An AI-based control method and a disinfection robot

Abstract

The present invention discloses a control method and a disinfection robot based on artificial intelligence, belonging to the field of automatic control technology. According to the air cleanliness and humidity in the environment, the present invention calculates the target spraying speed, and then, according to the target spraying speed, uses a BP neural network to predict the proportional coefficient, integral coefficient, and differential coefficient, and assigns the initial proportional coefficient, integral coefficient, and differential coefficient to the third-order PID model, facilitating the rapid response and convergence of the third-order PID model. Then, the third-order PID model is used for control, and according to the spraying speed change coefficient and spraying speed oscillation coefficient, the proportion, integral, and differential are adjusted to achieve the rapid and stable control of the spraying speed and improve the disinfection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic control, and particularly relates to a control method based on artificial intelligence and a disinfection robot. Background Art

[0002] With the development of technology, automation and intelligent technologies have been widely applied in multiple fields, including industries with extremely high environmental hygiene requirements such as medical treatment, food processing, and pharmaceuticals. In these industries, disinfection work is a key step to ensure product safety and public health. Traditional disinfection methods mostly rely on manual operation or simple automated equipment, and there are problems such as uneven disinfection and resource waste. When disinfecting the environment currently, a unified spraying speed of disinfectant water is adopted to spray disinfectant water comprehensively in the space, resulting in over-disinfection of areas with high air cleanliness and incomplete disinfection of areas with low air cleanliness, and there is a problem of low disinfection accuracy. Summary of the Invention

[0003] Aiming at the above deficiencies in the prior art, a control method based on artificial intelligence and a disinfection robot provided by the present invention solve the problem of low disinfection accuracy existing in the prior art.

[0004] In order to achieve the above invention purpose, the technical solution adopted by the present invention is as follows: A control method based on artificial intelligence, comprising the following steps:

[0005] S1. Calculate the target spraying speed according to the air cleanliness and humidity in the environment;

[0006] S2. Predict the proportional coefficient, integral coefficient, and differential coefficient by using a BP neural network according to the target spraying speed;

[0007] S3. Allocate the predicted proportional coefficient, integral coefficient, and differential coefficient to a third-order PID model;

[0008] S4. Perform control by using the third-order PID model, and adjust the predicted proportional coefficient, integral coefficient, and differential coefficient according to the spraying speed change coefficient and the spraying speed oscillation coefficient to obtain a third-order PID model with adjustment completed;

[0009] S5. Control the spraying speed by using the third-order PID model with adjustment completed.

[0010] Further, the S1 includes the following sub-steps:

[0011] S11. Calculate the first distance coefficient according to the distance between the air cleanliness in the environment and the target air cleanliness;

[0012] S12. Calculate the second distance coefficient according to the distance between the humidity in the environment and the target humidity;

[0013] S13. Calculate the target spraying speed according to the first distance coefficient and the second distance coefficient.

[0014] Further, the formula for calculating the target spraying speed in S13 is: , where v target is the target spraying speed, v0 is the initial spraying speed, α1 is the first weighting coefficient, α2 is the second weighting coefficient, max is to take the larger value, d1 is the first distance coefficient, and d2 is the second distance coefficient.

[0015] Further, the third-order PID model in S3 includes: a first PI sub-model, a second PI sub-model, a third PI sub-model, an adder A1, and a differentiator model;

[0016] The input end of the first PI sub-model is respectively connected to the input ends of the second PI sub-model and the third PI sub-model, and serves as the input end of the third-order PID model;

[0017] The input ends of the adder A1 are respectively connected to the output ends of the first PI sub-model, the second PI sub-model, and the third PI sub-model, and its output end is connected to the input end of the differentiator model;

[0018] The output end of the differentiator model serves as the output end of the third-order PID model.

[0019] Further, the expression of the differentiator model is: , , where U(t) is the control quantity output by the differentiator model at the t-th moment, H(t) is the control quantity output by the adder A1 at the t-th moment, u1(t) is the control quantity output by the first PI sub-model at the t-th moment, u2(t) is the control quantity output by the second PI sub-model at the t-th moment, u3(t) is the control quantity output by the third PI sub-model at the t-th moment, e(t) is the speed difference between the target spraying speed and the real-time spraying speed at the t-th moment, e(t - 1) is the speed difference between the target spraying speed and the real-time spraying speed at the (t - 1)-th moment, and K d is the predicted differential coefficient.

[0020] Further, S4 includes the following sub-steps:

[0021] S41. Perform control using the third-order PID model. During the control time T, collect the real-time spraying speed and construct a real-time spraying speed sequence;

[0022] S42. Divide the real-time spraying speed sequence into two parts to obtain a first spraying speed sequence and a second spraying speed sequence;

[0023] S43. Calculate the spraying speed change coefficient according to the first spraying speed sequence;

[0024] S44. Calculate the spraying speed oscillation coefficient according to the second spraying speed sequence;

[0025] S45. Adjust the predicted proportional coefficient and integral coefficient according to the spraying speed change coefficient;

[0026] S46. Adjust the predicted differential coefficient according to the spraying speed oscillation coefficient.

[0027] Furthermore, the formula for calculating the spraying speed change coefficient in S43 is: , where γ c is the spraying speed change coefficient, v n is the nth spraying speed in the first spraying speed sequence, v n-1 is the (n - 1)th spraying speed in the first spraying speed sequence, n is a positive integer, N is the number of elements in the first spraying speed sequence, | | is the absolute value, and S1 is the first normalization coefficient;

[0028] The formula for calculating the spraying speed oscillation coefficient in S44 is: , where γ s is the spraying speed oscillation coefficient, v m is the mth spraying speed in the second spraying speed sequence, M is the number of elements in the second spraying speed sequence, m is a positive integer, and S2 is the second normalization coefficient.

[0029] Furthermore, the formula for adjusting the predicted proportional coefficient in S45 is: , where, is the adjusted proportional coefficient, K p is the predicted proportional coefficient, γ c is the spraying speed change coefficient, γ c,th is the spraying speed change threshold, and e is the natural constant;

[0030] The formula for adjusting the predicted integral coefficient in S45 is: , where, is the adjusted integral coefficient, K i is the predicted integral coefficient.

[0031] Furthermore, the formula for adjusting the predicted differential coefficient in S46 is: , where, is the adjusted differential coefficient, K d is the predicted differential coefficient, γ s is the spraying speed oscillation coefficient, γ s,this the spraying speed oscillation threshold, and e is the natural constant.

[0032] An artificial intelligence-based disinfection robot implemented by an artificial intelligence-based control method, including: a target spraying speed calculation unit, a prediction unit, an allocation unit, an adjustment unit, and a control unit;

[0033] The target spraying speed calculation unit is used to calculate the target spraying speed according to the air cleanliness and humidity in the environment;

[0034] The prediction unit is used to predict the proportional coefficient, integral coefficient, and differential coefficient by using a BP neural network according to the target spraying speed;

[0035] The allocation unit is used to allocate the predicted proportional coefficient, integral coefficient, and differential coefficient to a third-order PID model;

[0036] The adjustment unit is used to perform control by using a third-order PID model, and adjust the predicted proportional coefficient, integral coefficient, and differential coefficient according to the spraying speed change coefficient and the spraying speed oscillation coefficient to obtain a third-order PID model with adjustment completed;

[0037] The control unit is used to control the spraying speed by using the third-order PID model with adjustment completed.

[0038] In summary, the beneficial effects of the present invention are as follows: The present invention calculates the target spraying speed according to the air cleanliness and humidity in the environment, and then predicts the proportional coefficient, integral coefficient, and differential coefficient by using a BP neural network according to the target spraying speed, and allocates the initial proportional coefficient, integral coefficient, and differential coefficient to the third-order PID model, which is convenient for the third-order PID model to respond and converge quickly. Then, the third-order PID model is used for control, and the proportional, integral, and differential are adjusted according to the spraying speed change coefficient and the spraying speed oscillation coefficient, so as to achieve fast and stable control of the spraying speed and improve the disinfection accuracy.

[0039] On the one hand, the present invention considers the air cleanliness and humidity in the environment, so as to allocate the corresponding target spraying speed, making the disinfection degree of different environments different and improving the disinfection accuracy. On the other hand, the coefficients of the PID model are predicted and adjusted, making the PID model fast and stable, and further improving the disinfection accuracy. Description of the Drawings

[0040] Figure 1 is a system block diagram of an artificial intelligence-based control method;

[0041] Figure 2 is a structural schematic diagram of a third-order PID model. Specific Embodiments

[0042] The specific embodiments of the present invention will be described below to facilitate those skilled in the art of this technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art of this technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions made using the concept of the present invention are within the scope of protection.

[0043] Example 1: As Figure 1 shown, an artificial intelligence-based control method includes the following steps:

[0044] S1. Calculate the target spraying speed according to the air cleanliness and humidity in the environment;

[0045] S2. Predict the proportional coefficient, integral coefficient, and differential coefficient using a BP neural network according to the target spraying speed;

[0046] S3. Assign the predicted proportional coefficient, integral coefficient, and differential coefficient to a third-order PID model;

[0047] S4. Perform control using the third-order PID model, and adjust the predicted proportional coefficient, integral coefficient, and differential coefficient according to the spraying speed change coefficient and spraying speed oscillation coefficient to obtain a third-order PID model with the adjustment completed;

[0048] S5. Control the spraying speed using the third-order PID model with the adjustment completed.

[0049] In this embodiment, the BP neural network includes an input layer, a hidden layer, and an output layer. The input layer is used to input the target spraying speed, and the output layer is used to output the predicted proportional coefficient, integral coefficient, and differential coefficient.

[0050] In this embodiment, the BP neural network is trained using samples composed of the target spraying speed and labels, and the labels are: proportional coefficient, integral coefficient, and differential coefficient. It is realized to allocate different proportional coefficients, integral coefficients, and differential coefficients according to different target spraying speeds, so that the PID model has the ability of rapid response.

[0051] The said S1 includes the following sub-steps:

[0052] S11. Calculate the first distance coefficient according to the distance between the air cleanliness in the environment and the target air cleanliness;

[0053] S12. Calculate the second distance coefficient according to the distance between the humidity in the environment and the target humidity;

[0054] S13. Calculate the target spraying speed according to the first distance coefficient and the second distance coefficient.

[0055] The formula for calculating the first distance coefficient in S11 is as follows: , where d1 is the first distance coefficient, C is the air cleanliness in the environment, and C target is the target air cleanliness;

[0056] The formula for calculating the second distance coefficient in S12 is as follows: , where d2 is the second distance coefficient, h is the humidity in the environment, and h target is the target humidity.

[0057] The formula for calculating the target spraying speed in S13 is as follows: , where v target is the target spraying speed, v0 is the initial spraying speed, α1 is the first weighting coefficient, α2 is the second weighting coefficient, max is to take the larger value, d1 is the first distance coefficient, and d2 is the second distance coefficient.

[0058] In the present invention, the air cleanliness is evaluated by measuring the concentration of suspended particles in the air. The greater the air cleanliness, the higher the concentration of suspended particles.

[0059] When the first distance coefficient is larger, the air cleanliness is higher and the concentration of suspended particles is higher. When the second distance coefficient is larger, the humidity is greater. In an environment with high humidity and high concentration of suspended particles, bacteria are more likely to grow. Therefore, the greater the target spraying speed allocated to it, so as to achieve the allocation of different amounts of disinfectant water according to different environments.

[0060] In this embodiment, the first weighting coefficient and the second weighting coefficient are specifically set according to requirements.

[0061] As Figure 2 shown, the third-order PID model in S3 includes: a first PI sub-model, a second PI sub-model, a third PI sub-model, an adder A1, and a differentiator model;

[0062] The input ends of the first PI sub-model are respectively connected to the input ends of the second PI sub-model and the third PI sub-model, and serve as the input ends of the third-order PID model;

[0063] The input ends of the adder A1 are respectively connected to the output ends of the first PI sub-model, the second PI sub-model, and the third PI sub-model, and its output end is connected to the input end of the differentiator model;

[0064] The output end of the differentiator model serves as the output end of the third-order PID model.

[0065] In this embodiment, the third-order PID model includes: two proportional coefficients, two integral coefficients, and one differential coefficient. Therefore, the output nodes of the output layer of the BP neural network are set to 5.

[0066] The expressions of the first PI sub-model, the second PI sub-model, and the third PI sub-model are as follows: , where u(t) is the control quantity output by the first PI sub-model, the second PI sub-model, and the third PI sub-model at the t-th moment, e(t) is the speed difference between the target spraying speed and the real-time spraying speed at the t-th moment, K p is the predicted proportionality coefficient, K i is the predicted integral coefficient, e(τ) is the speed difference between the target spraying speed and the real-time spraying speed at the τ-th moment, t is the time number, and τ is the integration variable.

[0067] The expression of the micro sub-model is as follows: , , where U(t) is the control quantity output by the micro sub-model at the t-th moment, H(t) is the control quantity output by the adder A1 at the t-th moment, u1(t) is the control quantity output by the first PI sub-model at the t-th moment, u2(t) is the control quantity output by the second PI sub-model at the t-th moment, u3(t) is the control quantity output by the third PI sub-model at the t-th moment, e(t) is the speed difference between the target spraying speed and the real-time spraying speed at the t-th moment, e(t - 1) is the speed difference between the target spraying speed and the real-time spraying speed at the (t - 1)-th moment, and K d is the predicted differential coefficient.

[0068] The present invention adopts a three-layer PI sub-model, and then uses an adder A1 to add the control quantities output by the three-layer PI sub-model, and combines the control quantity output by the micro sub-model to obtain the control quantity output by the third-order PID model. The present invention divides the PI part in the original PID into 3 layers, improves the response speed of the PID model to the speed difference, optimizes the performance of the traditional PID control, and also enhances the adaptability and robustness of the system, and is particularly suitable for occasions that require fast and accurate response.

[0069] The S4 includes the following sub-steps:

[0070] S41. Use the third-order PID model for control. During the control time T, collect the real-time spraying speed and construct a real-time spraying speed sequence;

[0071] In this embodiment, the time T is selected as a time less than 30S such as 10 seconds, 20 seconds, or 30 seconds;

[0072] S42. Divide the real-time spraying speed sequence into two parts to obtain a first-segment spraying speed sequence and a second-segment spraying speed sequence;

[0073] S43. Calculate the spraying speed change coefficient according to the first-segment spraying speed sequence;

[0074] S44. Calculate the spraying speed oscillation coefficient according to the second spraying speed sequence;

[0075] S45. Adjust the predicted proportional coefficient and integral coefficient according to the spraying speed change coefficient;

[0076] S46. Adjust the predicted differential coefficient according to the spraying speed oscillation coefficient.

[0077] When adjusting the predicted proportional coefficient and integral coefficient, one of the PI sub-models in the third-order PID model can be adjusted. Observe the spraying speed change coefficient and the spraying speed oscillation coefficient. If the spraying speed change coefficient is greater than the spraying speed change threshold and the spraying speed oscillation coefficient is less than the spraying speed oscillation threshold, the adjustment is completed. If the conditions are not met after adjusting one layer of the PI sub-model, then adjust the second layer of the PI sub-model and check again whether the conditions are not met. If the conditions are not met, then adjust the third layer of the PI sub-model.

[0078] After the present invention distributes the predicted proportional coefficient, integral coefficient, and differential coefficient to the third-order PID model and uses this third-order PID model for control, within time T, a real-time spraying speed sequence is obtained. Then, the real-time spraying speed sequence is divided into two parts. The first spraying speed sequence is the spraying speed collected in the first half of the time, and the second spraying speed sequence is the spraying speed collected in the second half of the time. Calculate the spraying speed change coefficient for the first spraying speed sequence to reflect the speed change rate, so as to facilitate testing the convergence and control ability of the third-order PID in a short time. Calculate the spraying speed oscillation coefficient for the second spraying speed sequence to reflect the volatility of the speed change, so as to facilitate testing the stability of the output after short-time control of the third-order PID.

[0079] The formula for calculating the spraying speed change coefficient in S43 is: , where γ c is the spraying speed change coefficient, v n is the nth spraying speed in the first spraying speed sequence, v n-1 is the (n - 1)th spraying speed in the first spraying speed sequence, n is a positive integer, N is the number of elements in the first spraying speed sequence, | | is the absolute value, and S1 is the first normalization coefficient.

[0080] The present invention obtains the spraying speed change coefficient through the sum of the absolute values of the differences between all adjacent two spraying speeds in the first spraying speed sequence, reflecting the change of the spraying speed during this period.

[0081] The formula for calculating the spraying speed oscillation coefficient in S44 is: , where γ s is the spraying speed oscillation coefficient, v mis the m-th spraying speed in the second spraying speed sequence, M is the number of elements in the second spraying speed sequence, m is a positive integer, and S2 is the second normalization coefficient.

[0082] In the present invention, the spraying speed oscillation coefficient is obtained by the sum of the absolute values of the differences between all spraying speeds in the second spraying speed sequence and the average spraying speed, reflecting the oscillation condition of the spraying speed during this period.

[0083] The formula for adjusting the predicted proportional coefficient in S45 is: , where is the adjusted proportional coefficient, K p is the predicted proportional coefficient, γ c is the spraying speed change coefficient, γ c,th is the spraying speed change threshold, and e is the natural constant.

[0084] When the spraying speed change coefficient is greater than or equal to the spraying speed change threshold, the response speed of the third-order PID model is fast. Therefore, the predicted proportional coefficient is retained. When the spraying speed change coefficient is less than the spraying speed change threshold, the response speed of the third-order PID model is slow. According to the gap between the spraying speed change threshold and the spraying speed change coefficient, the proportional coefficient is increased to accelerate the response speed of the third-order PID model.

[0085] The formula for adjusting the predicted integral coefficient in S45 is: , where is the adjusted integral coefficient, K i is the predicted integral coefficient.

[0086] When the spraying speed change coefficient is greater than or equal to the spraying speed change threshold, the response speed of the third-order PID model is fast. If the integral action is too strong, it will cause the control output to accumulate excessively, leading to system overshoot or long-term oscillation. According to the gap between the spraying speed change threshold and the spraying speed change coefficient, the integral coefficient is decreased, which can reduce the accumulation speed of the integral action on historical errors and help alleviate the overshoot problem caused by integral saturation. When the spraying speed change coefficient is less than the spraying speed change threshold, the response speed of the third-order PID model is slow, and the predicted integral coefficient is retained. Appropriate integration can still ensure that the steady-state error is gradually eliminated when the error is small, ensuring that the system finally stabilizes at the set value.

[0087] The formula for adjusting the predicted derivative coefficient in S46 is: , where is the adjusted derivative coefficient, K d is the predicted derivative coefficient, γ s is the spraying speed oscillation coefficient, γ s,th is the spraying speed oscillation threshold, and e is the natural constant.

[0088] When the spraying speed oscillation coefficient is less than the spraying speed oscillation threshold, the control quantity output by the third-order PID stabilizes the spraying speed at the target spraying speed. Therefore, the predicted differential coefficient is retained. When the spraying speed oscillation coefficient is greater than or equal to the threshold, the differential coefficient is increased according to the difference between the spraying speed oscillation threshold and the spraying speed oscillation coefficient, which can enhance the sensitivity to the current error, make the control output respond more quickly, increase the damping of the system, and help reduce overshoot and improve system stability.

[0089] In this embodiment, the spraying speed change threshold is the threshold set for the spraying speed change coefficient, and the spraying speed oscillation threshold is the threshold set for the spraying speed oscillation coefficient, which can be specifically set through experiments or experience.

[0090] Embodiment 2: A disinfection robot based on artificial intelligence, implemented based on the control method of artificial intelligence in Embodiment 1, including: a target spraying speed calculation unit, a prediction unit, a distribution unit, an adjustment unit, and a control unit;

[0091] The target spraying speed calculation unit is used to calculate the target spraying speed according to the air cleanliness and humidity in the environment;

[0092] The prediction unit is used to predict the proportional coefficient, integral coefficient, and differential coefficient by using a BP neural network according to the target spraying speed;

[0093] The distribution unit is used to distribute the predicted proportional coefficient, integral coefficient, and differential coefficient to the third-order PID model;

[0094] The adjustment unit is used to perform control by using the third-order PID model, and adjust the predicted proportional coefficient, integral coefficient, and differential coefficient according to the spraying speed change coefficient and the spraying speed oscillation coefficient to obtain a third-order PID model with adjustment completed;

[0095] The control unit is used to control the spraying speed by using the third-order PID model with adjustment completed.

[0096] The specific implementation manner of Embodiment 2 is the same as that of Embodiment 1.

[0097] The present invention calculates the target spraying speed according to the air cleanliness and humidity in the environment, then predicts the proportional coefficient, integral coefficient, and differential coefficient by using a BP neural network according to the target spraying speed, distributes the initial proportional coefficient, integral coefficient, and differential coefficient to the third-order PID model to facilitate the quick response and convergence of the third-order PID model, and then performs control by using the third-order PID model, and adjusts the proportion, integral, and differential according to the spraying speed change coefficient and the spraying speed oscillation coefficient to achieve the quick and stable control of the spraying speed and improve the disinfection accuracy.

[0098] On the one hand, the present invention takes into account the air cleanliness and humidity in the environment, so as to allocate the corresponding target spraying speed, making the disinfection degree different in different environments and improving the disinfection accuracy. On the other hand, the coefficients of the PID model are predicted and adjusted, making the PID model fast and stable and further improving the disinfection accuracy.

[0099] The proportional coefficient, integral coefficient and differential coefficient predicted by the present invention are equivalent to giving initial values to the PID model, and then adjusting the proportional coefficient, integral coefficient and differential coefficient according to the situation within the control time T to make it quickly adapt to the current control situation. The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A control method based on artificial intelligence, characterized in that: The following steps are involved: S1. Calculate the target spraying speed according to the air cleanliness and humidity in the environment; S2. According to the target spraying speed, the BP neural network is used to predict the proportional coefficient, integral coefficient and differential coefficient; S3, assigning the predicted proportional coefficient, integral coefficient and differential coefficient to the third-order PID model; S4, using a third-order PID model for control, adjusting the predicted proportional coefficient, integral coefficient and differential coefficient according to the spraying speed variation coefficient and the spraying speed oscillation coefficient, and obtaining an adjusted third-order PID model; S5, using the adjusted third-order PID model to control the spraying speed; The S4 comprises the following sub-steps: S41, using a third-order PID model for control, collecting real-time spraying speed within the controlled T time, and constructing a real-time spraying speed sequence; S42, dividing the real-time spraying speed sequence into two parts to obtain a first spraying speed sequence and a second spraying speed sequence; S43, calculating the spraying speed variation coefficient according to the first spraying speed sequence; S44, calculating the spraying speed oscillation coefficient according to the second spraying speed sequence; S45, adjusting the predicted proportional coefficient and integral coefficient according to the spraying speed variation coefficient; S46, adjusting the predicted differential coefficient according to the spraying speed oscillation coefficient; The formula for calculating the spraying speed variation coefficient in S43 is: , where γ c is the spraying speed variation coefficient, v n is the nth spraying speed in the first spraying speed sequence, v n-1 is the n-1th spraying speed in the first spraying speed sequence, n is a positive integer, N is the number of elements in the first spraying speed sequence, | | is the absolute value, and S1 is the first normalization coefficient; The formula for calculating the spraying speed oscillation coefficient in S44 is: , where γ s is the spraying velocity oscillation coefficient, v m is the mth spraying speed in the second spraying speed sequence, M is the number of elements in the second spraying speed sequence, m is a positive integer, and S2 is the second normalization coefficient; The formula for adjusting the predicted proportionality coefficient in S45 is: ,in, is the adjusted proportionality coefficient, K p is the predicted proportionality coefficient, γ c is the spraying speed variation coefficient, γ c,th is the spraying speed change threshold, e is a natural constant; The formula for adjusting the predicted integral coefficient in S45 is: ,in, is the adjusted integral coefficient, K i is the predicted integral coefficient; The formula for adjusting the predicted differential coefficient in S46 is: ,in, is the adjusted differential coefficient, K d is the predicted differential coefficient, γ s is the spraying speed oscillation coefficient, γ s,th is the spraying speed oscillation threshold, and e is a natural constant.

2. The artificial intelligence-based control method according to claim 1, characterized in that: The S1 comprises the following sub-steps: S11. Calculate a first distance coefficient according to the distance between the air cleanliness in the environment and the target air cleanliness; S12, calculating a second distance coefficient according to the distance between the humidity in the environment and the target humidity; S13. Calculate the target spraying speed according to the first distance coefficient and the second distance coefficient.

3. The artificial intelligence-based control method according to claim 2, characterized in that: The formula for calculating the target spraying speed in S13 is: , where v target is the target spraying speed, v0 is the initial spraying speed, α1 is the first weighting coefficient, α2 is the second weighting coefficient, max is the larger value, d1 is the first distance coefficient, and d2 is the second distance coefficient.

4. The artificial intelligence-based control method according to claim 1, characterized in that: The third-order PID model in S3 includes: a first PI sub-model, a second PI sub-model, a third PI sub-model, an adder A1 and a micro-molecule model; The input end of the first PI sub-model is connected to the input end of the second PI sub-model and the input end of the third PI sub-model respectively, and serves as the input end of the third-order PID model; The input end of the adder A1 is connected to the output end of the first PI sub-model, the output end of the second PI sub-model and the output end of the third PI sub-model respectively, and its output end is connected to the input end of the micro-molecule model; The output end of the micro-molecule model serves as the output end of the third-order PID model.

5. The artificial intelligence-based control method according to claim 4, characterized in that: The expression of the micromolecular model is: , , where U(t) is the control quantity output by the micro-molecular model at the tth moment, H(t) is the control quantity output by the adder A1 at the tth moment, u1(t) is the control quantity output by the first PI sub-model at the tth moment, u2(t) is the control quantity output by the second PI sub-model at the tth moment, u3(t) is the control quantity output by the third PI sub-model at the tth moment, e(t) is the speed difference between the target spraying speed at the tth moment and the real-time spraying speed, e(t-1) is the speed difference between the target spraying speed at the t-1th moment and the real-time spraying speed, and K d is the predicted differential coefficient.

6. An artificial intelligence-based disinfection robot, implemented based on the artificial intelligence-based control method according to any one of claims 1 to 5, characterized in that: include: a target spraying speed calculation unit, a prediction unit, a distribution unit, an adjustment unit and a control unit; The target spraying speed calculation unit is used to calculate the target spraying speed according to the air cleanliness and humidity in the environment; The prediction unit is used to predict the proportional coefficient, the integral coefficient and the differential coefficient according to the target spraying speed by using the BP neural network; The allocation unit is used to allocate the predicted proportional coefficient, integral coefficient and differential coefficient to the third-order PID model; The adjustment unit is used to adopt a third-order PID model for control, and adjust the predicted proportional coefficient, integral coefficient and differential coefficient according to the spraying speed variation coefficient and the spraying speed oscillation coefficient to obtain an adjusted third-order PID model; The control unit is used to control the spraying speed using an adjusted third-order PID model; The specific implementation process of the adjustment unit is as follows: a third-order PID model is used for control, and within the controlled T time, the real-time spraying speed is collected to construct a real-time spraying speed sequence; the real-time spraying speed sequence is divided into two parts to obtain a first-segment spraying speed sequence and a second-segment spraying speed sequence; according to the first-segment spraying speed sequence, a spraying speed variation coefficient is calculated; according to the second-segment spraying speed sequence, a spraying speed oscillation coefficient is calculated; according to the spraying speed variation coefficient, the predicted proportional coefficient and integral coefficient are adjusted; according to the spraying speed oscillation coefficient, the predicted differential coefficient is adjusted; the formula for calculating the spraying speed variation coefficient is: , where γ c is the spraying speed variation coefficient, v n is the nth spraying speed in the first spraying speed sequence, v n-1 is the n-1th spraying speed in the first spraying speed sequence, n is a positive integer, N is the number of elements in the first spraying speed sequence, | | is the absolute value, S1 is the first normalization coefficient; the formula for calculating the spraying speed oscillation coefficient is: , where γ s is the spraying velocity oscillation coefficient, v m is the mth spraying speed in the second spraying speed sequence, M is the number of elements in the second spraying speed sequence, m is a positive integer, S2 is the second normalization coefficient; the formula for adjusting the predicted proportional coefficient is: ,in, is the adjusted proportionality coefficient, K p is the predicted proportionality coefficient, γ c is the spraying speed variation coefficient, γ c,th is the spraying speed change threshold, e is a natural constant; the formula for adjusting the predicted integral coefficient is: ,in, is the adjusted integral coefficient, K i is the predicted integral coefficient; the formula for adjusting the predicted differential coefficient is: ,in, is the adjusted differential coefficient, K d is the predicted differential coefficient, γ s is the spraying speed oscillation coefficient, γ s,th is the spraying speed oscillation threshold, and e is a natural constant.

Images