trend model of the mobile robot to obtain the motion trend information at the next moment.

[0021] The process of adjusting the motion trajectory of the mobile robot based on the motion trend information includes:

[0022] The target location information, mobile robot location information, and motion trend information are marked on the work map within the work area to form a marked map;

[0023] Based on the target location information and motion trend information, the position of the dynamic operation target at the next moment is predicted to obtain the predicted position of the dynamic operation target at the next moment.

[0024] The mobile robot's trajectory is adjusted based on its position information, motion trend information, and the predicted position of the dynamic task target at the next moment.

[0025] The step of marking the target location information, mobile robot location information, and motion trend information on a work map within the work area to form a marked map includes:

[0026] A coordinate system is constructed on the work map within the work area, and the target position information and mobile robot position information are converted into target position coordinates and mobile robot position coordinates based on the coordinate system established on the work map within the work area.

[0027] The target location coordinates, the mobile robot location coordinates, the target movement trend, and the movement trend information are marked on the work map within the work area to form a marked map.

[0028] The process of marking the target location coordinates, the mobile robot location coordinates, the target movement trend, and the movement trend information on the work map within the work area includes:

[0029] The target location coordinates and the mobile robot location coordinates are marked on the work map within the work area to form an initial marked map;

[0030] The target movement trend and the movement trend information are converted into digital information and then marked on the initial marked map.

[0031] The process of predicting the position of the dynamic operation target at the next moment based on the target position information and movement trend information includes:

[0032] Obtain the target motion speed information from the motion trend information, and calculate the distance of the dynamic operation target motion at the next moment based on the target motion speed information;

[0033] Based on the target location information, the distance the dynamic operation target moves at the next moment, and the target movement direction information in the movement trend information, the position of the dynamic operation target at the next moment is predicted.

[0034] The step of predicting and adjusting the mobile robot's trajectory based on its position information, motion trend information, and dynamic task target at the next moment includes:

[0035] Based on the mobile robot's position information and the next moment prediction position of the dynamic task target, the robot's motion trend information, including the robot's motion direction information and robot motion speed information, is adjusted.

[0036] The movement trajectory of the mobile robot is adjusted based on the adjusted robot movement direction information and robot movement speed information.

[0037] Accordingly, the present invention also proposes a mobile robot, wherein the mobile robot is provided with a memory and a processor, the memory is used to store a computer program, and the processor is used to be coupled to the memory to execute the computer program in order to implement the method described above.

[0038] The method and system involved in this invention improve the operational efficiency of mobile robots by extracting feature quantities from acceleration vibration signals on mobile robots and then combining them with neural networks to predict the movement trends of the mobile robots. The method employing a combination of wavelet packet decomposition and energy spectral entropy directly reveals the feature information of the acceleration vibration signals on the mobile robot, facilitating feature extraction. This allows the neural network to learn and train on the acceleration vibration signals during the robot's operation, thereby predicting the robot's trajectory. This approach achieves high accuracy in identifying acceleration vibration signals and uses a neural network model to classify and predict these signals, thus improving the accuracy of trajectory prediction and resulting in high reliability and precision in the operation of the mobile robot. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of the structure of the mobile robot in an embodiment of the present invention;

[0041] Figure 2 This is a flowchart of the method for dynamically adjusting the working path of a mobile robot in an embodiment of the present invention;

[0042] Figure 3 This is a flowchart of a method for adjusting the motion trajectory of the mobile robot in an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Specifically, Figure 1 A schematic diagram of the structure of a mobile robot according to an embodiment of the present invention is shown. The mobile robot includes:

[0045] The data acquisition module is used to collect acceleration vibration signals during operation based on the acceleration sensors installed on the mobile robot;

[0046] The preprocessing module is used to preprocess the acceleration vibration signal to generate a wavelet packet energy spectrum entropy vector;

[0047] The neural network processing module is used to use the wavelet packet energy spectrum entropy vector as the input of the neural network to obtain the motion trend information of the mobile robot at the next moment.

[0048] The path adjustment module is used to adjust the motion trajectory of the mobile robot based on the motion trend information.

[0049] The mobile robot in this embodiment of the invention can improve its operational efficiency by extracting feature quantities from the acceleration vibration signals on the robot and combining them with a neural network to predict the robot's motion trend. The combination of wavelet packet decomposition and energy spectral entropy directly reveals the feature information of the acceleration vibration signals on the robot, facilitating feature extraction. This allows the neural network to learn and train on the acceleration vibration signals during the robot's operation, thereby predicting the robot's trajectory. This method achieves high accuracy in identifying acceleration vibration signals and uses a neural network model to classify and predict them, improving trajectory prediction accuracy and resulting in high reliability and precision in the mobile robot's operation.

[0050] Specifically, Figure 2 A flowchart illustrating a method for dynamically adjusting the work path of a mobile robot according to an embodiment of the present invention is shown. The method includes:

[0051] S201. Acceleration vibration signals during operation are collected based on the acceleration sensor installed on the mobile robot;

[0052] Specifically, the process of collecting acceleration vibration signals during operation based on the acceleration sensor installed on the mobile robot includes: collecting simulated acceleration vibration signals based on the acceleration sensor on the mobile robot; filtering the simulated acceleration vibration signals to remove high-frequency vibration signals; and performing analog-to-digital conversion on the simulated acceleration vibration signals after removing high-frequency vibration signals to generate digital acceleration vibration signals.

[0053] The mobile robot here is equipped with an accelerometer. The accelerometer can collect acceleration vibration signals at its location during the robot's movement, and then use these acceleration vibration signals to complete the final path adjustment process.

[0054] In this embodiment of the invention, a MEMS accelerometer can be selected to collect signals. The MEMS accelerometer has the advantages of small size, low price, strong output signal, and simple subsequent circuitry.

[0055] In this embodiment of the invention, an anti-aliasing filter circuit can be used for filtering, which can effectively filter out high-frequency vibration signals in the simulated acceleration vibration signal.

[0056] In this embodiment of the invention, the data is converted into digital quantities using an analog-to-digital converter (ADC). The sampling frequency of the ADC is set to above 5kHz. The sampled data is then output to the corresponding processor via a serial port according to a predefined format to extract the acceleration vibration signal.

[0057] S202. Preprocess the acceleration vibration signal to generate a wavelet packet energy spectrum entropy vector;

[0058] The preprocessing of the acceleration vibration signal to generate wavelet packet energy spectrum entropy values ​​includes: decomposing the acceleration vibration signal into wavelet packets using the wavelet packet decomposition principle; reconstructing the wavelet packet coefficients and calculating the proportion of energy contained in each frequency band component to the total signal energy; obtaining the energy distribution of each wavelet packet coefficient of the acceleration vibration signal by calculating the energy spectrum entropy of each frequency band, and using the wavelet packet energy spectrum entropy as a characteristic parameter of the acceleration vibration signal.

[0059] The reconstruction of wavelet packet coefficients and the determination of the proportion of energy contained in each frequency band component to the total signal energy include: obtaining different frequency band components of the signal through wavelet packet decomposition, obtaining the characteristic information of each frequency band based on the distribution of the acceleration vibration signal frequency bands, and analyzing the acceleration vibration signal using the wavelet packet energy spectrum.

[0060] It should be noted that wavelet packet decomposition of acceleration vibration signals is commonly used in the identification and judgment of fault vibrations under mechanical principles. The collected acceleration vibration signal forms a random variable with several probability distribution values. Each random variable has an overall feature entropy value. The larger the probability distribution value of the random variable, the larger its corresponding overall feature entropy value; the smaller the probability distribution value of the random variable, the smaller its corresponding overall feature entropy value.

[0061] For a random time series under random variables, there is a relationship between the energy of each frequency band under the random time series, and the energy spectral entropy under the random time series can reflect the distribution of energy of the time series signal in the frequency domain.

[0062] Wavelet packet decomposition can decompose the decomposition space into the sum of different sub-decomposition spaces according to different scales. If a higher resolution is required, each sub-decomposition space can be further decomposed, which unifies the multi-resolution subspace and each sub-decomposition space through the new space.

[0063] Here, wavelet packet decomposition is used to obtain different frequency band components of the signal. Based on the distribution of the acceleration vibration signal's frequency bands, the characteristic information of each frequency band is obtained. The wavelet packet energy spectrum is then used to analyze different acceleration vibration signals. Wavelet packet transform has good processing capabilities for non-stationary signals, and its application to the analysis of acceleration vibration signals yields good results. Wavelet packet decomposition can obtain different frequency band components of the signal, and the information distribution of these components differs. Based on the distribution of the acceleration vibration signal, the characteristic information of each frequency band can be obtained.

[0064] The energy of each frequency band component of the wavelet packet of an acceleration vibration signal can be calculated using Passerwale's theorem. The wavelet packet coefficients are used to analyze the energy of different frequency bands of the acceleration vibration signal. The energy corresponding to the wavelet packet coefficients is different for different acceleration vibration signals. Here, the wavelet packet energy spectrum can be used to analyze different acceleration vibration signals.

[0065] Energy spectral entropy is a quantitative description of the complexity of the energy distribution of an accelerated vibration signal in the frequency domain. It reflects the amount of information across the entire frequency range of the accelerated vibration signal, characterizing the information entropy at all frequencies without considering the detailed components of the spectrum. Here, wavelet packet transform is used to process the accelerated vibration signal. It has a wideband response for non-stationary signals, high frequency resolution at low frequencies, and high time resolution at high frequencies, making it suitable for analyzing non-stationary signals.

[0066] S203. Use the wavelet packet energy spectrum entropy vector as the input of the neural network to obtain the motion trend information of the mobile robot at the next moment.

[0067] Specifically, using the wavelet packet energy spectrum entropy vector as input to the neural network to obtain the motion trend information of the mobile robot at the next moment includes: acquiring sample entropy data of the mobile robot's operation for model training; inputting a portion of the sample entropy data into the constructed neural network for predicting the motion trend, and performing verification processing through another portion of the sample entropy data to train the motion trend model of the mobile robot; and inputting the wavelet packet energy spectrum entropy vector into the motion trend model of the mobile robot to obtain the motion trend information at the next moment.

[0068] Here, acceleration vibration data generated under different mobile robot motion trend adjustment environments can be used as training samples. The wavelet packet energy spectrum entropy vector of the acceleration vibration data generated under the motion trend adjustment environment is the sample entropy data of the mobile robot operation.

[0069] In the specific implementation process, a deep neural network for predicting motion trends is constructed, and the number of input layer neurons is selected according to the number of available feature columns; a certain preset percentage of sample entropy data is randomly selected from all sample entropy data as a training sample set to train the deep neural network, and the remaining sample entropy data is used as a validation set; the deep neural network is trained iteratively to obtain the neural network model of the trained mobile robot.

[0070] S204. Adjust the motion trajectory of the mobile robot based on the motion trend information.

[0071] This invention, in its embodiments, extracts feature quantities from the acceleration vibration signals of a mobile robot and then combines them with a neural network to predict the robot's motion trend. After obtaining the motion trend information, the mobile robot can then adjust its trajectory. Specifically... Figure 3 A flowchart illustrating a method for adjusting the motion trajectory of a mobile robot according to an embodiment of the present invention is shown, including:

[0072] S301. Mark the target location information, mobile robot location information, and motion trend information on the work map within the work area to form a marked map;

[0073] It should be noted that marking the target location information, mobile robot location information, and motion trend information on the work map within the work area to form a marked map includes: constructing a coordinate system on the work map within the work area, and converting the target location information and mobile robot location information into target location coordinates and mobile robot location coordinates based on the coordinate system established on the work map within the work area; and marking the target location coordinates, mobile robot location coordinates, target motion trend, and motion trend information on the work map within the work area to form a marked map.

[0074] It should be noted that the process of marking the target position coordinates, the mobile robot position coordinates, the target motion trend, and the motion trend information on the work map within the work area includes: marking the target position coordinates and the mobile robot position coordinates on the work map within the work area to form an initial marked map; converting the target motion trend and the motion trend information into digital information and marking them on the initial marked map.

[0075] Here, the target position coordinates and robot position coordinates are marked on the work map within the work area to form an initial marked map; then, the target motion trend and robot motion trend information are converted into digital information and marked on the initial marked map.

[0076] S302. Based on the target location information and motion trend information, predict the position of the dynamic operation target at the next moment to obtain the predicted position of the dynamic operation target at the next moment.

[0077] It should be noted that the process of predicting the position of the dynamic operation target at the next moment based on the target position information and the motion trend information includes: obtaining the target motion speed information in the motion trend information, and calculating the distance of the dynamic operation target at the next moment based on the target motion speed information; and predicting the position of the dynamic operation target at the next moment based on the target position information, the distance of the dynamic operation target at the next moment, and the target motion direction information in the motion trend information.

[0078] Here, it is necessary to obtain the target speed information in the target movement trend, and then calculate the distance of the dynamic operation target movement at the next moment based on the target speed information; and predict the position of the dynamic operation target at the next moment based on the target position information, the distance of the dynamic operation target movement at the next moment and the target movement direction information in the target movement trend information, so as to obtain the predicted position of the dynamic operation target at the next moment.

[0079] S303. Based on the mobile robot's position information, motion trend information, and the predicted position of the dynamic task target at the next moment, adjust the mobile robot's motion trajectory.

[0080] It should be noted that adjusting the mobile robot's trajectory based on the predicted position of the next moment of the mobile robot's position information, motion trend information, and dynamic task target includes: adjusting the robot's motion direction information and robot's motion speed information in the robot's motion trend information based on the mobile robot's position information and the predicted position of the next moment of the dynamic task target; and adjusting the mobile robot's trajectory based on the adjusted robot's motion direction information and robot's motion speed information. The method involved in this embodiment of the invention can improve the mobile robot's work efficiency by extracting the feature quantities of the acceleration vibration signal on the mobile robot and then combining it with a neural network to predict the mobile robot's motion trend. Here, the combination of wavelet packet decomposition and energy spectral entropy directly reflects the feature information of the acceleration vibration signal on the mobile robot, which is beneficial for feature information extraction. This allows the neural network to learn and train the acceleration vibration signal during the mobile robot's operation, thereby predicting the mobile robot's trajectory. This method has a high accuracy rate in recognizing acceleration vibration signals and can achieve classification and prediction of acceleration vibration signals through a neural network model, thus improving the accuracy of trajectory prediction and making the mobile robot's work efficiency highly reliable and accurate.

[0081] This invention also provides a mobile robot, which includes a memory and a processor. The memory stores a computer program, and the processor is coupled to the memory to execute the computer program to implement the methods described above.

[0082] This invention also provides a computer-readable storage medium, wherein when the computer program is executed by a processor, the processor causes the processor to perform the steps in the above-described method.

[0083] The embodiments of the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for dynamically adjusting the working path of a mobile robot, characterized in that, The method includes: Acceleration vibration signals are collected by the accelerometer installed on the mobile robot during operation. The collected acceleration vibration signals form a random variable, which has several probability distribution values. Each random variable has an overall characteristic entropy value. Preprocessing of the acceleration vibration signal generates a wavelet packet energy spectrum entropy vector; The wavelet packet energy spectrum entropy vector is used as the input of the neural network to obtain the motion trend information of the mobile robot at the next moment. The movement trajectory of the mobile robot is adjusted based on the motion trend information; The step of using the wavelet packet energy spectrum entropy vector as input to the neural network to obtain the motion trend information of the mobile robot at the next moment includes: Obtain sample entropy data for mobile robot tasks used for model training; A portion of the sample entropy data is input into the constructed neural network for predicting motion trends, and another portion of the sample entropy data is used to perform verification processing to train the motion trend model of the mobile robot. The wavelet packet energy spectrum entropy vector is input into the motion trend model of the mobile robot to obtain the motion trend information at the next moment. The process of adjusting the motion trajectory of the mobile robot based on the motion trend information includes: The target location information, mobile robot location information, and motion trend information are marked on the work map within the work area to form a marked map; Based on the target location information and motion trend information, the position of the dynamic operation target at the next moment is predicted to obtain the predicted position of the dynamic operation target at the next moment. The mobile robot's trajectory is adjusted based on its position information, motion trend information, and the predicted position of the dynamic task target at the next moment.

2. The method for dynamically adjusting the working path of a mobile robot as described in claim 1, characterized in that, The acceleration vibration signals collected during operation by the acceleration sensor installed on the mobile robot include: Simulated acceleration vibration signals are collected using accelerometers on mobile robots; The simulated acceleration vibration signal is filtered to remove high-frequency vibration signals; The analog acceleration vibration signal after filtering out high-frequency vibration signals is processed by analog-to-digital conversion to generate a digital acceleration vibration signal.

3. The method for dynamically adjusting the working path of a mobile robot as described in claim 1, characterized in that, The preprocessing of the acceleration vibration signal to generate the wavelet packet energy spectrum entropy vector includes: The acceleration vibration signal is decomposed into wavelet packets using the wavelet packet decomposition principle. The wavelet packet coefficients are reconstructed, and the proportion of energy contained in each frequency band component to the total signal energy is calculated. The energy distribution of each wavelet packet coefficient of the acceleration vibration signal is obtained by calculating the energy spectral entropy of each frequency band, and the wavelet packet energy spectral entropy is used as a characteristic parameter of the acceleration vibration signal.

4. The method for dynamically adjusting the working path of a mobile robot as described in any one of claims 1-3, characterized in that, The step of marking the target location information, mobile robot location information, and motion trend information on a work map within the work area to form a marked map includes: A coordinate system is constructed on the work map within the work area, and the target position information and mobile robot position information are converted into target position coordinates and mobile robot position coordinates based on the coordinate system established on the work map within the work area. The target location coordinates, the mobile robot location coordinates, and the motion trend information are marked on the work map within the work area to form a marked map.

5. The method for dynamically adjusting the working path of a mobile robot as described in claim 4, characterized in that, The process of marking the target location coordinates, the mobile robot location coordinates, and the motion trend information on the work map within the work area includes: The target location coordinates and the mobile robot location coordinates are marked on the work map within the work area to form an initial marked map; The motion trend information is converted into digital information and then marked on the initial marked map.

6. The method for dynamically adjusting the working path of a mobile robot as described in claim 4, characterized in that, The process of predicting the position of the dynamic operation target at the next moment based on the target position information and movement trend information includes: Obtain the target motion speed information from the motion trend information, and calculate the distance of the dynamic operation target motion at the next moment based on the target motion speed information; Based on the target location information, the distance the dynamic operation target moves at the next moment, and the target movement direction information in the movement trend information, the position of the dynamic operation target at the next moment is predicted.

7. The method for dynamically adjusting the working path of a mobile robot as described in claim 4, characterized in that, The step of predicting the next moment's position and adjusting the mobile robot's trajectory based on the mobile robot's position information, motion trend information, and dynamic task target includes: Based on the mobile robot's position information and the next moment prediction position of the dynamic task target, the robot's motion trend information, including the robot's motion direction information and the robot target's motion speed information, is adjusted. The movement trajectory of the mobile robot is adjusted based on the adjusted robot movement direction information and the robot target movement speed information.

8. A mobile robot, characterized in that, The mobile robot is provided with a memory and a processor. The memory is used to store computer programs, and the processor is used to couple with the memory to execute the computer programs in order to implement the method according to any one of claims 1 to 7.

Images