Device for intelligently rectifying deviation of cleaning robot through cooperation of gyroscope and stepping motor

Abstract

The invention discloses an intelligent correction cleaning robot device with a gyroscope matched with a stepping motor, and the device comprises the gyroscope which is installed on a middle support and is used for monitoring the relative horizontal position of an upper transmission part and a lower transmission part in real time; the stepping motor is mounted on the upper and lower transmission brackets, is connected in series with the walking wheels through a transmission shaft and is used for driving the cleaning robot to advance; the control panel receives an offset signal of the gyroscope, performs offset training based on an LSTM neural network model, and generates a stride frequency adjustment instruction; the helical gear pair is installed on the transmission shaft, and the motion direction of the stepping motor is changed to adapt to different path requirements. The walking wheels and the depended wheels provide advancing power and assist in deviation correction. The device can effectively cope with deviation caused by topographic relief and inclination in the cleaning process, the stability of a cleaning path and the cleaning efficiency are remarkably improved, and the device has the advantages of being compact in structure and convenient and fast to install and is suitable for intelligent cleaning application to various complex terrains.

Description

Technical Field

[0001] The present invention relates to the field of intelligent cleaning, and more specifically to a device for an intelligent deviation-correcting cleaning robot using a gyroscope in conjunction with a stepping motor. Background Art

[0002] With the rapid development of the photovoltaic industry, photovoltaic panel cleaning has become a key link in ensuring the efficiency of photovoltaic power generation. However, photovoltaic panels are usually installed in open-air environments. After long-term exposure, they are prone to accumulate dust and dirt, which seriously affects the power generation efficiency. Therefore, the application of photovoltaic panel cleaning robots is becoming more and more widespread. However, during the cleaning process, due to the difference in friction on the surface of the photovoltaic panel and the uneven walking path, the cleaning robot may tilt or deviate during operation. This phenomenon will cause the robot's cleaning path to deviate, resulting in a decrease in the cleaning effect, and may even cause the robot to get stuck on the surface of the photovoltaic panel and fail to complete the cleaning task.

[0003] In order to ensure the stable operation of the cleaning robot on the photovoltaic panel, some attempts have been made in the prior art to use sensors and motors to adjust the path. However, traditional sensors can usually only detect simple deviation information, lack high-precision real-time feedback, and are difficult to deal with complex terrain and path differences. Stepper motors are widely used in a variety of precision positioning systems due to their precise control capabilities, but using stepper motors alone cannot completely solve the deviation problem caused by the robot during the cleaning process. Summary of the invention

[0004] In response to the above problems, this patent proposes a cleaning robot device that uses a gyroscope and a stepper motor for intelligent deviation correction. The gyroscope monitors the robot's posture changes such as yaw and tilt in real time during the cleaning process, and combined with the precise adjustment of the stepper motor, the cleaning robot can automatically correct its deviation on the predetermined path, thereby effectively avoiding the deviation caused by factors such as friction differences and changes in path conditions, ensuring that the cleaning robot is always in the best cleaning state. The specific plan is as follows:

[0005] A device for a cleaning robot that uses a gyroscope to cooperate with a stepper motor for intelligent deviation correction, the device comprising: a gyroscope, which is installed on an intermediate bracket and is used to monitor the relative horizontal positions of upper and lower transmission components of the cleaning robot in real time; a stepper motor, which is installed on upper and lower transmission brackets; a control board, which receives an offset signal from the gyroscope, performs offset training using an LSTM neural network model and generates a step frequency adjustment instruction; a bevel gear, which is used to change the movement direction of the stepper motor; a walking wheel and a supporting wheel, which assist in deviation correction and provide driving force; the device detects the horizontal deviation of the upper and lower transmission motor ends through the gyroscope, the control board analyzes the offset information and sends a pulse signal to the stepper motor, and the step frequency is adjusted through a transmission shaft and a walking wheel to achieve intelligent deviation correction.

[0006] Furthermore, the steps for using the LSTM neural network model to train the offset are as follows:

[0007] S1: Initialization phase, initializing the weight parameters of the LSTM neural network;

[0008] S2: Input the training offset value to the LSTM neural network to adjust the initial bias of the model to ensure that the model can capture the offset characteristics;

[0009] S3: Perform forward calculations on each neuron in the LSTM network step by step to obtain the output value of each node;

[0010] S4: Use the time back propagation algorithm to calculate the cumulative error of the network;

[0011] S5: Based on the accumulated error, calculate the gradient value of each weight in the neural network to obtain the optimal adjustment direction;

[0012] S6: Use the calculated gradient value to update the network weights, thereby gradually optimizing the LSTM model;

[0013] S7: Determine whether the model has converged, that is, whether the model error has reached the preset convergence standard. If the model has converged, the training ends; otherwise, return to the previous step and continue the training process.

[0014] Furthermore, the bevel gear is installed on the transmission shaft to change the rotation direction of the stepper motor, so that the power of the stepper motor is transmitted to the walking wheel through the transmission shaft, and the movement direction of the cleaning robot is adjusted when necessary.

[0015] Furthermore, the stepper motor is connected in series with the transmission shaft and the travel wheel.

[0016] Furthermore, the supporting wheels include an active supporting wheel installed on the upper transmission bracket and a driven supporting wheel installed on the lower transmission bracket; the active supporting wheel generates an active torque and moves together with the walking wheel.

[0017] Furthermore, the helical gear is installed on the travel path of the transmission shaft as a transmission pair device.

[0018] Furthermore, the control board can receive the deviation signal from the gyroscope in real time when the cleaning robot is tilted or the path is deviated, and calculate the step frequency adjustment amount of the stepper motor according to a preset correction algorithm to achieve dynamic correction.

[0019] Furthermore, during the offset training process of the LSTM neural network model, the weight of each neuron is optimized through multiple iterations and back propagation to improve the model's adaptability to different path offset modes, thereby ensuring accurate correction effects on various complex terrains.

[0020] Beneficial effects:

[0021] The present invention provides a device involving a gyroscope and a stepping motor for an intelligent deviation correction cleaning robot, which has the following beneficial effects:

[0022] (1) The robot's tilt and path deviation are monitored in real time by a gyroscope, and precise deviation correction is achieved through a stepper motor and transmission system. Combined with the dynamic learning function of the LSTM neural network, it can effectively cope with interference from different tilt angles and complex terrain, ensuring the stability and accuracy of the cleaning path.

[0023] (2) This device uses an LSTM neural network model and achieves adaptive adjustment through offset training and time back propagation algorithm. The device can automatically optimize the neural network weights according to different offsets and environmental changes, so that the cleaning robot can maintain the best moving state in a variety of environments and improve its intelligence level.

[0024] (3) The control board in the device can quickly respond to the offset signal from the gyroscope and control the stepper motor by adjusting the step frequency to correct the path deviation in the shortest time. This design effectively reduces the repeated movement caused by the offset during the cleaning process, thereby improving the cleaning efficiency, reducing power consumption, and extending the working time of the equipment.

[0025] (4) The device adopts a compact structural design, and the components such as gyroscope and stepper motor are reasonably arranged, which not only simplifies the mechanical structure of the device, but also makes installation and maintenance more convenient. In particular, the introduction of helical gear transmission pairs in the transmission system realizes the flexible change of transmission direction, further enhancing the structural adaptability of the device.

[0026] (5) The device is designed with a combination of active and passive wheels, which assists the movement of the walking wheels by distributing the active torque, thus achieving path stability of the cleaning robot on photovoltaic panels or other complex surfaces. This design can effectively avoid path deviation caused by tilt or obstacle interference of the robot, and improve the uniformity of the cleaning effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the device structure of a gyroscope cooperating with a stepper motor to perform an intelligent deviation correction cleaning robot.

[0028] Figure 2 This is the LSTM training iteration flow chart.

[0029] Figure 3 It is the working layout of the cleaning robot on the photovoltaic array.

[0030] Figure 4 is the working state A of the cleaning robot on the photovoltaic array.

[0031] Figure 5 It is the working state B of the cleaning robot on the photovoltaic array.

[0032] Figure 6 It is a flow chart of the cleaning robot’s deviation correction process.

[0033] Figure ID:

[0034] 1. Gyroscope; 2. Stepper motor; 3. Transmission shaft; 4. Bevel gear; 5. Control board; 6. Upper transmission bracket; 7. Lower transmission bracket; 8. Intermediate bracket; 9. Travel wheel; 10. Active supporting wheel; 11. Driven supporting wheel. DETAILED DESCRIPTION

[0035] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0036] Figure 1 The figure is a schematic diagram of the structure of a gyroscope combined with a stepper motor to perform intelligent deviation correction on a cleaning robot. Figure 1 As shown: the deviation correction device consists of a gyroscope 1 (used to monitor the yaw angle of the cleaning robot in real time), a stepper motor 2 (used to adjust the direction of the cleaning robot to keep it on a predetermined path), a transmission shaft 3 (transmitting power), a bevel gear 4 (changing the direction of movement), a control board 5 (used to receive gyroscope signals, analyze the yaw angle, and control the stepper motor for adjustment), an upper transmission bracket 6, a lower transmission bracket 7, an intermediate bracket 8, a walking wheel 9, an active supporting wheel 10, a driven supporting wheel 11, etc.

[0037] The gyroscope 1 of the correction device is installed at a fixed position on the intermediate bracket 8, a stepper motor 2 is installed on the lower transmission bracket 7, and is connected in series with the transmission shaft 3 and a group of running wheels 9, and the driven wheel 11 is freely fixed on the lower transmission bracket 7. Another stepper motor 2 is installed on the upper transmission bracket 6, and is connected in series with the transmission shaft 3 and another group of running wheels 9. The helical gear 4 transmission pair is installed on the path of the transmission shaft 3 to change the transmission direction. The active wheel 10 generates active torque and moves together with the running wheel 9.

[0038] Figure 2 This is the LSTM training iteration flow chart. Figure 2 As shown:

[0039] The steps for offset training using the LSTM neural network model are as follows:

[0040] S1: Initialization phase, initializing the weight parameters of the LSTM neural network;

[0041] S2: Input the training offset value to the LSTM neural network to adjust the initial bias of the model to ensure that the model can capture the offset characteristics;

[0042] S3: Perform forward calculations on each neuron in the LSTM network step by step to obtain the output value of each node;

[0043] S4: Use the time back propagation algorithm to calculate the cumulative error of the network;

[0044] S5: Based on the accumulated error, calculate the gradient value of each weight in the neural network to obtain the optimal adjustment direction;

[0045] S6: Use the calculated gradient value to update the network weights, thereby gradually optimizing the LSTM model;

[0046] S7: Determine whether the model has converged, that is, whether the model error has reached the preset convergence standard. If the model has converged, the training ends; otherwise, return to the previous step and continue the training process.

[0047] like Figure 6 The following is a main diagram of the cleaning robot correcting the deviation based on sensor data, which includes the following main steps:

[0048] function cleanRobotCorrectionAlgorithm():

[0049] # Initialize sensor data and control parameters

[0050] initializeData Init()

[0051] initializeControlParameters()

[0052] Data_Send

[0053] while cleaningTaskNotCompleted():

[0054] #Read sensor data

[0055] sensorData=read Offet_Data()

[0056] #Calculate attitude angle and position deviation

[0057] attitudeAngles=calculateAttitudeAngles(Ofet_Data)

[0058] positionDeviation=calculatePositionDeviation(Ofet_Action)

[0059] #Determine whether correction is needed

[0060] if needCorrection(attitudeAngles,positionDeviation):

[0061] #Generate control instructions

[0062] controlCommands=generateControlCommands(P_calc I_calc)

[0063] #Execute control instructions

[0064] executeControlCommands(controlCommands)

[0065] #Update sensor data and control parameters

[0066] updateSensorData()

[0067] updateControlParameters()

[0068] #Cleaning task completed

[0069] endCleaningTask()

[0070] The two stepper motors 2 placed on the lower transmission bracket 7 and the upper transmission bracket 6 are finally connected to the control board 5 according to the wiring rules to train the offset value, as follows Figure 3 As shown, the deceleration point and acceleration point critical values ​​are initially preset to prepare for the training offset value. When the equipment is running, 1 gyroscope 1 monitors the relative horizontal position of the upper and lower transmission ends. When the upper transmission motor end leads the lower transmission motor end, Figure 4 , the information is fed back to the control board, and the control board inputs the training offset value to the LSTM neural network, so that the stepper motor 2 receives the corresponding electric pulse signal, transmits the step frequency through the transmission shaft 3, and feeds back to the walking wheel 9 and the active supporting wheel 10, and performs the upper transmission motor end correction action. The driven supporting wheel 11 is located at the lower transmission end to assist the mechanical correction. Similarly, the gyroscope 1 monitors the relative horizontal position of the upper and lower transmission ends. When the lower transmission motor side leads the upper transmission motor end, as shown in Figure 5, the control board 5 trains the same principle action, but the reverse cumulative error of the timing sequence must be counted. Similarly, the stepper motor 2 receives the corresponding electric pulse signal, transmits the step frequency through the transmission axis 3, and feeds back to the walking wheel 9. The driven wheel 11 is located at the lower transmission end to assist the mechanical correction.

[0071] As a further improvement, the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A device for an intelligent deviation-correcting cleaning robot using a gyroscope in conjunction with a stepping motor, characterized in that: The device comprises: a gyroscope, which is installed on an intermediate bracket and is used to monitor the relative horizontal positions of the upper and lower transmission components of the cleaning robot in real time; a stepper motor, which is installed on the upper and lower transmission brackets; a control board, which receives the offset signal of the gyroscope, uses an LSTM neural network model to perform offset training and generate a step frequency adjustment instruction; a bevel gear, which is used to change the movement direction of the stepper motor; a walking wheel and a supporting wheel, which assist in deviation correction and provide driving force; the device detects the horizontal deviation of the upper and lower transmission motor ends through the gyroscope, the control board analyzes the deviation information and sends a pulse signal to the stepper motor, and adjusts the step frequency through the transmission shaft and the walking wheel to realize intelligent deviation correction.

2. The device of the gyroscope cooperating with the stepping motor to perform intelligent deviation correction cleaning robot according to claim 1 is characterized in that: The steps for offset training using the LSTM neural network model are as follows: S1: Initialization phase, initializing the weight parameters of the LSTM neural network; S2: Input the training offset value to the LSTM neural network to adjust the initial bias of the model to ensure that the model can capture the offset characteristics; S3: Perform forward calculations on each neuron in the LSTM network step by step to obtain the output value of each node; S4: Use the time back propagation algorithm to calculate the cumulative error of the network; S5: Based on the accumulated error, calculate the gradient value of each weight in the neural network to obtain the optimal adjustment direction; S6: Use the calculated gradient value to update the network weights, thereby gradually optimizing the LSTM model; S7: Determine whether the model has converged, that is, whether the model error has reached the preset convergence standard. If the model has converged, the training ends; otherwise, return to the previous step and continue the training process.

3. The device of the gyroscope cooperating with the stepping motor to perform intelligent deviation correction cleaning robot according to claim 1, characterized in that: The bevel gear is installed on the transmission shaft and is used to change the rotation direction of the stepper motor, so that the power of the stepper motor is transmitted to the walking wheel through the transmission shaft, and the movement direction of the cleaning robot is adjusted when necessary.

4. The device of the gyroscope cooperating with the stepping motor to perform intelligent deviation correction cleaning robot according to claim 1, characterized in that: The stepper motor is connected in series with the transmission shaft and the travel wheel.

5. The device of the gyroscope cooperating with the stepping motor to perform intelligent deviation correction cleaning robot according to claim 1, characterized in that: The supporting wheels include an active supporting wheel installed on the upper transmission bracket and a driven supporting wheel installed on the lower transmission bracket; the active supporting wheel generates an active torque and moves together with the walking wheel.

6. The device of the gyroscope cooperating with the stepping motor to perform intelligent deviation correction cleaning robot according to claim 1, characterized in that: The helical gear is installed on the transmission shaft path as a transmission pair device.

7. The device of the gyroscope cooperating with the stepping motor to perform intelligent deviation correction cleaning robot according to claim 1, characterized in that: The control board can receive the deviation signal from the gyroscope in real time when the cleaning robot is tilted or the path is offset, and calculate the step frequency adjustment amount of the stepping motor according to a preset correction algorithm to achieve dynamic correction.

8. The device of the gyroscope cooperating with the stepping motor to perform intelligent deviation correction cleaning robot according to claim 2, characterized in that: During the offset training process of the LSTM neural network model, the weight of each neuron is optimized through multiple iterations and back propagation to improve the model's adaptability to different path offset modes, thereby ensuring accurate correction effects on various complex terrains.

Images