CN115562299A - Navigation method and device of mobile robot, mobile robot and medium - Google Patents
Navigation method and device of mobile robot, mobile robot and medium Download PDFInfo
- Publication number
- CN115562299A CN115562299A CN202211326593.1A CN202211326593A CN115562299A CN 115562299 A CN115562299 A CN 115562299A CN 202211326593 A CN202211326593 A CN 202211326593A CN 115562299 A CN115562299 A CN 115562299A
- Authority
- CN
- China
- Prior art keywords
- mobile robot
- controller parameter
- parameter
- controller
- long
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 230000015654 memory Effects 0.000 claims description 47
- 230000006870 function Effects 0.000 claims description 24
- 239000013598 vector Substances 0.000 claims description 23
- 238000004590 computer program Methods 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 13
- 230000007787 long-term memory Effects 0.000 claims description 8
- 238000004891 communication Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 230000006403 short-term memory Effects 0.000 description 6
- 238000012937 correction Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 229920002595 Dielectric elastomer Polymers 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 229910001285 shape-memory alloy Inorganic materials 0.000 description 2
- 229920000431 shape-memory polymer Polymers 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 229920000831 ionic polymer Polymers 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000002905 metal composite material Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 239000002520 smart material Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
- G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0276—Control of position or course in two dimensions specially adapted to land vehicles using signals provided by a source external to the vehicle
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Electromagnetism (AREA)
- Manipulator (AREA)
Abstract
The invention discloses a navigation method and a navigation device for a mobile robot, the mobile robot and a medium. The method comprises the following steps: after a mobile robot executes current moving operation, inputting current sensor data of the mobile robot into a pre-trained target network model to obtain output first controller parameters; determining a second controller parameter based on the current actual track parameter and the current preset track parameter of the mobile robot; controlling the mobile robot to perform a next movement operation based on the first controller parameter and the second controller parameter. The embodiment of the invention solves the problem that the existing navigation method of the mobile robot depends on a single deviation rectifying algorithm, and improves the navigation accuracy of the mobile robot.
Description
Technical Field
The present invention relates to the field of robot technology, and in particular, to a method and an apparatus for navigating a mobile robot, and a medium.
Background
With the rapid development of science and technology, a robot system with a moving function, which is composed of a sensor, an automatic controller and other mechanisms, becomes one of the popular research directions, and a mobile robot is widely applied to the fields of production industry, building industry, monitoring industry and the like.
In actual scene application, a mobile robot generally needs to perform navigation movement according to a preset track, but under the influence of performance parameters of a body of the mobile robot and an external environment, a track error usually exists between the actual track of the mobile robot and the preset track, so that the mobile robot cannot accurately reach a target position to execute task operation.
The navigation method of the existing mobile robot mainly depends on a single deviation correction algorithm, so that the navigation accuracy of the mobile robot is not high.
Disclosure of Invention
The embodiment of the invention provides a navigation method and device of a mobile robot, the mobile robot and a medium, which aim to solve the problem that the existing navigation method of the mobile robot depends on a single deviation correction algorithm and improve the navigation accuracy of the mobile robot.
According to an embodiment of the present invention, there is provided a navigation method of a mobile robot, the method including:
after a mobile robot executes current moving operation, inputting current sensor data of the mobile robot into a pre-trained target network model to obtain output first controller parameters;
determining a second controller parameter based on the current actual track parameter and the current preset track parameter of the mobile robot;
controlling the mobile robot to perform a next movement operation based on the first controller parameter and the second controller parameter.
According to another embodiment of the present invention, there is provided a navigation apparatus of a mobile robot, the apparatus including:
the first controller parameter determining module is used for inputting the current sensor data of the mobile robot into a pre-trained target network model after the mobile robot executes the current moving operation to obtain an output first controller parameter;
the second controller parameter determining module is used for determining second controller parameters based on the current actual track parameters and the current preset track parameters of the mobile robot;
and the mobile robot control module is used for controlling the mobile robot to execute the next moving operation based on the first controller parameter and the second controller parameter.
According to another embodiment of the present invention, there is provided a mobile robot including: a drive controller and at least one sensor;
wherein the sensor is used for collecting sensor data;
the drive controller comprises at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to enable the at least one processor to perform the method of navigating a mobile robot according to any of the embodiments of the present invention.
According to another embodiment of the present invention, there is provided a computer-readable storage medium storing computer instructions for causing a processor to implement a navigation method of a mobile robot according to any one of the embodiments of the present invention when the computer instructions are executed.
According to the technical scheme of the embodiment of the invention, after the mobile robot executes the current movement operation, the current sensor data of the mobile robot is input into the pre-trained target network model to obtain the output first controller parameter, the second controller parameter is determined based on the current actual track parameter and the current preset track parameter of the mobile robot, the mobile robot is controlled to execute the next movement operation based on the first controller parameter and the second controller parameter, and the controller parameters are respectively obtained based on two dimensions, so that the problem that the existing navigation method of the mobile robot depends on a single deviation correction algorithm is solved, and the navigation accuracy of the mobile robot is improved.
It should be understood that the statements in this section are not intended to identify key or critical features of the embodiments of the present invention, nor are they intended to limit the scope of the invention. Other features of the present invention will become apparent from the following description.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a flowchart of a navigation method of a mobile robot according to an embodiment of the present invention;
fig. 2 is a flowchart of another navigation method for a mobile robot according to an embodiment of the present invention;
FIG. 3 is an architecture diagram of a target network model according to an embodiment of the present invention;
fig. 4 is a flowchart illustrating an exemplary navigation method of a mobile robot according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a navigation device of a mobile robot according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a mobile robot according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a driving controller according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Moreover, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Fig. 1 is a flowchart of a navigation method of a mobile robot according to an embodiment of the present invention, where the navigation method is applicable to a situation of performing navigation tracking on a mobile robot, and the navigation method may be executed by a navigation device of the mobile robot, where the navigation device of the mobile robot may be implemented in a form of hardware and/or software, and the navigation device of the mobile robot may be configured in the mobile robot. As shown in fig. 1, the method includes:
and S110, after the mobile robot executes the current moving operation, inputting the current sensor data of the mobile robot into a pre-trained target network model to obtain the output first controller parameters.
The mobile robot in this embodiment may be a soft robot or a mechanical robot, and is particularly suitable for a soft robot. The soft robot is made of soft intelligent material, different from the mechanical robot, and the driving mode of the soft robot mainly depends on the intelligent material, such as the intelligent material generally being Dielectric Elastomer (DE), ionic polymer metal composite material (IPMC), shape Memory Alloy (SMA), shape Memory Polymer (SMP), etc. The smart material used in the soft robot is not limited herein.
Because a mechanical robot can build a mechanical or kinematic model by relying on a traditional mechanical algorithm, most of the prior art directly calculates controller parameters of the mechanical robot based on sensor data, but the driving characteristics of a software robot are not suitable for the traditional mechanical algorithm, and the sensor data and the controller parameters do not have a curve rule which can be fitted.
By way of example, the current sensor data includes, but is not limited to, current velocities of three axes, current accelerations of three axes, current angular velocities of three axes, current angles of three axes, current temperatures, current wind speeds, and the like, wherein the three axes include an x-axis, a y-axis, and a z-axis. The current sensor data collected by the mobile robot is not limited, and a user can set a required sensor on the mobile robot according to actual requirements.
Exemplary model architectures of the target Network model include, but are not limited to, CNN (Convolutional Neural Networks), FCN (full Convolutional Neural Networks), residual error Networks (resnets), DNN (Deep Neural Networks), RNN (cyclic Neural Networks), multi-channel LSTM (Long Short-Term Memory), transform Networks, and so on, and the model architecture of the target Network model is not limited herein.
The first controller parameter includes, but is not limited to, an amplitude of the current, a frequency of the current, a duty cycle of the current, a duration of the current, an amplitude of the voltage, a frequency of the voltage, a duty cycle of the voltage, a duration of the voltage, and the like, and is not limited herein, and the first controller parameter can be customized by a user according to actual needs.
And S120, determining a second controller parameter based on the current actual track parameter and the current preset track parameter of the mobile robot.
Specifically, the current actual trajectory parameter is used to characterize an actual trajectory parameter of the mobile robot after performing the current movement operation, and the actual trajectory parameter includes, but is not limited to, an actual position coordinate, an actual heading angle, and the like. The actual trajectory parameters are not limited herein.
Specifically, the current preset trajectory parameter is used to represent a preset trajectory parameter of the mobile robot after performing the current movement operation, and for example, the preset trajectory parameter includes, but is not limited to, a preset position coordinate, a preset heading angle, and the like. The preset trajectory parameters are not limited herein.
In an optional embodiment, the determining the second controller parameter based on the current actual trajectory parameter and the current preset trajectory parameter of the mobile robot comprises: and determining a second controller parameter by adopting a preset closed-loop control algorithm based on the current actual track parameter and the current preset track parameter of the mobile robot.
Exemplary preset closed-loop control algorithms include, but are not limited to, a bit-based control algorithm, a PID control algorithm, a robust control algorithm, a green algorithm, and an LQR (Linear Quadratic Regulator) algorithm, among others. The predetermined closed-loop-control algorithm is not limited herein.
On the basis of the foregoing embodiment, optionally, the method further includes: obtaining current actual trajectory parameters acquired by the vision equipment, and/or determining the current actual trajectory parameters based on current sensor data.
In an optional embodiment, a visual device is installed on the mobile robot or in the scene environment, and is used for determining the current actual trajectory parameters based on the acquired trajectory images.
In another alternative embodiment, the current actual trajectory parameter is determined based on, in particular, the last actual trajectory parameter and the current sensor data.
In another optional embodiment, the current actual trajectory parameters acquired by the vision device and the mean of the current actual trajectory parameters determined based on the current sensor data are used as the current actual trajectory parameters of the mobile robot.
And S130, controlling the mobile robot to execute the next moving operation based on the first controller parameter and the second controller parameter.
In an optional embodiment, the average value of the first controller parameter and the second controller parameter is used as a target controller parameter, and the mobile robot is controlled to perform the next moving operation based on the target controller parameter. In particular, the target controller parameters are used to characterize parametric data input to a drive controller on the mobile robot, which includes, for example and without limitation, a controlled power supply and a PWM module.
In this case, for example, the voltage amplitude in the first controller parameter is 5V, the voltage amplitude in the second controller parameter is 7V, and the voltage amplitude in the target controller parameter is 6V.
According to the technical scheme of the embodiment, after the mobile robot executes the current moving operation, the current sensor data of the mobile robot is input into a pre-trained target network model to obtain the output first controller parameter, the second controller parameter is determined based on the current actual track parameter and the current preset track parameter of the mobile robot, the mobile robot is controlled to execute the next moving operation based on the first controller parameter and the second controller parameter, the controller parameters are respectively obtained based on two dimensions, the problem that the existing navigation method of the mobile robot depends on a single deviation correction algorithm is solved, and the navigation accuracy of the mobile robot is improved.
Fig. 2 is a flowchart of another navigation method for a mobile robot according to an embodiment of the present invention, and the embodiment further refines the target network model in the above embodiment. As shown in fig. 2, the method includes:
s210, inputting the training sensor data in the training sensor data set into an initial network model to obtain at least one output predictive controller parameter.
For example, the training sensor data set may be sensor data collected by a sensor during the movement of the mobile robot.
And S220, determining a loss function based on each predictive controller parameter and the standard controller parameter corresponding to each predictive controller parameter.
Here, the standard controller parameter may be, for example, parameter data input into the driving controller during the movement of the mobile robot.
In an optional embodiment, the standard controller parameter corresponding to the current training sensor data acquired at the current time is a standard controller parameter acquired at the next time.
Exemplary loss functions include, but are not limited to, a square loss function, a logarithmic loss function, an exponential loss function, a logistic regression loss function, a Huber loss function, a cross-entropy loss function, and a Kullback-Leibler divergence loss function, among others. The loss function is not limited herein.
And S230, based on the loss function, adjusting model parameters of the initial network model until the loss function is converged, and obtaining a trained target network model.
On the basis of the foregoing embodiment, optionally, the method further includes: inputting each verification sensor in the verification sensor data set into a target network model to obtain at least one output prediction controller parameter, and determining performance parameters of the target network model based on each prediction controller parameter and a standard controller parameter corresponding to each prediction controller parameter; and under the condition that the performance parameters do not meet the preset parameter standard, adjusting the super parameters of the target network model, and continuing to train the adjusted target network model.
Exemplary hyper-parameters include, but are not limited to, the number of network layers, the number of network nodes, the number of iterations, and the learning rate, among others.
The advantage of this arrangement is that the generalization ability of the trained target network model can be improved.
In this embodiment, the target network model is a multi-channel long-and-short-term memory model. In an optional embodiment, the target network model includes an output module and at least one long-short time memory module, where the first long-short time memory module is configured to output a first long-short time feature vector based on input current sensor data, the ith long-short time memory module is configured to output an ith long-short time feature vector based on the input current sensor data and an i-1 th long-short time feature vector output by the i-1 th long-short time memory module, and the output module is configured to output a first controller parameter based on an nth long-short time feature vector output by the nth long-short time memory module; and i is an integer which is greater than 1 and less than n, and n represents the number of long-time memory modules in the target network model.
In an optional embodiment, the ith long-short-term memory module includes an input unit and a long-short-term memory unit, the input unit includes a first linear layer, a self-attention layer and a second linear layer, the first linear layer is used for performing linear coding on input current sensor data and outputting first sensor characteristics, the self-attention layer is used for outputting fused sensor data based on the first sensor characteristics output by the first linear layer, the second linear layer is used for performing linear coding on the fused sensor data output by the attention layer and outputting second sensor characteristics, and the long-short-term memory unit is used for outputting an ith long-short-term feature vector based on the second sensor characteristics output by the second linear layer and an i-1 th long-short-term feature vector output by the i-1 th long-term memory module.
Wherein, the self-attention layer is used for realizing multi-channel data fusion. Illustratively, the self-attention layer satisfies the formula:
wherein Q represents a query vector matrix, K represents a key vector matrix, V represents a value vector matrix, d k Representing a vector for a query.
Specifically, the long and short term memory unit includes a memory unit c, an input gate i, a forgetting gate f and an output gate o. The updating process of the long-time memory unit and the short-time memory unit at the time t meets the following formula:
f t ＝σ(W f ·[h t-1 ,x t ]+b f )；i t ＝σ(W i ·[h t-1 ,x t ]+b i )；
o t ＝σ(W o ·[h t-1 ,x t ]+b o )；h t ＝o t ⊙tanh(c t )
wherein x is t Representing input data of the long-and-short-term memory cell at time t, f t 、i t 、c t And o t Respectively representing the values of the forgetting gate, the input gate, the memory unit and the output gate at the time t,
represents a candidate memory state value, h, of the memory cell t Long and short term feature vectors representing the outputs of the long and short term memory cells, sigma represents a logistic sigmoid function, W f 、W i 、W c 、W o 、b f 、b i 、b c And b o A weight matrix is represented.
Fig. 3 is an architecture diagram of a target network model according to an embodiment of the present invention. Specifically, the target network model comprises an output module and long and short term memory modules corresponding to n moments respectively, wherein the output module comprises a linear layer and an activation layer, each long and short term memory module comprises an input unit and a long and short term memory unit, and the input unit comprises a first linear layer, a self-attention layer and a second linear layer. The first controller parameters output by the target network model include a voltage parameter for voltage modulation, a current parameter for current modulation, and a duty cycle for PWM modulation.
And S240, after the mobile robot executes the current moving operation, inputting the current sensor data of the mobile robot into a pre-trained target network model to obtain the output first controller parameters.
And S250, determining a second controller parameter based on the current actual track parameter and the current preset track parameter of the mobile robot.
And S260, controlling the mobile robot to execute the next moving operation based on the first controller parameter and the second controller parameter.
In an optional embodiment, based on the first controller parameter and the second controller parameter, controlling the mobile robot to perform a next movement operation includes: determining a controller parameter error based on the first controller parameter and the second controller parameter; and determining a target controller parameter based on the first controller parameter and the controller parameter error, and controlling the mobile robot to execute the next moving operation based on the target controller parameter.
Wherein, in particular, the controller parameter error is used to characterize proportional difference data between the first controller parameter and the second controller parameter.
Therein, an exemplary, target controller parameter α * Satisfies the formula:
α * ＝α 1 ±λ|α 1 -α 2 |
wherein alpha is 1 Denotes a first controller parameter, lambda denotes a scaling factor, alpha 2 Representing a second controller parameter.
Specifically, the target controller parameter is equal to the sum of the errors of the first controller parameter and the controller parameter when the first controller parameter is smaller than the second controller parameter, and the target controller parameter is equal to the difference between the errors of the first controller parameter and the controller parameter when the first controller parameter is larger than the second controller parameter.
In another alternative embodiment, the target controller parameter is determined based on the second controller parameter and the controller parameter error, and the mobile robot is controlled to perform the next move operation based on the target controller parameter.
The advantage of this arrangement is that the accuracy of the target controller parameters can be further improved, thereby further improving the navigation accuracy of the mobile robot.
Fig. 4 is a flowchart of an embodiment of a navigation method of a mobile robot according to an embodiment of the present invention, which is exemplarily illustrated by taking the mobile robot as a soft robot, and specifically, the soft robot is provided with a driving controller, a sensor and a vision device.
After the software robot executes the current movement operation and before the software robot reaches the end point, the current actual track data acquired by the visual equipment is sent to the PID algorithm module, so that the PID algorithm module determines and determines a second controller parameter alpha based on the current preset track parameter and the current actual track parameter 2 . Inputting the current sensor data acquired by the sensor into a target network model to obtain an output first controller parameter alpha 1 . Will be based on the first controller parameter alpha 1 And a second controller parameter alpha 2 Determined target controller parameter alpha * The input is input into a driving controller on the soft robot, so that the driving controller controls the soft robot to execute the next moving operation.
According to the technical scheme of the embodiment, each training sensor data in a training sensor data set is input into an initial network model to obtain at least one output prediction controller parameter, a loss function is determined based on each prediction controller parameter and a standard controller parameter corresponding to each prediction controller parameter, model parameters of the initial network model are adjusted based on the loss function until a trained target network model is obtained when the loss function is converged, the training problem of the target network model is solved, the multi-channel LSTM model is selected as the target network model, the adaptability of the target network model and a navigation scene of the mobile robot is guaranteed, and the navigation accuracy of the mobile robot is further improved.
Fig. 5 is a schematic structural diagram of a navigation device of a mobile robot according to an embodiment of the present invention. As shown in fig. 5, the apparatus includes: a first controller parameter determination module 310, a second controller parameter determination module 320, and a mobile robot control module 330.
The first controller parameter determining module 310 is configured to, after the mobile robot executes a current moving operation, input current sensor data of the mobile robot into a pre-trained target network model to obtain an output first controller parameter;
a second controller parameter determining module 320, configured to determine a second controller parameter based on a current actual trajectory parameter and a current preset trajectory parameter of the mobile robot;
and a mobile robot control module 330 for controlling the mobile robot to perform a next moving operation based on the first controller parameter and the second controller parameter.
According to the technical scheme of the embodiment, after the mobile robot executes the current moving operation, the current sensor data of the mobile robot is input into a pre-trained target network model to obtain the output first controller parameter, the second controller parameter is determined based on the current actual track parameter and the current preset track parameter of the mobile robot, the mobile robot is controlled to execute the next moving operation based on the first controller parameter and the second controller parameter, the controller parameters are respectively obtained based on two dimensions, the problem that the existing navigation method of the mobile robot depends on a single deviation correction algorithm is solved, and the navigation accuracy of the mobile robot is improved.
On the basis of the above embodiment, optionally, the apparatus further includes:
the target network model training module is used for inputting the data of each training sensor in the training sensor data set into the initial network model to obtain at least one output predictive controller parameter;
determining a loss function based on each predictive controller parameter and a standard controller parameter corresponding to each predictive controller parameter respectively;
and adjusting model parameters of the initial network model based on the loss function until the loss function is converged, so as to obtain a trained target network model.
On the basis of the foregoing embodiment, optionally, the target network model includes an output module and at least one long-short-term memory module, where the first long-short-term memory module is configured to output a first long-short-term feature vector based on input current sensor data, the ith long-short-term memory module is configured to output an ith long-short-term feature vector based on the input current sensor data and an i-1 th long-short-term feature vector output by the i-1 th long-short-term memory module, and the output module is configured to output a first controller parameter based on an nth long-short-term feature vector output by the nth long-short-term memory module; and i is an integer which is greater than 1 and less than n, and n represents the number of long-time memory modules in the target network model.
On the basis of the foregoing embodiment, optionally, the ith long-short-term memory module includes an input unit and a long-term memory unit, where the input unit includes a first linear layer, a self-attention layer, and a second linear layer, the first linear layer is configured to perform linear coding on input current sensor data and output a first sensor feature, the self-attention layer is configured to output fused sensor data based on the first sensor feature output by the first linear layer, the second linear layer is configured to perform linear coding on the fused sensor data output by the self-attention layer and output a second sensor feature, and the long-short-term memory unit is configured to output an ith long-term feature vector based on a second sensor feature output by the second linear layer and an i-1 th long-short-term feature vector output by the i-1 th long-short-term memory module.
On the basis of the foregoing embodiment, optionally, the mobile robot control module 330 is specifically configured to:
determining a controller parameter error based on the first controller parameter and the second controller parameter;
and determining a target controller parameter based on the first controller parameter and the controller parameter error, and controlling the mobile robot to execute a next moving operation based on the target controller parameter.
On the basis of the foregoing embodiment, optionally, the second controller parameter determining module 320 is specifically configured to:
and determining a second controller parameter by adopting a preset closed-loop control algorithm based on the current actual track parameter and the current preset track parameter of the mobile robot.
On the basis of the above embodiment, optionally, the apparatus further includes:
and the current actual track parameter determining module is used for acquiring the current actual track parameters acquired by the visual equipment and/or determining the current actual track parameters based on the current sensor data.
The navigation device of the mobile robot provided by the embodiment of the invention can execute the navigation method of the mobile robot provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method.
Fig. 6 is a schematic structural diagram of a mobile robot according to an embodiment of the present invention. The mobile robot includes a drive controller 410 and at least one sensor 420, the sensor 420 for collecting sensor data. Fig. 6 exemplifies a case where two sensors 420 are mounted on the mobile robot. Illustratively, the at least one sensor 420 includes, but is not limited to, a odometer, a gyroscope, an inertial sensor, a pressure sensor, a temperature sensor, and the like. The components shown in the embodiments of the present invention, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
On the basis of the above embodiment, optionally, the mobile robot further includes a vision device 430 for acquiring current actual trajectory parameters. Illustratively, the vision device 430 may be a vision sensor.
Fig. 7 is a schematic structural diagram of a driving controller according to an embodiment of the present invention. As shown in fig. 7, the drive controller 410 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, and the like, wherein the memory stores a computer program executable by the at least one processor 11, and the processor 11 can perform various suitable actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the drive controller 410 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to the bus 14.
A plurality of components in the drive controller 410 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, or the like; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the drive controller 410 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, or the like. The processor 11 performs the various methods and processes described above, such as the navigation method of the mobile robot.
In some embodiments, the navigation method of the mobile robot may be implemented as a computer program that is tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the drive controller 410 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the navigation method of the mobile robot described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the navigation method of the mobile robot by any other suitable means (e.g. by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
An embodiment of the present invention further provides a computer-readable storage medium, where a computer instruction is stored, where the computer instruction is used to enable a processor to execute a navigation method for a mobile robot, and the method includes:
after the mobile robot executes the current moving operation, inputting the current sensor data of the mobile robot into a pre-trained target network model to obtain an output first controller parameter;
determining a second controller parameter based on the current actual track parameter and the current preset track parameter of the mobile robot;
and controlling the mobile robot to execute the next moving operation based on the first controller parameter and the second controller parameter.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service are overcome.
It should be understood that various forms of the flows shown above, reordering, adding or deleting steps, may be used. For example, the steps described in the present invention may be executed in parallel, sequentially, or in different orders, and are not limited herein as long as the desired results of the technical solution of the present invention can be achieved.
The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A method of navigating a mobile robot, comprising:
after a mobile robot executes current moving operation, inputting current sensor data of the mobile robot into a pre-trained target network model to obtain output first controller parameters;
determining a second controller parameter based on the current actual track parameter and the current preset track parameter of the mobile robot;
controlling the mobile robot to perform a next movement operation based on the first controller parameter and the second controller parameter.
2. The method of claim 1, wherein the method for training the target network model comprises:
inputting each training sensor data in the training sensor data set into an initial network model to obtain at least one output predictive controller parameter;
determining a loss function based on each of the predictive controller parameters and a standard controller parameter corresponding to each of the predictive controller parameters, respectively;
and adjusting model parameters of the initial network model based on the loss function until the loss function is converged to obtain a trained target network model.
3. The method according to claim 1 or 2, wherein the target network model comprises an output module and at least one long-short-term memory module, wherein the first long-short-term memory module is used for outputting a first long-short-term feature vector based on input current sensor data, the ith long-short-term memory module is used for outputting an ith long-short-term feature vector based on the input current sensor data and an ith-1 long-short-term feature vector output by the ith-1 long-term memory module, and the output module is used for outputting a first controller parameter based on an nth long-short-term feature vector output by the nth long-short-term memory module; and i is an integer which is greater than 1 and less than n, and n represents the number of long-time memory modules in the target network model.
4. The method according to claim 3, wherein the ith long-short-term memory module includes an input unit and a long-short-term memory unit, the input unit includes a first linear layer, a self-attention layer and a second linear layer, the first linear layer is used for performing linear coding on input current sensor data and outputting first sensor features, the self-attention layer is used for outputting fused sensor data based on the first sensor features output by the first linear layer, the second linear layer is used for performing linear coding on the fused sensor data output by the self-attention layer and outputting second sensor features, and the long-short-term memory unit is used for outputting an ith long-short-term feature vector based on the second sensor features output by the second linear layer and the ith-1 long-short-term feature vectors output by the ith-1 long-short-term memory module.
5. The method of claim 1, wherein controlling the mobile robot to perform a next movement operation based on the first controller parameter and the second controller parameter comprises:
determining a controller parameter error based on the first controller parameter and the second controller parameter;
determining a target controller parameter based on the first controller parameter and the controller parameter error, and controlling the mobile robot to perform a next movement operation based on the target controller parameter.
6. The method of claim 1, wherein determining a second controller parameter based on a current actual trajectory parameter and a current preset trajectory parameter of the mobile robot comprises:
and determining a second controller parameter based on the current actual track parameter and the current preset track parameter of the mobile robot by adopting a preset closed-loop control algorithm.
7. The method of claim 1, further comprising:
and acquiring current actual track parameters acquired by the vision equipment, and/or determining the current actual track parameters based on the current sensor data.
8. A navigation device of a mobile robot, comprising:
the first controller parameter determining module is used for inputting the current sensor data of the mobile robot into a pre-trained target network model after the mobile robot executes the current moving operation to obtain the output first controller parameter;
the second controller parameter determining module is used for determining second controller parameters based on the current actual track parameters and the current preset track parameters of the mobile robot;
and the mobile robot control module is used for controlling the mobile robot to execute the next moving operation based on the first controller parameter and the second controller parameter.
9. A mobile robot, characterized in that the mobile robot comprises: a drive controller and at least one sensor;
wherein the sensor is used for collecting sensor data;
the drive controller comprises at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to enable the at least one processor to perform the method of navigating a mobile robot of any one of claims 1-7.
10. A computer-readable storage medium, characterized in that it stores computer instructions for causing a processor, when executed, to implement the navigation method of a mobile robot of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211326593.1A CN115562299A (en) | 2022-10-27 | 2022-10-27 | Navigation method and device of mobile robot, mobile robot and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211326593.1A CN115562299A (en) | 2022-10-27 | 2022-10-27 | Navigation method and device of mobile robot, mobile robot and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115562299A true CN115562299A (en) | 2023-01-03 |
Family
ID=84768098
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211326593.1A Pending CN115562299A (en) | 2022-10-27 | 2022-10-27 | Navigation method and device of mobile robot, mobile robot and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115562299A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117348577A (en) * | 2023-12-05 | 2024-01-05 | 青岛宇方机器人工业股份有限公司 | Production process simulation detection method, device, equipment and medium |
-
2022
- 2022-10-27 CN CN202211326593.1A patent/CN115562299A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117348577A (en) * | 2023-12-05 | 2024-01-05 | 青岛宇方机器人工业股份有限公司 | Production process simulation detection method, device, equipment and medium |
| CN117348577B (en) * | 2023-12-05 | 2024-03-12 | 青岛宇方机器人工业股份有限公司 | Production process simulation detection method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210316447A1 (en) | Systems, devices, and methods for distributed artificial neural network computation | |
| EP3602409A1 (en) | Selecting actions using multi-modal inputs | |
| KR20190028531A (en) | Training machine learning models for multiple machine learning tasks | |
| EP3610418A1 (en) | Distributional reinforcement learning | |
| EP4010847A1 (en) | Training action selection neural networks using hindsight modelling | |
| CN110323981A (en) | A kind of method and system controlling permanent magnetic linear synchronous motor | |
| CN108762077B (en) | Mobile robot rolling time domain estimation method with communication constraint | |
| CN115562299A (en) | Navigation method and device of mobile robot, mobile robot and medium | |
| Mercat et al. | Kinematic single vehicle trajectory prediction baselines and applications with the NGSIM dataset | |
| CN112571420A (en) | Dual-function model prediction control method under unknown parameters | |
| Zhang et al. | An intelligent unmanned control method for redunant moving agent | |
| Zhu et al. | Improved PER-DDPG based nonparametric modeling of ship dynamics with uncertainty | |
| Vertzberger et al. | Adaptive attitude estimation using a hybrid model-learning approach | |
| Zheng et al. | An rnn-based learnable extended kalman filter design and application | |
| CN116736748B (en) | Method for constructing controller of robot and robot | |
| US20210405659A1 (en) | Deep model reference adaptive controller | |
| CN115099163A (en) | Charge state determining method and device, electronic equipment and storage medium | |
| CN114118371A (en) | Intelligent agent deep reinforcement learning method and computer readable medium | |
| CN109951874B (en) | Method for tracking moving unknown node in real time in sensor network | |
| Ma et al. | Trajectory Tracking Prediction of Multiple-UAVs Formation Based on Gated Cyclic Convolution Neural Network | |
| Sun et al. | Unmanned aerial vehicles control study using deep deterministic policy gradient | |
| Fan et al. | RNN-UKF: Enhancing Hyperparameter Auto-Tuning in Unscented Kalman Filters through Recurrent Neural Networks | |
| CN115600875B (en) | Environmental parameter calibration method and device, electronic equipment and storage medium | |
| CN117539153B (en) | Upper limb rehabilitation robot self-adaptive control method and system based on definite learning | |
| CN117784601B (en) | Omnidirectional mobile robot tracking control method based on event triggering and disturbance compensation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |