Wireless multi-source data fusion method for multi-component mixing spray gun control

Abstract

The invention provides a wireless multi-source data fusion method for multi-component mixing spray gun control, which comprises the following steps of: constructing a time sequence vector V, carrying out dimension raising processing and preprocessing on the time sequence vector, sending a signal to a wireless channel, superposing modulation signals of different sensor nodes to obtain a superposed signal, and fusing the superposed signal into a multi-component mixing spray gun; the method comprises the following steps of: acquiring a superposed signal, constructing a mean square error according to the superposed signal and a preset spraying decision function, constructing and solving an optimization problem to obtain an optimal sensor emission signal, sending the optimal sensor emission signal to a decision controller, and directly performing a spraying control decision according to the received signal by the decision controller to control spraying parameters of a multi-component mixing spray gun; according to the spray gun control system, dimension raising processing and preprocessing are conducted on the signals collected by the sensors, the preprocessed signals are directly transmitted to the decision controller, the decision controller directly receives the decision result, the spray gun is controlled more rapidly, and the control precision and effect of the spray gun are improved.

Description

Technical Field

[0001] The invention relates to the technical field of mixed spray gun control methods, and in particular to a wireless multi-source data fusion method for multi-component mixed spray gun control. Background Art

[0002] During the spraying process of a multi-component mixed spray gun, the spraying effect of the spray gun needs to be accurately controlled based on multiple parameters such as temperature, humidity, wind speed, wind direction, and particle concentration. Currently, multiple sensors are used to sample these parameters and wirelessly transmit the sampled data to the decision controller, which calculates the decision result for controlling the spray gun based on the received data. However, this approach has the following problems:

[0003] On the one hand, the sampling data between different types of sensors have low correlation. Simply combining the sampling data of various sensors will not construct a time series vector with insufficient dimensions, making it difficult to fully reflect the true state of the spraying environment, affecting the accuracy of decision-making. On the other hand, each sensor transmits data wirelessly independently. In the case of limited channel resources, the data transmission capacity and rate are limited. At the same time, after receiving the data from each sensor, the decision controller needs to restore and recalculate the data, resulting in a large delay from sampling to generating decision results, affecting the real-time nature of the spraying effect.

[0004] Therefore, the present invention proposes a wireless multi-source data fusion method for multi-component mixing spray gun control to solve the above problems. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a wireless multi-source data fusion method for multi-component mixing spray gun control. By performing dimensionality up-processing and preprocessing on the signals collected by each sensor, the preprocessed signals are directly transmitted to the decision controller. The decision controller directly receives the decision result, thereby achieving faster control of the spray gun and improving the control accuracy and effect of the spray gun.

[0006] In order to solve the above technical problems, the technical solution of the present invention is: a wireless multi-source data fusion method for multi-component mixing spray gun control, the innovation of which is that the method comprises the following steps:

[0007] Step 1: Obtain the perceived variables of the temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, and particle concentration sensor at time t, and construct a time series vector V, where each element of the time series vector is a sampling value of a sensor type at a sampling time of a sampling node;

[0008] Step 2: Perform dimension-upgrading processing on the time series vector constructed in step 1, where each element of the dimension-upgraded time series vector V' includes the sensor type x, the sampling node number y and the sampling time t;

[0009] Step 3: For each element in the time series vector after dimensionality increase, a preset preprocessing function f(V′) is used to perform preprocessing to obtain a preprocessed time series vector;

[0010] Step 4: Use the preset calibration coefficient θ(x, y, t) to control the transmit power p(x, y, t) of each sensor node, and use the binary phase shift keying modulation method to send the preprocessed timing vector signal to the wireless channel;

[0011] Step 5: In the wireless channel, the modulated signals from different sensor nodes are superimposed, and the superimposed signal is obtained at the channel receiving end. The mean square error is constructed by the superimposed signal and the preset spraying decision function, and the mean square error corresponds to the spraying effect;

[0012] Step 6: construct an optimization problem with minimizing the mean square error as the objective function and a preset transmit power value range as a constraint condition;

[0013] Step 7: Determine the optimal transmission power by solving the optimization problem, so as to obtain the optimal sensor transmission signal;

[0014] Step 8: Send the optimal sensor transmission signal to the decision controller, and the decision controller directly makes a spray control decision based on the received signal to control the spray parameters of the multi-component mixing spray gun.

[0015] Furthermore, the expression of the time series vector V constructed in step 1 is:

[0016] Formula 1: V=[V(x=1,t), V(x=2,t), V(x=3,t), V(x=4,t), V(x=5,t)]

[0017] Among them, V(x=1, t), V(x=2, t), V(x=3, t), V(x=4, t), and V(x=5, t) are the perception variables of the temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, and particle concentration sensor at time t, respectively.

[0018] Furthermore, the process of dimensionality upgrading of the time series vector in step 2 includes:

[0019] Step 2-1: Each type of sensor uses N sampling nodes, N ≥ 2, and the sampling correlation coefficient ρ between different sampling nodes of the same type of sensor is calculated to determine the measurement accuracy of the space. When ρ < 1, the sampling result can more accurately reflect the measurement accuracy of the space;

[0020] Step 2-2, perform dimension-upgrading processing on the time series vector V;

[0021] Step 2-3: Get the time series vector V' after dimensionality increase, which contains the sampling data of each sensor at different times and locations. The expression is:

[0022] Formula 2:

[0023] Furthermore, the expression of the preprocessing function f(V′) in step 3 is:

[0024] Formula 3:

[0025] Wherein, th is a given threshold. The value of each element in the dimension-upgraded time series V' is compared with the given threshold th. If the value of the element is greater than the given threshold th, it is set to 1. If the value of the element is less than the given threshold th, it is set to 0. The function matrix F(x, y, t) is constructed with f(V') as the element. The dimension of F(x, y, t) is the same as that of Formula 2.

[0026] Furthermore, the expression of the superimposed signal in step 5 is:

[0027] Formula 4:

[0028] Where y′ represents the superimposed signal, h(x, y, t) represents the transmission channel coefficient between the sensor node V′(x, y, t) and the decision controller, and n represents the received noise power at the decision controller.

[0029] Furthermore, the expression of the mean square error in step 5 is:

[0030] Formula 5: MSE = y′-f′

[0031] Where MSE represents mean square error, and f′ represents the ideal spraying decision function without delay.

[0032] Furthermore, the expression of the optimization problem in step 6 is:

[0033] Objective function: min{MSE}

[0034] Constraints: 0<p(x, y, t)<Th

[0035] Wherein, Th is the maximum allowed threshold of p(x, y, t).

[0036] Furthermore, in step 7, when solving the optimization problem, the Lagrange multiplier method is used to convert the optimization problem into a KKT problem. First, it is determined whether the constraints of the optimization problem are met. If so, the optimal solution of the KKT problem is solved by MATLAB simulation software to obtain the optimal sensor calibration coefficient. The optimal sensor calibration coefficient is configured to the corresponding sensor. The sensor sends a preprocessed signal according to the optimal calibration coefficient. The superimposed signal received by the decision controller is the optimal fusion calculation result. If the constraints are not met, the calibration coefficient of the sensor transmit power is adjusted, and steps 4 to 7 are repeated until the constraints are met.

[0037] Furthermore, the process of controlling the spraying parameters of the two-component mixing spray gun in step 8 includes:

[0038] Step 8-1, determine whether the spray control decision meets the preset control parameter conditions. If so, control the spray parameters of the multi-component mixing spray gun, including but not limited to spray speed, spray time and spray angle. If not, return to step 6 and rebuild the optimization problem for solution;

[0039] Step 8-2, obtain the real-time spraying effect data of the multi-component mixing spray gun, compare it with the preset target spraying effect, obtain the spraying effect deviation value, introduce the deviation value into the objective function of the optimization problem in step 6, dynamically optimize the sensor emission signal, and realize closed-loop control of the spraying effect.

[0040] The technical solution of the present invention includes the following beneficial effects:

[0041] (1) The present invention uses multiple sampling nodes to perform dimensionality-upgrading processing on sensor data, and calculates the sampling correlation coefficients between different sampling nodes of the same type of sensors, thereby improving the spatial measurement accuracy, preprocessing multi-source sensor data, achieving data cleaning and format unification, ensuring data quality and consistency, and transmitting the preprocessed data in BPSK form. The preprocessed data is superimposed in the air to form a calculation result. The controller directly receives the calculation result rather than the original perception data of each sensor, thereby greatly reducing the time delay from perception to control and improving the accuracy of the control effect.

[0042] (2) The present invention obtains the best sensor emission signal by solving the optimization problem, and directly obtains the best calculation result at the decision controller without the need for additional signal reception, recovery and calculation, thereby obtaining a high-precision and low-latency decision result, controlling the movement of the spray gun, and improving the control accuracy and effect of the spray gun. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1The present invention is a flow chart of a wireless multi-source data fusion method for multi-component mixing spray gun control. DETAILED DESCRIPTION

[0044] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0045] like Figure 1 As shown, this embodiment provides a wireless multi-source data fusion method for multi-component mixing spray gun control, comprising the following steps:

[0046] Step 1: Obtain the perceived variables of the temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, and particle concentration sensor at time t, and construct a time series vector V, where each element of the time series vector is a sampling value of a sensor type at a sampling time of a sampling node;

[0047] The expression of the time series vector V is:

[0048] Formula 1: V=[V(x=1,t), V(x=2,t), V(x=3,t), V(x=4,t), V(x=5,t)]

[0049] Among them, V(x=1, t), V(x=2, t), V(x=3, t), V(x=4, t), and V(x=5, t) are the perception variables of the temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, and particle concentration sensor at time t, respectively.

[0050] For example, at time t=1, the data collected by the temperature sensor is 25°C, the data collected by the humidity sensor is 60%, the wind speed sensor collects a wind speed of 1m / s, the wind direction sensor collects a wind direction of southeast, and the particle concentration sensor collects a particle concentration of 85μg / m 3 , then the corresponding 5-dimensional time series vector V = [25, 60, 1, 135, 85] is obtained. When the wind direction is represented here, the angle rotated clockwise from the north direction as the starting point is used as the wind direction data.

[0051] Step 2: Perform dimension-upgrading processing on the time series vector constructed in step 1, where each element of the dimension-upgraded time series vector V' includes the sensor type x, the sampling node number y and the sampling time t;

[0052] The process of dimensionality increase of time series vector includes:

[0053] Step 2-1: Each type of sensor uses N sampling nodes, N ≥ 2, and the sampling correlation coefficient ρ between different sampling nodes of the same type of sensor is calculated to determine the measurement accuracy of the space. When ρ < 1, the sampling result can more accurately reflect the measurement accuracy of the space;

[0054] Step 2-2: Perform dimension-upgrading on the time series vector V

[0055] Step 2-3: Get the time series vector V' after dimensionality increase, which contains the sampling data of each sensor at different times and locations. The expression is:

[0056] Formula 2:

[0057] In this embodiment, each type of sensor uses 5 sampling nodes, V′(1, 1, 1) represents the sampling value of the first temperature sensor at time t=1, V′(1, 2, t) represents the sampling value of the second temperature sensor at time t=1, and so on, to obtain a time series vector with a dimension of 25, that is:

[0058]

[0059] Step 3: For each element in the time series vector after dimensionality increase, a preset preprocessing function f(V′) is used to perform preprocessing to obtain a preprocessed time series vector;

[0060] The expression of the preprocessing function f(V′) is:

[0061] Formula 3:

[0062] Wherein, th is a given threshold. The value of each element in the dimension-upgraded time series V' is compared with the given threshold th. If the value of the element is greater than the given threshold th, it is set to 1. If the value of the element is less than the given threshold th, it is set to 0. The function matrix F(x, y, t) is constructed with f(V') as the element. The dimension of F(x, y, t) is the same as that of Formula 2.

[0063] Step 4: Use the preset calibration coefficient θ(x, y, t) to control the transmit power p(x, y, t) of each sensor node, and use the binary phase shift keying modulation method to send the preprocessed timing vector signal to the wireless channel;

[0064] Step 5: In the wireless channel, the modulated signals from different sensor nodes are superimposed, and the superimposed signal is obtained at the channel receiving end. The mean square error is constructed by the superimposed signal and the preset spraying decision function, and the mean square error corresponds to the spraying effect;

[0065] The expression of the superposition signal is:

[0066] Formula 4:

[0067] Wherein, y′ represents the superimposed signal, h(x, y, t) represents the transmission channel coefficient between the sensor node V′x, y, t) and the decision controller, and n represents the receiving noise power at the decision controller. In this embodiment, n is 10 dBm.

[0068] The expression of mean square error is:

[0069] Formula 5: MSE = y′-f′

[0070] Where MSE represents mean square error, and f′ represents the ideal spraying decision function without delay.

[0071] Step 6: construct an optimization problem with minimizing the mean square error as the objective function and a preset transmit power value range as a constraint condition;

[0072] The optimization problem is expressed as:

[0073] Objective function: min{MSE}

[0074] Constraints: 0<p(x, y, t)<Th

[0075] Wherein, Th is the maximum allowed threshold of p(x, y, t).

[0076] Step 7: Determine the optimal transmission power by solving the optimization problem, so as to obtain the optimal sensor transmission signal;

[0077] When solving the optimization problem, the Lagrange multiplier method is used to transform the optimization problem into a KKT problem. First, determine whether the constraints of the optimization problem are met. If so, the optimal solution of the KKT problem is solved by the fmincon function in the MATLAB simulation software optimization toolbox to obtain the optimal sensor calibration coefficient. The optimal sensor calibration coefficient is configured to the corresponding sensor. The sensor sends the preprocessed signal according to the optimal calibration coefficient. The superimposed signal received by the decision controller is the optimal fusion calculation result. If the constraints are not met, adjust the calibration coefficient of the sensor transmit power and repeat steps 4 to 7 until the constraints are met.

[0078] Step 8: Send the optimal sensor transmission signal to the decision controller, and the decision controller directly makes a spray control decision based on the received signal to control the spray parameters of the multi-component mixing spray gun;

[0079] The process of controlling the spraying parameters of a two-component mixing spray gun includes:

[0080] Step 8-1, determine whether the spray control decision meets the preset control parameter conditions. If so, control the spray parameters of the multi-component mixing spray gun, including but not limited to spray speed, spray time and spray angle. If not, return to step 6 and rebuild the optimization problem for solution;

[0081] Step 8-2, obtain the real-time spraying effect data of the multi-component mixing spray gun, compare it with the preset target spraying effect, obtain the spraying effect deviation value, introduce the deviation value into the objective function of the optimization problem in step 6, dynamically optimize the sensor emission signal, and realize closed-loop control of the spraying effect.

[0082] The simulation results show that when the calibration coefficient of the temperature sensor is 2, the calibration coefficient of the humidity sensor is 9, the calibration coefficient of the wind speed sensor is 5, the calibration coefficient of the wind direction sensor is 8, and the calibration coefficient of the particle concentration sensor is 1, the mean square error is the smallest. The superimposed signal at this time is the optimal multi-source sensor fusion calculation result. This result is used to control the spray gun, which can achieve high-precision spraying with an average error of less than 1 mm.

[0083] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A wireless multi-source data fusion method for multi-component mixing spray gun control, characterized in that: The method comprises the following steps: Step 1: Obtain the perceived variables of the temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, and particle concentration sensor at time t, and construct a time series vector V, where each element of the time series vector is a sampling value of a sensor type at a sampling time of a sampling node; Step 2: Perform dimension-upgrading processing on the time series vector constructed in step 1, where each element of the dimension-upgraded time series vector V' includes the sensor type x, the sampling node number y and the sampling time t; Step 3: For each element in the time series vector after dimensionality increase, a preset preprocessing function f(V′) is used to perform preprocessing to obtain a preprocessed time series vector; Step 4: Use the preset calibration coefficient θ(x, y, t) to control the transmit power p(x, y, t) of each sensor node, and use the binary phase shift keying modulation method to send the preprocessed timing vector signal to the wireless channel; Step 5: In the wireless channel, the modulated signals from different sensor nodes are superimposed, and the superimposed signal is obtained at the channel receiving end. The mean square error is constructed by the superimposed signal and the preset spraying decision function, and the mean square error corresponds to the spraying effect; Step 6: construct an optimization problem with minimizing the mean square error as the objective function and a preset transmit power value range as a constraint condition; Step 7: Determine the optimal transmission power by solving the optimization problem, so as to obtain the optimal sensor transmission signal; Step 8: Send the optimal sensor transmission signal to the decision controller, and the decision controller directly makes a spray control decision based on the received signal to control the spray parameters of the multi-component mixing spray gun.

2. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 1 is characterized in that: The expression of the time series vector V constructed in step 1 is: Formula 1: V=[V(x=1,t), V(x=2,t), V(x=3,t), V(x=4,t), V(x=5,t)] Among them, V(x=1, t), V(x=2, t), V(x=3, t), V(x=4, t), and V(x=5, t) are the perception variables of the temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, and particle concentration sensor at time t, respectively.

3. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 2 is characterized in that: The process of dimensionality upgrading of the time series vector in step 2 includes: Step 2-1: Each type of sensor uses N sampling nodes, N ≥ 2, and the sampling correlation coefficient ρ between different sampling nodes of the same type of sensor is calculated to determine the measurement accuracy of the space. When ρ < 1, the sampling result can more accurately reflect the measurement accuracy of the space; Step 2-2, perform dimension-upgrading processing on the time series vector V; Step 2-3: Get the time series vector V' after dimensionality increase, which contains the sampling data of each sensor at different times and locations. The expression is: Formula 2:

4. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 3 is characterized in that: The expression of the preprocessing function f(V′) in step 3 is: Formula 3: Wherein, th is a given threshold. The value of each element in the dimension-upgraded time series V' is compared with the given threshold th. If the value of the element is greater than the given threshold th, it is set to 1. If the value of the element is less than the given threshold th, it is set to 0. The function matrix F(x, y, t) is constructed with f(V′) as the element. The dimension of F(x, y, t) is the same as that of Formula 2.

5. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 4 is characterized in that: The expression of the superimposed signal in step 5 is: Formula 4: Where y′ represents the superimposed signal, h(x, y, t) represents the transmission channel coefficient between the sensor node V′(x, y, t) and the decision controller, and n represents the received noise power at the decision controller.

6. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 5 is characterized in that: The expression of the mean square error in step 5 is: Formula 5: MSE = y′-f′ Where MSE represents mean square error, and f′ represents the ideal spraying decision function without delay.

7. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 6 is characterized in that: The expression of the optimization problem in step 6 is: Objective function: min{MSE} Constraints: 0 <p(x,y,t)<Th Wherein, Th is the maximum allowed threshold of p(x, y, t).

8. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 7, characterized in that: In step 7, when solving the optimization problem, the Lagrange multiplier method is used to convert the optimization problem into a KKT problem. First, it is determined whether the constraints of the optimization problem are met. If so, the optimal solution of the KKT problem is solved by MATLAB simulation software to obtain the optimal sensor calibration coefficient. The optimal sensor calibration coefficient is configured to the corresponding sensor. The sensor sends a preprocessed signal according to the optimal calibration coefficient. The superimposed signal received by the decision controller is the optimal fusion calculation result. If the constraints are not met, the calibration coefficient of the sensor transmit power is adjusted, and steps 4 to 7 are repeated until the constraints are met.

9. The wireless multi-source data fusion method for multi-component mixing spray gun control according to claim 8, characterized in that: The process of controlling the spraying parameters of the two-component mixing spray gun in step 8 includes: Step 8-1, determine whether the spray control decision meets the preset control parameter conditions. If so, control the spray parameters of the multi-component mixing spray gun, including but not limited to spray speed, spray time and spray angle. If not, return to step 6 and rebuild the optimization problem for solution; Step 8-2, obtain the real-time spraying effect data of the multi-component mixing spray gun, compare it with the preset target spraying effect, obtain the spraying effect deviation value, introduce the deviation value into the objective function of the optimization problem in step 6, dynamically optimize the sensor emission signal, and realize closed-loop control of the spraying effect.

Images