Multi-frequency channel analysis and remote control system based on Internet of Things

Abstract

The invention provides a multi-frequency channel analysis and remote control system based on the Internet of Things, and relates to the technical field of the Internet of Things, multi-channel Wi-Fi modules are deployed at the front part and the rear part of target equipment, a dry ground reference model and 2.4 GHz, 5GHz and 6GHz three-frequency-band multi-stage analysis are combined, and channel information of an uncleaned area and a cleaned area is obtained in real time. If obvious attenuation or Doppler shift is detected, ground liquid abnormal information is generated and uploaded to the cloud; and the cloud performs classification analysis based on the three-frequency attenuation feature vector library, judges types of oil stains, water stains or foams and the like, and matches a cleaning strategy library to generate a control instruction. The rear multi-channel Wi-Fi acquisition can assess the post-cleaning effect and decide whether to clean for the second time, so that closed-loop management from detection, classification to re-cleaning is formed. Through layered detection and cloud cooperation, various kinds of liquid are effectively distinguished, cleaning parameters are dynamically adjusted, and the intelligent cleaning efficiency and quality are improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of Internet of Things, and in particular to a multi-frequency channel analysis and remote control system based on the Internet of Things. Background Art

[0002] At present, various industrial and commercial places (such as storage and logistics centers, supermarkets, workshops, parking lots, etc.) have increasingly higher requirements for the efficiency and quality of floor cleaning. Traditional ride-on floor scrubbers are usually manually controlled for driving and cleaning parameters, which is time-consuming and laborious to operate. In addition, this type of ride-on floor scrubber based on the driver's subjective experience is difficult to accurately judge the distribution of floor stains and the cleaning effect in a timely manner. In addition, there are also some ride-on floor scrubbers that default to the most comprehensive and maximum cleaning mode, but the resulting waste of cleaning resources and greatly shortened battery life are also problems to be solved. Summary of the invention

[0003] In view of the deficiencies of the prior art, the present application provides a multi-frequency channel analysis and remote control system based on the Internet of Things, including: Two sets of multi-channel Wi-Fi modules are respectively deployed above the front brush plate and the rear squeegee of the target device, for collecting channel information; the channel information is divided into front channel information and rear channel information, and includes channel information of the following frequency bands: 2.4 GHz, 5 GHz and 6 GHz; wherein the target device has an active driving function; An analysis module is used to perform multi-level analysis on the channel information using a preset dry ground reference model and generate a detection result; the detection result is uploaded to the cloud for matching analysis to generate ground liquid anomaly information; the multi-level analysis includes at least one of amplitude analysis, phase analysis and Doppler frequency shift analysis; A classification module, used to perform category analysis in the cloud using a preset three-frequency attenuation feature vector library based on the ground liquid anomaly information to generate liquid type information; The control module is used to generate control information for controlling the target device in the cloud based on the ground liquid abnormality information and the liquid type information, using a preset cleaning strategy library, and send the control information to the target device.

[0004] As an optional implementation manner, the collecting channel information includes: The front channel information obtained in the same sampling period is matched with the driving speed and steering angle of the target device, and marked as ground channel data obtained in the front uncleaned area; At the same time, the rear channel information is matched with the ground position that has been cleaned to evaluate the cleaning effect.

[0005] As an alternative implementation, the detection results corresponding to the multi-level analysis include at least one of: a first detection result corresponding to the 2.4 GHz frequency band, a second detection result corresponding to the 5 GHz frequency band, and a third detection result corresponding to the 6 GHz frequency band.

[0006] As an alternative implementation, the multi-level analysis includes: Performing a first comparative analysis on the channel information of the 2.4 GHz frequency band in the front channel information by using the dry ground reference model to generate a first detection result; In response to the first detection result having a first preset feature, enabling the acquisition of channel information of the 5 GHz frequency band in the multi-channel Wi-Fi module; Wherein, the first comparative analysis includes: amplitude difference; the first preset feature includes: the amplitude difference of consecutive preset frames is greater than a first preset threshold.

[0007] As an alternative implementation, the multi-level analysis further includes: Performing a second comparative analysis on the channel information of the 5 GHz frequency band in the front channel information by using the dry ground reference model to generate a second detection result; In response to the second detection result having a second preset feature, enabling the acquisition of channel information of the 6 GHz frequency band in the multi-channel Wi-Fi module; Wherein, the second comparative analysis includes: the fused difference of amplitude and phase; the second preset feature includes: the fused difference of consecutive preset frames is greater than a second preset threshold.

[0008] As an alternative implementation, the multi-level analysis further includes: Performing a third comparative analysis on the channel information of the 6 GHz frequency band in the front channel information by using the dry ground reference model to determine whether there is a third preset feature in the channel information and generate a third detection result; Wherein, the third comparative analysis includes: Doppler frequency shift analysis; the third preset feature includes: in the subcarrier information of the 6 GHz frequency band, it is detected that the Doppler frequency shift amount of consecutive preset values of sampling periods is greater than a third preset threshold.

[0009] As an alternative implementation, the classifying and analyzing the ground liquid anomaly information in the cloud by using a preset three-frequency attenuation feature vector library to generate liquid type information includes: Generating a multi-dimensional attenuation vector based on the ground liquid anomaly information; Comparatively analyzing the multi-dimensional attenuation vector with the preset three-frequency attenuation feature vector library to generate the liquid type information; Among them, the three-frequency attenuation feature vector library includes: reference fingerprints of clear water, oily liquid, foam-containing cleaner, and transparent film water stains; The comparison and analysis of the multi-dimensional attenuation vector with the preset three-frequency attenuation feature vector library includes: When the distance between the multi-dimensional attenuation vector and any reference fingerprint is less than the preset threshold, it is determined that the liquid type is consistent with the category corresponding to the reference fingerprint.

[0010] As an optional implementation manner, when generating control information for controlling the target device by using the preset cleaning strategy library, it further includes: The cloud determines whether to perform a cleaning effect evaluation based on the control information; In response to the cloud determining to perform a cleaning effect evaluation, the cleaning effect evaluation instruction and the control information are sent to the target device together; The evaluation of the cleaning effect includes: Based on the preset post-cleaning reference model, a comparative analysis is performed on the amplitude attenuation of the 2.4 GHz frequency band in the rear channel information.

[0011] As an optional implementation manner, it further includes: An IMU sensor is deployed on the body of the target device to collect three-axis acceleration and three-axis angular velocity data, and calculate the real-time inclination angle of the target device during operation through a fusion algorithm; Based on the real-time inclination angle, the slope grade is determined; The dry ground reference model further includes reference channel information at different slope grades; The first preset threshold, the second preset threshold, and the third preset threshold are dynamically adjusted based on the reference channel information at different slope grades.

[0012] Compared with the prior art, the present application can more accurately detect various oil stains, water stains and foam residues by deploying multi-channel Wi-Fi modules at the front and rear of the floor washer respectively and supporting the acquisition of channel information in three frequency bands of 2.4 GHz, 5 GHz and 6 GHz simultaneously. Secondly, the present invention adopts hierarchical matching of a "dry ground reference model" and a "post-cleaning reference model", which can detect the uncleaned and cleaned areas before and after operation respectively, and combine cloud big data to perform anomaly recognition and liquid type classification on the detection results; thus forming a closed loop from detection, classification to strategy issuance, and triggering cleaning effect evaluation or secondary cleaning operations as needed. Furthermore, by deploying an IMU sensor on the fuselage, the system can automatically adjust the thresholds of different frequency bands according to the slope level, enhance the stability on complex road surfaces such as slopes and depressions, and reduce false alarms caused by changes in the body attitude. Finally, the present invention adopts a "local + cloud" dual-layer architecture. On the one hand, it relies on the powerful storage and computing capabilities of the cloud to maintain a "three-frequency attenuation feature vector library", quickly complete liquid type recognition and flexibly update the fingerprint library; on the other hand, it enables the vehicle-mounted analysis module to focus on real-time detection and preliminary determination, significantly reducing the dependence on local computing power. Through the above comprehensive technical means, the present invention has achieved significant improvements in detection accuracy, liquid type recognition ability, scene adaptability, cleaning quality and energy consumption efficiency, and can better meet the high-standard automation operation requirements of large industrial or commercial sites for ride-on floor washers. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 FIG. is a schematic diagram of a multi-frequency channel analysis and remote control system based on the Internet of Things provided by an embodiment of the present application; Figure 2 FIG. is a flowchart of a method for collecting the channel information provided by an embodiment of the present application; Figure 3 FIG. is a schematic diagram of a method for generating liquid type information provided by an embodiment of the present application; Figure 4 FIG. is a schematic diagram of the use of channel information provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0015] See Figure 1 , Figure 1 FIG. is a schematic diagram of a multi-frequency channel analysis and remote control system based on the Internet of Things provided by an embodiment of the present disclosure, including: a multi-channel Wi-Fi module 10, an analysis module 20, a classification module 30 and a control module 40, wherein: Two groups of multi-channel Wi-Fi modules 10 are respectively deployed above the scrubbing brush near the front of the floor washer and above the water suction squeegee near the rear, for collecting channel information; the channel information is divided into front channel information and rear channel information, and includes channel information in the following frequency bands: 2.4 GHz, 5 GHz, and 6 GHz; wherein, the floor washer has an autonomous driving function; An analysis module 20, which is used to perform multi-level analysis on the channel information by using a preset dry ground reference model, and generate a detection result; upload the detection result to the cloud for matching analysis to generate ground liquid anomaly information; the multi-level analysis includes at least one of amplitude analysis, phase analysis, and Doppler frequency shift analysis; A classification module 30, which is used to perform category analysis on the ground liquid anomaly information in the cloud by using a preset three-frequency attenuation feature vector library to generate liquid type information; A control module 40, which is used to generate control information for controlling the floor washer in the cloud based on the ground liquid anomaly information and the liquid type information, and send the control information to the floor washer by using a preset cleaning strategy library.

[0016] In the system of the present application, by respectively deploying multi-channel Wi-Fi modules 10 at the front and rear of the floor washer, real-time channel information of the unwashed area in front and the washed area in the rear can be obtained. Through multi-level analysis of the three frequency bands of 2.4 GHz, 5 GHz, and 6 GHz, it can be initially determined whether there are liquids such as oil stains, water stains, and foam on the ground. When the analysis result exceeds the set threshold, the detection data is uploaded to the cloud, and with the help of a pre-established dry ground reference model and multi-band attenuation feature vectors, the ground liquid anomaly information is accurately matched. At the same time, based on the classification algorithm, the cloud can further identify the liquid type (such as clean water, oily liquid, foam-containing cleaner, etc.), and match the corresponding cleaning strategies (scrubbing brush pressure, water suction motor power, detergent dosage, etc.), and finally send them to the floor washer for execution. This process can also combine the rear channel information to evaluate the cleaning effect of the ground, forming a closed-loop control from detection, decision-making to verification.

[0017] Among them, the 2.4 GHz, 5 GHz, and 6 GHz frequency bands have different wavelengths, penetration, and reflection characteristics: The 2.4 GHz frequency band has a wavelength of approximately 12.5 cm, a wider coverage area, but is susceptible to environmental interference; the 5 GHz frequency band has its wavelength shortened to approximately 6 cm, with better spatial resolution and sensitivity to medium-thickness water stains or oil stains; while the 6 GHz frequency band has an even shorter wavelength, about 5 cm or less, and a bandwidth of up to 80 MHz, 160 MHz, or even higher, making it more advantageous in detecting the distribution of fine liquids such as thin water films and transparent oil stains. Physically, electromagnetic waves in each frequency band will experience varying degrees of absorption and scattering when encountering a liquid medium, and the specific attenuation amount is closely related to the dielectric constant and thickness of the liquid. To compare the amplitude and phase differences of the front or rear ground in dry and slippery states, this application first establishes a "dry ground reference model" by recording the reference amplitude, phase, and Doppler distribution of each frequency band in a dry environment. During actual operation, the real-time observation data is then compared with the reference value by taking the difference. If continuous amplitude attenuation, phase drift, or Doppler frequency shift exceeding the threshold is detected, it can be determined that there is a medium change (such as oil stains, water stains, or foam) at that location on the ground. Among them, 2.4 GHz is suitable for rough detection and large-scale discovery of potential liquid anomalies. 5 GHz is more refined than 2.4 GHz in terms of reflection and phase processing and can distinguish medium-thickness or obvious oil residues. 6 GHz combines subcarrier information analysis, and the Doppler rate and high bandwidth help capture more concealed situations such as extremely thin water films and transparent foam liquids. This differential physical characteristic provides a solid basis for multi-level detection and, in cooperation with the cloud matching algorithm, further improves the identification of the type and severity of ground liquids.

[0018] In specific implementation, the ride-on floor scrubber is equipped with a seat and a steering wheel. The operator can, while ensuring control over the vehicle's driving speed and steering, automatically collect ground information with the multi-channel Wi-Fi module provided by the present invention. Since this floor scrubber usually operates in enclosed or semi-enclosed places, to minimize communication interference and ensure the stable acquisition of multi-band information, corresponding wire harness slots and mounting fixtures are reserved inside the body to secure the 2.4 GHz, 5 GHz, and 6 GHz tri-band antennas and the waterproof housing, and to avoid resonance or electromagnetic coupling with the vehicle-mounted high-current drive wire harness. For environments with large slopes and vibrations, the sealing and anti-vibration of the connection ports between the antenna and the main control board can be further strengthened to prevent excessive signal fluctuations from affecting data reliability.

[0019] To adapt to the differences in ground operations at the front and rear, a set of multi-channel Wi-Fi modules is deployed at the front end (above the brush disk) and the rear end (above the water squeegee) of the floor washer. The front-end module obtains the 2.4GHz, 5GHz, and 6GHz channel data of the area yet to be cleaned. After preliminary multi-level analysis (amplitude, phase, or Doppler) of the channel data, if the amplitude or phase continuously exceeds the threshold, the relevant detection results will be sent to the cloud through the vehicle network to assist the cloud in generating "ground liquid anomaly information". The rear-end module captures the residual situation on the ground just after cleaning: if significant attenuation still occurs in the 2.4GHz amplitude analysis, it can prompt the operator of the floor washer to pay attention, or the cloud can send instructions such as secondary water suction or increasing the amount of detergent spraying to the control module to ensure that the final ground dryness and cleanliness meet the standards.

[0020] For the analysis module 20, an embedded system (such as a main control board with an ARM architecture) can be used to perform basic amplitude / phase threshold comparison locally to determine whether to upload a larger amount of CSI (Channel State Information) to the cloud. After the vehicle starts, the analysis module automatically loads the "dry ground reference model" and adapts to the current temperature and humidity conditions through a certain period of initial calibration (such as detecting the channel information of the blank area). Corresponding amplitude thresholds, phase drift thresholds, and Doppler frequency shift thresholds are configured for 2.4GHz, 5GHz, and 6GHz respectively. Only when the continuous number of frames exceeds the standard will it be determined that "detection result = abnormal tendency". In this way, during actual operation, the floor washer does not need to continuously send all the original data to the cloud, but only uploads more intensive data when the detection results reveal a greater possibility of abnormality. After receiving it, the cloud combines more historical databases and large-scale scenario statistical information to calculate and generate more accurate "ground liquid anomaly information". This "local combined with cloud" double-layer structure effectively reduces the dependence on the computing power of the vehicle-mounted terminal.

[0021] Regarding the classification module 30, since ride-on floor washers usually face multiple types of liquids (such as cooking oil, engine oil, antifreeze, foam-containing detergents, etc.), it is not convenient to maintain a large fingerprint feature library solely relying on the local area. In this embodiment, a "three-frequency attenuation feature vector library" is established in the cloud. When receiving the "ground liquid anomaly information", it will call a multi-dimensional similarity algorithm (such as Euclidean distance or cosine similarity) to compare with the fingerprint library, so as to quickly determine different categories such as "oily residue" or "clean water / foam", and mark the result as "liquid type information". This "liquid type information" is of great significance to the floor washer. For example, if it is determined to be oil stain, control information will be sent to make the machine body spray degreaser, increase the brush disk pressure or water suction force; if it is determined to be foam, more attention will be paid to the water suction operation and the driving speed will be reduced.

[0022] For the control module 40, it is also deployed in the cloud and connected to the classification module 30. It forms "control information" from the parameterized instructions (brush RPM, water absorption vacuum, vehicle speed limit, detergent dosing concentration, etc.) that the floor washer needs to execute through the "cleaning strategy library", and then sends it back to the floor washer side. For example, if the category analysis determines it as "thick oil stain", the system will specify a higher pressure and deceleration travel strategy to ensure that the machine can fully scrub and suck; if it is thin film foam, it will increase the vacuum suction and appropriately increase the water volume for cleaning. The human-machine interface built into the ride-on floor washer will display the instructions issued by the cloud in real time and prompt the driver with operation suggestions such as "slow down the vehicle speed" or "increase the turning radius" when necessary, improving safety and operation quality. Since it is a ride-on model, the driver can flexibly observe special areas during the cleaning process and make fine adjustments according to the detection results at the front, forming a closed-loop operation of "human-machine cooperation" with the cloud scheduling of the present invention.

[0023] In addition, the advantages of the "ride-on" model in this embodiment are also reflected in operation comfort and coverage efficiency: The driver only needs to concentrate on driving and obstacle avoidance, and the cleaning strategies (such as brush and water absorption adjustment) are automatically optimized by the multi-channel Wi-Fi detection + cloud intelligent analysis system of the present invention throughout the process. Whether the ground of the site is inclined (the slope can be corrected with an IMU sensor), the system can dynamically adjust the threshold and cleaning strategy to maintain excellent detection accuracy and floor washing effect. In this way, the advantages of combining the "large-scale operation and high manual mobility" of the ride-on floor washer with the "Internet of Things remote data processing" are fully utilized, greatly improving the cleaning operation efficiency and finished product quality under complex working conditions (such as engine oil stains, paint stains, foam liquid, etc.).

[0024] As an alternative embodiment, refer to Figure 2 , Figure 2 which is a flowchart of a method for collecting the channel information provided by the embodiment of the present application, including steps S101 to S102, where: S101: Corresponding the front channel information obtained within the same sampling period with the traveling speed and steering angle of the floor washer, and marking it as the ground channel data obtained in the unwashed area ahead; S102: At the same time, matching the rear channel information with the ground position that has been washed to evaluate the cleaning effect.

[0025] In order to more accurately mark the ground channel information collected by the front and rear multi-channel Wi-Fi modules, corresponding to the unwashed and washed areas respectively, the system also obtains motion parameters such as the traveling speed and steering angle on the ride-on floor washer, so as to perform position identification on the front channel information within the same sampling period.

[0026] Specifically, the main control board of the floor washer can obtain its motion state in the following ways: First, it uses on-vehicle speed sensors to collect the rotational speeds of the left and right wheels or the information of the speedometer in real time; second, it uses a steering wheel sensor or a steering angle encoder to capture the steering angle of the vehicle; third, on high-end models, it can calculate the current tilt angle and attitude in combination with an IMU (Inertial Measurement Unit). The above motion parameters and the 2.4GHz, 5GHz, and 6GHz channel information collected by the front multi-channel Wi-Fi module are all marked with a unified time stamp and registered in the local analysis module, so as to mark the ground position observed in this sampling period, namely the "uncleaned area ahead". In this way, when it is detected that the amplitude or phase difference continues to exceed the threshold, the system can identify which uncleaned ground area is suspected to have liquid.

[0027] In contrast, the 2.4GHz, 5GHz, and 6GHz channel information collected by the rear multi-channel Wi-Fi module is regarded as the corresponding data of the "cleaned ground position".

[0028] In this embodiment, to improve the matching accuracy, the cumulative displacement of the floor washer's driving speed and driving direction can be used, or the time difference between the rear and the front can be calculated by combining the known rudder angle and wheel diameter, so that the channel information obtained by the rear multi-channel Wi-Fi module can be automatically aligned to the area where the floor washer is currently located or has just cleaned. Once it is found in the rear information that the amplitude difference or Doppler analysis is still much higher than the dry reference value, it means that there are still obvious residual water stains or oil foam on the just-cleaned ground; based on this, the system can issue a secondary water suction or extended brushing instruction from the cloud or the local main control board. If the rear detection result shows that it has basically returned to normal, it indicates that the cleaning is relatively thorough.

[0029] Through such corresponding markings of "front channel information + driving speed / steering angle", the present invention can accurately locate the "uncleaned area being detected" at the same moment, and the automatic matching of "rear channel information + cleaned ground position" can evaluate the cleaning effect in real time, thus realizing the closed-loop management of the front and rear end information, and more flexible and high-precision detection and positioning can be obtained in large-area cleaning tasks or multi-curve site environments.

[0030] In a further embodiment of the system of the present invention, in order to enable the multi-channel Wi-Fi module to generate detection results corresponding to multi-level analysis for the channel information collected in the three frequency bands of 2.4 GHz, 5 GHz, and 6 GHz, the analysis module will distinguish the first detection result, the second detection result, and the third detection result at the vehicle-mounted end, so as to obtain a preliminary judgment of "abnormal ground liquid" by the subsequent cloud or local. Specifically, the main control board of the floor washer loads the dry ground reference model parameters of different frequency bands during initialization, and sequentially activates the detection channels of each frequency band according to the actual working environment. If only one or two frequency bands are used in this detection cycle, the analysis module can also select "at least one" frequency band for operation according to the preset strategy.

[0031] During the actual acquisition process, the following examples illustrate the generation methods of the three detection results: The first detection result (2.4 GHz frequency band): When the floor washer is driving normally, the front multi-channel Wi-Fi module defaults to preferentially performing amplitude differential comparison on the 2.4 GHz channel information. If the amplitude difference values of several consecutive frames exceed the first threshold and remain at a high level after filtering random jitters, the first detection result can be determined to be generated. At this stage, the analysis module only processes the basic data mainly based on amplitude or RSSI, and can achieve a rough detection of large-scale significant oil stains, water stains, etc. with a relatively low calculation amount.

[0032] The second detection result (5 GHz frequency band): When the analysis module captures that the first detection result indicates that there may be "mild or moderate abnormality", or the user enables the "medium detection" mode in the settings, the 5 GHz frequency band will be enabled for amplitude + phase fusion differential analysis. If the phase drift or the subcarrier amplitude continuously exceeds the second threshold in multiple frames, the second detection result is generated. This result is usually applicable to identifying medium-thickness oil stains or complex water stains. Since the 5 GHz wavelength is shorter and the environmental interference is relatively less, it can more accurately detect the fine pits or sticky residual liquids on the ground.

[0033] The third detection result (6 GHz frequency band): If the second detection result shows obvious over-standard or the system judges that higher-precision confirmation is required, the analysis module will enable the 6 GHz frequency band in the next detection cycle and perform in-depth comparison on the corresponding subcarrier amplitude, phase, and Doppler information. When the Doppler frequency shift value or the phase drift rate continuously exceeds the third threshold in multiple frames, it is determined that the third detection result is generated. This result usually corresponds to the situation of transparent oil stains with high dielectric constant or extremely thin water films, that is, the "precision inspection" stage. The analysis module can upload the third detection result to the cloud after stamping it with the time stamp and location information, so that the cloud can perform advanced fusion of subsequent ground liquid abnormality and type identification.

[0034] In this mode, the detection results "one, two, three" corresponding to the multi-level analysis do not necessarily all appear completely in the same cleaning operation: If the 2.4GHz rough detection is sufficient for certain scenarios, the system also allows only the first detection result to be output; If the situation is complex and 5GHz and 6GHz are required for further confirmation, the second and third detection results are generated in sequence. This maintains the flexibility for multiple frequency bands and the detection depth corresponding to different stains and different precision requirements. Through the IMU or speed sensor information described above, the system can also record the inclination angle and movement trajectory of the corresponding ground, enabling the three detection results to accurately match the ground conditions of a certain section or time period.

[0035] In a further embodiment of the present invention, in order to make full use of the advantages of the 2.4GHz frequency band in large-scale preliminary detection, in the multi-level analysis process of the analysis module of the system, the first comparative analysis is first performed on the 2.4GHz channel information collected by the front multi-channel Wi-Fi module. This analysis is mainly based on the amplitude difference principle: After the vehicle starts, the analysis module loads the 2.4GHz reference amplitude corresponding to the "dry ground reference model", and collects the current ground amplitude information in real time and performs a difference operation with this reference amplitude. If the difference result exceeds the first preset threshold for a continuous preset number of frames (such as 3 or 5 consecutive frames), it is determined that the first preset feature appears. In a typical scenario, the first preset threshold can be a certain ratio of the attenuation amount relative to the dry ground reference amplitude (such as more than 3dB or 5dB), and it is required to remain stably exceeded within several sampling periods to filter out occasional noise or short-term vibrations.

[0036] Once the analysis module detects the "first preset feature", it indicates that the 2.4GHz frequency band has sufficiently indicated the existence of obvious stains or water stain risks on the ground in this area. At this time, the system will automatically enable the collection of 5GHz channel information in the multi-channel Wi-Fi module according to the predefined strategy, thus entering the subsequent more refined detection stage. By this method of "first performing rough detection by amplitude difference and then enabling the high-frequency band after the first preset feature appears", the present invention can quickly detect potential abnormalities during the normal driving of the floor washer without having to maintain a high sampling rate for all frequency bands all the time, which can effectively reduce the load of local computing and network transmission.

[0037] In this process, the analysis module will also mark the time stamp and ground position for each 2.4GHz amplitude difference result (the position mark can be calculated from the vehicle speed and direction angle), so that if it is necessary to trace back the suspected stain area again later or to match the differential data of the higher frequency band with it. In some low-demand scenarios, if the 2.4GHz amplitude detection is sufficient, the system can also be set not to enter the 5GHz detection; However, in high-standard situations, once the first detection result indicates an abnormality, subsequent frequency bands must be enabled for further hierarchical analysis to improve the accuracy of liquid identification.

[0038] Among them, in the multi-level analysis process of the present invention, the reason for performing "first comparison analysis" only based on amplitude difference in the 2.4 GHz frequency band without deeply collecting or processing phase information is mainly based on the following considerations: On the one hand, the 2.4 GHz wavelength is relatively long (about 12.5 cm) and has a larger coverage range, and is often used for preliminary scanning or "coarse detection". At this time, the system pays more attention to the obvious amplitude attenuation caused by large areas or significant stains. On the other hand, the 2.4 GHz frequency band is affected by multi-source interference (such as Bluetooth, microwave ovens, etc.) and relatively complex multipath scattering in common enclosed places. If high-precision analysis of phase signals is performed, a large number of random fluctuations or phase mutations will occur, greatly increasing the filtering and stabilization processing of local algorithms and also increasing the risk of misjudgment. Moreover, for a typical ride-on floor washer in daily operation scenarios, if the 2.4 GHz can quickly detect significant attenuation (such as a large area of oil stains or thick water accumulation) in the form of amplitude difference, it is sufficient to prompt the system to enter the detection of higher-precision 5 GHz and 6 GHz frequency bands, rather than paying high computing power and transmission bandwidth at the primary stage to analyze the microscopic changes at the phase or sub-carrier level. In summary, adopting the relatively simple detection method of amplitude difference at 2.4 GHz is not only beneficial to reducing the real-time operation burden of the analysis module, but also can discover significant liquid anomalies at an early enough moment, meeting the hierarchical detection idea of "coarse detection first, then fine detection" of the present invention.

[0039] In a further embodiment of the present invention, in order to still not determine the specific conditions of ground stains or water stains after the 2.4 GHz coarse detection in the early stage, the system sets a second comparison analysis in the analysis module, mainly for the channel information of the 5 GHz frequency band.

[0040] Specifically, when the first detection result indicates the existence of a "first preset feature", the main control board actively enables the front multi-channel Wi-Fi module to collect 5 GHz frequency band data in the next sampling period and performs the fusion difference of amplitude and phase in combination with the dry ground reference model. At this stage, the analysis module not only pays attention to the amplitude attenuation amount, but also statistically analyzes the phase drift of each sub-carrier or the main carrier to obtain a more refined judgment.

[0041] In a relatively typical scenario, if the vehicle is passing through a moderately oil-contaminated or deep water accumulation area, although amplitude difference alone can detect anomalies, it is difficult to distinguish the viscosity or coverage thickness. At this time, through the "amplitude + phase" fusion difference in the 5GHz band, additional phase shift characteristics can be observed. If the phase drift continuously exceeds the second preset threshold in several frames of data, the analysis module determines that the "second detection result" is generated and records the occurrence of the "second preset feature". For example, the analysis module can be set to determine that the liquid viscosity or thickness is relatively significant when the 5GHz amplitude and phase fusion difference exceeds a certain value (such as several degrees to a dozen degrees) and lasts for M sampling periods.

[0042] Once the second preset feature is detected, the system believes that there are likely to be relatively complex forms such as medium-thickness oil stains and thick foams on the ground, and a higher-precision or shorter-wavelength detection method is required to finally confirm or distinguish "oil stains and transparent films", etc. Then the analysis module immediately enables the 6GHz band channel information collection of the multi-channel Wi-Fi module to prepare for the subsequent third detection result. Since 6GHz has more advantages in wavelength and bandwidth than 5GHz, it can further capture minute attenuation or Doppler frequency shift, thus making up for the deficiencies of the previous two bands in detecting extremely thin and extremely concealed liquids.

[0043] In this process, the "second preset feature" is equivalent to a threshold and consistency determination of the analysis module for 5GHz data: the comprehensive value of the amplitude difference and the phase difference needs to exceed the second preset threshold in a certain number of sampling frames to indicate that the anomaly in this area is significant enough to require in-depth detection. If the 5GHz analysis result does not continuously exceed the threshold, it means that the liquid anomaly may not be serious, and it may not be necessary to enter the 6GHz detection in a pre-set scenario, saving vehicle computing power and network bandwidth.

[0044] Through such a design, the present invention realizes "fusion difference of amplitude and phase" in the 5GHz band, that is, more accurately grasps medium-thickness or difficult-to-detect water stains and oil stains, and then controls whether to enter the 6GHz precise detection process through the "second preset feature", fully reflecting the hierarchical idea of multi-level detection and also ensuring the dynamic balance between performance and resource consumption.

[0045] Among them, in the multi-level analysis process of the present invention, the 5GHz band has advantages such as shorter wavelength, higher spatial resolution, and relatively less external interference compared to the 2.4GHz band. However, at the same time, the ground oil stains or water stains also show diversity in terms of thickness, viscosity, and surface morphology. Although certain information can be obtained based on amplitude attenuation alone, there may be missed or misjudged cases when encountering moderately viscous oil stains or local pits. Based on this, the present application adopts a fusion difference method of "combining amplitude difference and phase difference" in the 5GHz band to improve the overall detection fineness and anti-interference ability. The main considerations are as follows: On the one hand, for medium-thickness oil stains or large-area water stains on the ground, the amplitude attenuation characteristics are often obvious, which can quickly prompt potential pollution areas. On the other hand, if there are local depressions, thick liquids, viscous foams, etc. on the ground, the phase signal is more sensitive to the reflection and refraction behaviors of these irregular media, and can capture the wavefront phase offset more accurately, thus making up for the deficiency that amplitude analysis is vulnerable to random multipath or small attenuation disturbances and ignores tiny anomalies. Phase difference detection can perform cumulative or filtering judgments on the phase changes of the target section in several consecutive samples to exclude accidental interference and highlight real medium changes.

[0046] Therefore, in the 5GHz frequency band, if only based on amplitude difference, although most liquid coverage scenarios can be detected, it may cause recognition errors for edge distributions or semi-viscous liquids. Using only phase difference will increase the burden on hardware synchronization and data filtering, and may also miss the situation where the amplitude is obvious but the phase changes weakly. By fusing the two, when the phase offset and amplitude attenuation conditions are met, it can be determined that the "second preset feature" has reached sufficient confidence, and then trigger the higher-precision detection at 6GHz. And if one of the amplitude and phase is in a boundary state, the numerical reference provided by the other can also be used for "complementary" judgment to reduce missed detections or false detections caused by interfering noise. This combined analysis method of amplitude + phase difference is exactly the key to giving full play to the characteristics of "medium precision" and "moderate energy loss" in the 5GHz frequency band, enabling multi-level detection to obtain higher accuracy with a controllable amount of computation.

[0047] In a further embodiment of the present invention, when the analysis module still detects an obvious excess threshold in the second comparative analysis stage (5GHz frequency band) and cannot fully determine the type or thickness of the ground liquid, the system will enable the multi-channel Wi-Fi module to perform a third comparative analysis on the corresponding ground area in the 6GHz frequency band to generate a third detection result. In this stage, in addition to amplitude and phase, the present invention especially uses Doppler frequency shift analysis to improve the detection accuracy for thin-film liquids, transparent oil stains or foamy solutions.

[0048] Specifically, in the next sampling period, the main control board of the floor washer will deeply obtain the subcarrier channel information in the 6GHz frequency band, and compare it with the reference amplitude, phase, Doppler and other data at the 6GHz wavelength in the "dry ground reference model". If the system observes that the Doppler frequency shift amount in this section is always greater than the third preset threshold in a continuous preset number of sampling periods (for example, continuous N frames), it can be determined that the third preset feature appears, and the analysis module outputs the "third detection result". The present invention believes that the characteristics of short wavelength and high bandwidth in the 6GHz frequency band make it easier to capture minute fluctuations: when there is an extremely thin water film or semi-transparent oil stain on the ground moving relative to the floor washer, subtle changes will occur in the subcarrier phase and frequency, accumulating to form a significant Doppler frequency shift amount. If such changes continuously exceed the dry reference range in several periods, it indicates that there is a high dielectric constant medium on the ground and it is not easily captured by the rough amplitude / phase detection of the previous 2.4GHz or 5GHz.

[0049] In a typical scenario, assume that when the floor washer passes through a small area of transparent oil film, the 2.4GHz amplitude detection may not significantly exceed the threshold, and the 5GHz phase difference is not obvious either, but there will be a phenomenon of excessive accumulation of subcarrier frequency shift in the 6GHz Doppler analysis. The system then determines that "third detection result = abnormal". At this time, the analysis module will upload this result to the cloud. The cloud generates "ground liquid abnormal information" based on this, and through the classification module cooperating with the three-frequency attenuation vector library, it is confirmed that it may be a high dielectric oil film or a relatively special foam residue, and then the control module is ordered to allocate corresponding cleaning strategies (such as driving at a reduced speed, increasing the water absorption power, etc.).

[0050] It should be noted that the "third comparative analysis" does not necessarily trigger in all operations. Only when the "second detection result" in 5GHz shows that the "second preset feature" is significant, will the system activate the 6GHz frequency band to obtain more delicate data. Once the Doppler frequency shift amount of the 6GHz subcarrier continuously exceeds the "third preset threshold" multiple times within a multi-frame range, it indicates that the electromagnetic wave changes brought about by the thin film liquid or high dielectric solution are obvious enough. If this amount appears sporadically or does not continuously exceed the standard, the analysis module can determine that "the third preset feature is not reached", thus saving unnecessary deep detection and data upload costs.

[0051] Through the above design, the present invention uses Doppler frequency shift detection in the 6GHz frequency band to take into account amplitude / phase reference, effectively identify thin film oil stains or transparent foams without significantly increasing local computing power, and realize the final positioning and confirmation function of the aforementioned "coarse detection (2.4GHz) - medium detection (5GHz) - fine detection (6GHz)" hierarchical detection system, improving the detection accuracy and stability for extremely special liquid scenarios.

[0052] As an alternative implementation, see Figure 3, Figure 3 A schematic diagram of a method for generating liquid type information provided by an embodiment of the present application, including steps S201 to S202, where: S201: Generate a multi-dimensional attenuation vector based on the ground liquid anomaly information; S202: Compare and analyze the multi-dimensional attenuation vector with a preset three-frequency attenuation feature vector library to generate the liquid type information; Among them, the three-frequency attenuation feature vector library includes: reference fingerprints of clean water, oily liquid, foam-containing cleaner, and transparent film water stains; The comparing and analyzing the multi-dimensional attenuation vector with a preset three-frequency attenuation feature vector library includes: When the distance between the multi-dimensional attenuation vector and any reference fingerprint is less than a preset threshold, it is determined that the liquid type is consistent with the category corresponding to the reference fingerprint.

[0053] In a further embodiment of the present invention, in order to perform a higher-level type recognition on the ground where liquid anomalies have been determined to exist, the cloud will generate a multi-dimensional attenuation vector based on the "ground liquid anomaly information" and compare and analyze it with a preset "three-frequency attenuation feature vector library" to finally output the "liquid type information".

[0054] Specifically, after the system completes multi-level analysis and uploads the "ground liquid anomaly information", the cloud classification module will collect the differential results (including amplitude attenuation, phase shift value, Doppler frequency shift amount, etc.) and corresponding confidence levels from three frequency bands of 2.4 GHz, 5 GHz, and 6 GHz, and integrate this information into a multi-dimensional attenuation vector. In a practical example, this vector can be defined as in the form of, where represents the amplitude attenuation (or attenuation ratio), represents the phase shift statistical value, characterizes the accumulation or peak value of the Doppler frequency shift; independent components are provided respectively in the three frequency bands, making the vector dimension higher (such as the amplitude attenuation and phase results of 2.4 GHz, 5 GHz, and 6 GHz respectively, totaling 6 to 9 dimensions), and finally forming a comprehensive representation of the ground position.

[0055] In the "three-frequency attenuation feature vector library" of the present invention, the system pre-stores reference fingerprints of common liquid categories, such as "clean water", "oily liquid", "foam-containing cleaner", and "transparent film water stains", etc. Each fingerprint contains several statistical features for the possible value ranges of amplitude attenuation, phase shift, or Doppler frequency shift in the three frequency bands; it is also possible to cluster the multi-dimensional vectors of each liquid type based on experimental data or scenario measurement methods to obtain a more accurate fingerprint description.

[0056] Once the cloud generates the "multi-dimensional attenuation vector", it can be compared one by one with several reference vectors in the fingerprint database through similarity (or distance) calculation. For similarity measurement, Euclidean distance, cosine similarity or other common algorithms can be used; if the distance of a certain reference fingerprint is the smallest and lower than the preset threshold, the system will determine that the liquid type of the abnormal ground matches the corresponding fingerprint.

[0057] Exemplarily, if the multi-dimensional attenuation vector shows high amplitude attenuation, moderate phase drift, and low Doppler change at 2.4 GHz, 5 GHz, and 6 GHz, the cloud can compare and find that its distance from the fingerprint of "oily liquid" is the shortest (e.g., lower than the threshold of 0.15), and then determine it as oil stain; if the Doppler component is particularly significant, while the amplitude and phase are relatively medium, it may correspond to the fingerprint of transparent foam or high dielectric film, and more specifically output the liquid type "foam-containing detergent" or "transparent oil film". After completing this "comparative analysis", the cloud automatically marks the recognition result as "liquid type information" and sends it to the control module together with the "abnormal ground liquid information", so that the control module can schedule the corresponding cleaning plan from the cleaning strategy library (e.g., increasing suction, accelerating the injection agent, reducing the driving speed, etc.).

[0058] In this process, the "distance between the multi-dimensional attenuation vector and the reference fingerprint" being less than the preset threshold means that the ground liquid highly coincides with the reference category in terms of attenuation characteristics in each frequency band; if the minimum distance is still greater than the preset thresholds of all categories, the system can output "unknown liquid" or "mixed liquid", prompting manual intervention or supplementing the fingerprint database. Since the ride-on floor scrubber often faces various unknown stains, maintaining and expanding the fingerprint database in the cloud (such as self-learning based on historical big data) can significantly improve the accuracy and expandability of the present invention in liquid classification.

[0059] In a further implementation manner of the present invention, in order to better evaluate the cleaning effect in the cloud decision-making link and perform secondary cleaning or adjust the cleaning strategy when necessary, when the system generates control information using the preset cleaning strategy library, a "cleaning effect evaluation" determination link is also set.

[0060] In specific implementation, when the cloud determines operation parameters such as brush pressure and water absorption strength based on the abnormal ground liquid information, liquid type information, etc., it will first generate a basic cleaning instruction and send it to the floor scrubber; subsequently, the cloud will determine whether to verify the actual cleaning effect according to the historical records or scenario requirements (such as heavy oil stains, foam agent residues, etc.) in the cleaning strategy library. If the cloud determines to "evaluate the cleaning effect", it will attach a "cleaning effect evaluation instruction" to this control information and send it back to the main control system of the floor scrubber.

[0061] After the floor washer has completed the corresponding cleaning operations (such as increasing the brush pressure, slowing down the vehicle speed, increasing the suction, etc.), it will re-analyze the channel information collected by the rear multi-channel Wi-Fi module according to the effect evaluation instructions attached by the cloud. For this purpose, the system has established a "post-cleaning reference model" in this embodiment, which is slightly different from the "dry ground reference model": the former can accommodate the ground state with mild moisture or trace detergent residue as the "qualified" standard. For example, in this post-cleaning reference model, the amplitude attenuation reference value in the 2.4 GHz frequency band may only be allowed to fluctuate within a threshold of 2 - 3 dB to indicate that the ground is still in a "relatively dry and safe" condition.

[0062] After the rear multi-channel Wi-Fi module of the floor washer collects data in the 2.4 GHz frequency band, the local analysis module will perform a comparative analysis with the "post-cleaning reference model": if the amplitude attenuation is much lower than the reference value, it means the ground is too wet and there is obvious residue; if the relative value continuously exceeds a certain evaluation threshold (such as 3 dB) and maintains for several sampling periods, it can be determined that the "cleaning effect is insufficient". The system will generate an "abnormal rear cleaning evaluation" message at this time and report it to the cloud. After receiving this anomaly, the cloud will call the cleaning policy library again to issue commands such as "secondary water suction" or "repeated brushing"; on the contrary, if the amplitude attenuation basically matches the post-cleaning reference model, it means the "cleaning result is qualified" and no additional remedial measures are required.

[0063] Through this process of "the cloud determines whether to conduct a cleaning effect evaluation - issues an evaluation instruction - the local uses the post-cleaning reference model to compare and analyze the rear 2.4 GHz amplitude attenuation - decides on subsequent operations", without increasing redundant detections at the floor washer end, the present invention can flexibly decide whether to conduct an effect verification according to the operation scenario, and quickly feedback in combination with the rear information, forming an integrated closed-loop from cloud policy generation to on-site effect evaluation and then to secondary correction, greatly improving the cleaning quality and resource utilization efficiency.

[0064] Exemplarily, please refer to Figure 4 FIG. [FIG ID], which is a schematic diagram of the use of channel information provided by an embodiment of the present application. The channel information is divided into front channel information and rear channel information. The front channel information is used for the detection results in each frequency band generated by multi-level detection through multi-level analysis based on amplitude, phase, and Doppler frequency shift, while the rear channel information is used for cleaning effect evaluation. In an optional embodiment, a preliminary evaluation in the 2.4 GHz frequency band is sufficient.

[0065] In a further embodiment of the present invention, in order to improve the detection accuracy for inclined ground or uneven road sections, the system further includes an IMU sensor (Inertial Measurement Unit), which is deployed on the body of the floor washer and is used to collect triaxial acceleration and triaxial angular velocity data. The IMU data calculates the real-time inclination angle of the floor washer during operation through a fusion algorithm (such as complementary filtering or Kalman filtering), enabling the analysis module to more accurately identify the changes in Wi-Fi channel information obtained by the body at different slope levels.

[0066] In a specific implementation, when the floor washer moves to a ground area with a slope, the IMU sensor will detect a corresponding offset in the roll angle or pitch angle. If the offset exceeds a preset threshold (such as 5°, 10°, etc.), the analysis module will consider this information together with the vehicle's driving speed or direction angle, and then determine that the current ground is slope level X (for example, 0° - 5° is regarded as a mild slope, 5° - 10° is regarded as a moderate slope, and above 10° is a steeper grade). On this basis, the system reserves a set of correction factors or additional reference data in the "dry ground reference model" for each slope level, indicating that the Wi-Fi interference and attenuation laws of the same ground may vary under different inclinations. For example, when facing a steeper inclination angle, the influence of the body on the signal reflection path will increase, and additional offsets in Doppler or phase are also likely to occur.

[0067] Specifically, if the IMU determines that the current slope level is high, the analysis module will call the "reference channel information at the corresponding slope level" for differential comparison in multi-level analysis. Correspondingly, the first preset threshold (mainly for 2.4GHz amplitude difference), the second preset threshold (mainly for 5GHz amplitude + phase fusion difference), and the third preset threshold (mainly for 6GHz Doppler analysis) can all be relaxed or tightened by a certain proportion according to the slope level.

[0068] Exemplarily, if the slope is large, there are naturally certain deviations in the amplitude, phase, or Doppler data. To prevent misjudgment, the system can slightly increase the threshold to make the "abnormal" determination more confident; conversely, if the slope is small or the ground is relatively flat, the threshold can be tightened to more sensitively capture small amplitude attenuation.

[0069] In a typical application scenario, when the floor washer drives from a flat area onto an inclined ground with an inclination of about 6°, the IMU sensor detects the change in the pitch angle and updates the "current slope level = moderate slope" in the analysis module. Subsequently, the corresponding reference amplitude and phase are selected or interpolated in the dry ground reference model, and the first preset threshold (such as the base value of 3 dB) is increased by 0.5 dB to adapt to the additional interference brought by the slope environment. If the attenuation amount detected at 2.4 GHz exceeds this dynamic threshold thereafter, the system will still determine that "the first detection result = abnormal tendency" and enter the next frequency band or subsequent analysis. Through this "reference channel information under the slope level" and "dynamic adjustment of the threshold", the present invention can not only reduce the error caused by the fuselage attitude at the slope, but also maintain a more stable detection accuracy under different terrain conditions.

[0070] Meanwhile, when receiving the detection result after IMU compensation from the cloud control module, different cleaning strategies can also be selected according to the slope level (such as reducing the driving speed and increasing the power of the water suction unit at a steeper slope to avoid the accumulation of liquid sliding down, etc.), further improving the cleaning safety and quality. Through the combination of the IMU sensor and dynamic threshold optimization, the present invention realizes the adaptation to various complex terrains and expands the reliability and practical value of the intelligent detection of the ride-on floor washer under various working conditions.

[0071] Those skilled in the art can understand that in the above method of the specific implementation manner, the writing order of each step does not mean a strict execution order that constitutes any limitation to the implementation process. The specific execution order of each step should be determined according to its function and possible internal logic. It should be understood that determining B according to A does not mean determining B only according to A, but also B can be determined according to A and / or other information.

[0072] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present invention can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0073] In the description of this specification, the description referring to terms such as "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

Claims

1. Multi-frequency channel analysis and remote control system based on the Internet of Things, characterized in that: include: Two sets of multi-channel Wi-Fi modules are respectively deployed above the front brush plate and the rear squeegee of the target device, for collecting channel information; the channel information is divided into front channel information and rear channel information, and includes channel information of the following frequency bands: 2.4 GHz, 5 GHz and 6 GHz; wherein the target device has an active driving function; An analysis module is used to perform multi-level analysis on the channel information using a preset dry ground reference model and generate a detection result; the detection result is uploaded to the cloud for matching analysis to generate ground liquid anomaly information; the multi-level analysis includes at least one of amplitude analysis, phase analysis and Doppler frequency shift analysis; A classification module, for performing category analysis in the cloud using a preset three-frequency attenuation feature vector library based on the ground liquid anomaly information to generate liquid type information; The control module is used to generate control information for controlling the target device in the cloud based on the ground liquid abnormality information and the liquid type information, using a preset cleaning strategy library, and send the control information to the target device.

2. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 1 is characterized in that: The acquisition channel information includes: The front channel information obtained in the same sampling period is matched with the driving speed and steering angle of the target device, and marked as ground channel data obtained in the front uncleaned area; At the same time, the rear channel information is matched with the ground position that has been cleaned to evaluate the cleaning effect.

3. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 2 is characterized in that: The detection results corresponding to the multi-level analysis include: at least one of: a first detection result corresponding to the 2.4 GHz frequency band, a second detection result corresponding to the 5 GHz frequency band, and a third detection result corresponding to the 6 GHz frequency band.

4. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 3 is characterized in that: The multi-level analysis includes: For the channel information of the 2.4 GHz frequency band in the front channel information, using the dry ground reference model, a first comparative analysis is performed to generate a first detection result; In response to the first detection result having a first preset feature, enabling 5 GHz frequency band channel information collection in the multi-channel Wi-Fi module; Among them, the first comparative analysis includes: amplitude difference; the first preset feature includes: the amplitude difference of consecutive preset frames is greater than a first preset threshold.

5. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 4 is characterized in that: The multi-level analysis also includes: Based on the channel information of the 5 GHz frequency band in the front channel information, a second comparative analysis is performed using the dry ground reference model to generate a second detection result; In response to the second detection result having a second preset feature, enabling 6 GHz frequency band channel information collection in the multi-channel Wi-Fi module; Among them, the second comparative analysis includes: fusion difference of amplitude and phase; the second preset feature includes: fusion difference of consecutive preset frames is greater than a second preset threshold.

6. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 5, characterized in that: The multi-level analysis also includes: Based on the channel information of the 6 GHz frequency band in the front channel information, a third comparative analysis is performed using the dry ground reference model to determine whether a third preset feature exists in the channel information, and a third detection result is generated; The third comparative analysis includes: Doppler frequency shift analysis; the third preset feature includes: in the subcarrier information of the 6 GHz frequency band, it is detected that the Doppler frequency shift amount of a continuous preset value of sampling periods is greater than a third preset threshold.

7. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 1, characterized in that: Based on the ground liquid anomaly information, the preset three-frequency attenuation feature vector library is used to perform category analysis in the cloud to generate liquid type information, including: Based on the ground liquid anomaly information, generating a multi-dimensional attenuation vector; Comparing and analyzing the multi-dimensional attenuation vector with a preset three-frequency attenuation characteristic vector library to generate the liquid type information; The three-frequency attenuation feature vector library includes: reference fingerprints of clean water, oily liquid, foam-containing detergent and transparent film water stains; The comparing and analyzing the multi-dimensional attenuation vector with a preset three-frequency attenuation feature vector library includes: When the distance between the multi-dimensional attenuation vector and any reference fingerprint is less than a preset threshold, it is determined that the liquid type is consistent with the category corresponding to the reference fingerprint.

8. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 2, characterized in that: The use of the preset cleaning strategy library to generate control information for controlling the target device also includes: The cloud determines whether to perform a cleaning effect evaluation based on the control information; In response to the cloud determining to perform a cleaning effect evaluation, sending a cleaning effect evaluation instruction together with the control information to the target device; The evaluation of cleaning effect includes: Based on a preset post-cleaning benchmark model, a comparative analysis is performed on the amplitude attenuation of the 2.4 GHz frequency band in the rear channel information.

9. The multi-frequency channel analysis and remote control system based on the Internet of Things according to claim 6, characterized in that: Also includes: An IMU sensor is deployed on the body of the target device, and is used to collect three-axis acceleration and three-axis angular velocity data, and calculate the real-time inclination angle of the target device during operation through a fusion algorithm; Based on the real-time inclination angle, determining a slope grade; The dry ground reference model also includes reference channel information at different slope levels; The first preset threshold, the second preset threshold and the third preset threshold are dynamically adjusted based on the reference channel information at the different slope levels.

Images