A wiper control method and vehicle

Abstract

The present application relates to the technical field of vehicles, and discloses a wiper control method and a vehicle, the method comprising: acquiring rainfall characteristics collected by each detection unit in a rainfall detection unit array, wherein the rainfall detection unit array is arranged on the vehicle, and the rainfall detection unit array comprises a main detection unit and at least one slave detection unit; performing validity detection on the rainfall characteristics collected by each detection unit to obtain valid rainfall characteristics corresponding to each detection unit; fusing the valid rainfall characteristics corresponding to each detection unit to obtain fused rainfall characteristics; and controlling the wiper of the vehicle to perform a wiping action corresponding to the fused rainfall characteristics. The present application solves the problem of misjudgment of rainfall signals and inaccurate wiper control caused by local shielding, abnormal installation or failure of the existing rainfall sensor.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, specifically to a windshield wiper control method and a vehicle. Background Technology

[0002] Rain sensors are the core sensing component of automotive intelligent wiper systems, and their detection accuracy directly affects the accuracy of wiper control and driving safety. Currently, automotive rain sensors are mainly divided into two categories: one is the wired fixed installation type, which is fixed to the area below the rearview mirror inside the windshield by adhesive or mechanical means, requiring the removal of the interior trim and professional wiring, making installation and maintenance cumbersome; the other is the single-node wireless type, which simplifies wiring, but is limited by the single-point installation location and cannot cover the sides of the windshield, side windows, and rear window areas, and is prone to detection blind spots due to local obstructions (such as leaves, snow, and stains).

[0003] However, existing rain sensors cannot identify abnormal conditions when the sensor is partially blocked or slightly loosened, and directly output incorrect rainfall signals, causing wipers to malfunction. Simply stacking multiple sensors makes it impossible for each sensor to distinguish between invalid signals caused by installation gap interference, partial obstruction, or its own malfunction and the real rainfall signal, resulting in unreliable fusion results and poor wiper control accuracy. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a windshield wiper control method and a vehicle to solve the problem of misjudgment of rain signals and inaccurate windshield wiper control caused by existing rain sensors due to partial obstruction, abnormal installation, or malfunction.

[0005] In a first aspect, embodiments of the present invention provide a windshield wiper control method, the method comprising: The rainfall characteristics collected by each detection unit in the rainfall detection unit array are acquired, wherein the rainfall detection unit array is deployed on a vehicle and includes a main detection unit and at least one slave detection unit; The validity of the rainfall characteristics collected by each detection unit is tested to obtain the valid rainfall characteristics corresponding to each detection unit; The effective rainfall features corresponding to each detection unit are fused to obtain the fused rainfall features; The vehicle's windshield wipers are controlled to perform the wiping action corresponding to the fused rainfall characteristics.

[0006] Furthermore, the acquisition of rainfall features collected by each detection unit in the rainfall detection unit array includes: Acquire the raw rainfall signal collected by each of the detection units; Obtain the compensation coefficient corresponding to each detection unit, wherein the compensation coefficient is used to compensate the original rainfall signal collected by the corresponding detection unit in order to eliminate the baseline drift of each detection unit during the acquisition process; For each detection unit, the original rainfall signal is compensated according to the compensation coefficient to obtain a compensated rainfall signal, and the corresponding rainfall features are extracted from the compensated rainfall signal.

[0007] Furthermore, obtaining the compensation coefficient corresponding to each of the detection units includes: Obtain the deployment status of the main detection unit and the at least one slave detection unit; When all the deployment states are in the deployed state, the main detection unit broadcasts a networking request and establishes a target communication network with the at least one slave detection unit that responds to the networking request. The main detection unit controls the target communication network to enter a calibration state. The reference signals collected by the main detection unit and the at least one slave detection unit in the calibration state are acquired, and the compensation coefficient of each detection unit is calculated based on the reference signals.

[0008] Furthermore, the step of performing validity testing on the rainfall characteristics collected by each detection unit to obtain the valid rainfall characteristics corresponding to each detection unit includes: Identify the abnormal factors that affect the effectiveness of rainfall detection by the detection unit, and obtain at least one judgment index corresponding to each of the abnormal factors; Based on the at least one determination index, abnormal features in the rainfall characteristics are identified, and the abnormal features are removed from the rainfall characteristics to obtain the effective rainfall characteristics.

[0009] Furthermore, the abnormal factors include: gap abnormalities, obstruction abnormalities, and malfunction abnormalities; The judgment indicators include the signal variance of the rainfall signal, the signal peak value of the rainfall signal, and the communication status of the detection unit. The signal variance is the judgment indicator corresponding to the gap anomaly type, the signal peak value is the judgment indicator corresponding to the occlusion anomaly type, and the communication status is the judgment indicator corresponding to the fault anomaly type.

[0010] Furthermore, the step of identifying abnormal features in the rainfall characteristics based on the at least one determination index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics, includes: When the abnormal factor is an intermittent abnormality, the average variance is calculated based on the signal variance in each rainfall feature to obtain the baseline variance; Identify the first anomalous feature in the rainfall characteristics whose signal variance is greater than the baseline variance, and remove the first anomalous feature from the rainfall characteristics to obtain the effective rainfall characteristics.

[0011] Furthermore, the step of identifying abnormal features in the rainfall characteristics based on the at least one determination index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics, includes: When the abnormal factor is an obstruction abnormality, a second abnormal feature in the rainfall characteristics is identified whose signal peak value is a first preset value; For detection units other than those corresponding to candidate anomaly features, calculate the average value of the signal peak value in their corresponding rainfall features to obtain the baseline peak value; Detect whether the reference peak value is greater than or equal to the second preset value; When the baseline peak value is greater than or equal to the second preset value, the second abnormal feature is removed from the rainfall feature to obtain the effective rainfall feature.

[0012] Furthermore, the step of identifying abnormal features in the rainfall characteristics based on the at least one determination index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics, includes: When the abnormal factor is a fault, a third abnormal feature is identified in the rainfall feature that has data missing or data exceeding the limit for a continuous preset number of cycles. The third abnormal feature is removed from the rainfall feature to obtain the effective rainfall feature.

[0013] Furthermore, the step of identifying abnormal features in the rainfall characteristics based on the at least one determination index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics, includes: When the abnormal factors include gap abnormality type, shading abnormality type and fault abnormality type, the average variance is calculated based on the signal variance in each rainfall feature to obtain the baseline variance; Identify the first anomalous feature in the rainfall features whose signal variance is greater than the benchmark variance, and remove the first anomalous feature from the rainfall features to obtain the first remaining feature; Identify a second abnormal feature among the first remaining features whose signal peak value is a first preset value; For detection units other than those corresponding to candidate abnormal features, calculate the average value of the signal peaks in the corresponding first remaining features to obtain the baseline peak value; Detect whether the reference peak value is greater than or equal to the second preset value; When the benchmark peak value is greater than or equal to the second preset value, the second abnormal feature is removed from the first remaining feature to obtain the second remaining feature; Identify a third abnormal feature in the second remaining features that has data missing or data exceeding limits for a continuous preset number of periods, and remove the third abnormal feature from the second remaining features to obtain the effective rainfall feature.

[0014] Furthermore, each detection unit in the rain detection unit array is magnetically attached to the vehicle via a magnetic structure. The magnetic structure includes a magnetic base, a magnetic circuit assembly, a sealing assembly, and a buffer pad. The magnetic circuit assembly is disposed inside the magnetic base and is used to adjust the magnetic attraction force. The sealing assembly is disposed at the edge of the magnetic base and is used to form a seal with the vehicle. The buffer pad is disposed on the magnetic base and is used to conform to the vehicle to form a buffer.

[0015] In a second aspect, embodiments of the present invention provide a windshield wiper control device, the device comprising: An acquisition module is used to acquire rainfall characteristics collected by each detection unit in the rainfall detection unit array, wherein the rainfall detection unit array is deployed on a vehicle and includes a main detection unit and at least one slave detection unit. The detection module is used to detect the validity of the rainfall characteristics collected by each detection unit, and obtain the valid rainfall characteristics corresponding to each detection unit; The fusion module is used to fuse the effective rainfall features corresponding to each of the detection units to obtain fused rainfall features; The control module is used to control the vehicle's windshield wipers to perform the wiping action corresponding to the fused rainfall characteristics.

[0016] Thirdly, embodiments of the present invention provide a vehicle, including: a controller and a windshield wiper, the controller including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method described in the first aspect or any corresponding embodiment thereof.

[0017] Fourthly, embodiments of the present invention provide a computer device, including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method described in the first aspect or any corresponding embodiment thereof.

[0018] Fifthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions for causing a computer to perform the method described in the first aspect or any corresponding embodiment thereof.

[0019] The method provided in this application has the following beneficial effects: The method provided in this application acquires rainfall features from a rainfall detection unit array consisting of a main detection unit and at least one slave detection unit deployed on a vehicle. This achieves full-view rainfall information coverage across multiple areas of the vehicle, avoiding blind spots caused by single-point installation. By validating the rainfall features acquired by each detection unit, invalid data caused by anomalies such as adsorption gaps, partial obstruction, and node failures can be identified and eliminated, ensuring the authenticity and reliability of the information used for subsequent fusion and significantly reducing the risk of misjudgment. By fusing the valid rainfall features corresponding to each detection unit and integrating the rainfall information reflected by each valid rainfall feature, the fused rainfall features more accurately reflect the actual rainfall intensity and distribution. Finally, the method controls the vehicle's windshield wipers to perform wiping actions corresponding to the fused rainfall features, achieving precise adaptive speed adjustment from intermittent, low speed, medium speed to high speed, avoiding false triggering or delayed response of the wipers, and significantly improving the safety and comfort of driving in rainy weather. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating a windshield wiper control method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the star-shaped networking and communication architecture of the rainfall detection unit array according to an embodiment of the present invention; Figure 3 This is a flowchart illustrating another windshield wiper control method according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the validity detection and weighted fusion process of rainfall features according to an embodiment of the present invention; Figure 5 This is a flowchart illustrating another windshield wiper control method according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the magnetic flexible sealing adsorption layered structure of the detection unit according to an embodiment of the present invention; Figure 7 This is a structural block diagram of a windshield wiper control device according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] According to embodiments of the present invention, a windshield wiper control method and a vehicle are provided. It should be noted that the steps shown in the flowcharts in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0024] This embodiment provides a windshield wiper control method. Figure 1 This is a flowchart of a windshield wiper control method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain rainfall characteristics collected by each detection unit in the rainfall detection unit array, wherein the rainfall detection unit array is deployed on the vehicle and includes a main detection unit and at least one slave detection unit.

[0025] It should be noted that the installation position of the rain detection unit array can be flexibly configured according to the coverage requirements of the vehicle's windshield, forming various geometric layouts: when the main detection unit is attached to the core area of ​​the inner surface of the windshield (below the rearview mirror), and two secondary detection units are attached to the inner surfaces of the left and right sides of the windshield respectively, the three detection units form a triangular array covering the windshield's field of vision; when two additional secondary detection units are added to the sides of the windshield and attached to the inner surfaces of the left and right windshields respectively, the four detection units form a square or rectangular array, achieving cross-coverage of the windshield and lateral areas; if the number of secondary detection units is further increased to five or more, for example, deployed on the inner surfaces of the windshield, side windows, and rear window respectively, a rhomboid or polygonal array is formed, in which the main detection unit is located in the core position of the windshield, and the remaining secondary detection units are distributed around it, thereby eliminating rain detection blind spots and achieving segmented monitoring of the vehicle's entire windshield.

[0026] In addition, a two-point linear layout can be adopted, which involves deploying only one main detection unit and one secondary detection unit. For example, the main detection unit is attached to the core area of ​​the windshield, and one secondary detection unit is attached to the inner surface of the rear windshield, forming a linear layout covering both the front and rear windshields. This is suitable for small vehicles that are not sensitive to lateral rain. Alternatively, a four-point asymmetrical layout can be adopted, deploying the main detection unit and three secondary detection units, located on the left side of the windshield, the right side of the windshield, and the rear windshield, respectively, while omitting them on the side windows. This forms an irregular quadrilateral layout covering both the windshield and the rear windshield, taking into account both front and rear visibility. Alternatively, the ability of the main node to connect up to eight auxiliary nodes can be utilized to deploy the main detection unit and seven to eight secondary detection units, attached to the middle of the windshield, both sides of the windshield, the left and right side windows, the rear windshield, and the inner surface of the sunroof glass, forming a star-shaped surrounding layout centered on the main detection unit, achieving 360-degree full glass coverage. A linear array layout can also be used, where one main detection unit and multiple secondary detection units are deployed horizontally along the upper or lower edge of the vehicle's windshield, forming a linear array arranged in a line, specifically for accurately identifying the distribution density of raindrops in the horizontal direction. All of the above layouts support a star topology with a main detection unit and up to eight slave detection units, meeting the distributed rainfall detection needs of different vehicle models and user scenarios.

[0027] Figure 2 This is a schematic diagram of the star-shaped networking and communication architecture of the rain detection unit array in this embodiment of the invention. The architecture includes: a main detection unit 1 deployed below the rearview mirror inside the windshield as the central node, which establishes a star topology network with secondary detection units 2 and 3 respectively arranged on both sides of the windshield or secondary detection units 3 and 4 in the side rearview mirror area via a Bluetooth 5.0 BLE wireless communication link. The main detection unit is responsible for data aggregation, fusion decision-making and windshield wiper control command issuance, while the secondary detection units are only responsible for local rain data collection and encrypted uploading, realizing wireless collaboration of distributed rain detection. At the same time, the WiFi icon indicates the communication link between the main detection unit and the vehicle domain controller / vehicle system, ensuring the transmission of control commands and status feedback.

[0028] In this embodiment of the application, obtaining the rainfall characteristics collected by each detection unit in the rainfall detection unit array includes: Step A1: Obtain the raw rainfall signal collected by each detection unit.

[0029] Specifically, a detection unit refers to a sensor node deployed on the inner surface of a vehicle's windshield for detecting rainfall. It includes a main detection unit and at least one slave detection unit. The main detection unit is responsible for network control and data fusion decisions, while the slave detection units are responsible for assisting in the collection of rainfall data. The raw rainfall signal refers to the uncompensated voltage signal directly collected by the dual-channel differential rainfall detection unit within each detection unit. This includes the rainfall signal channel voltage Vs (the voltage obtained by emitting 940nm infrared light from the infrared emitting component, which is then scattered by raindrops and received and converted by the signal receiving channel) and the reference channel voltage Vr (the voltage obtained by receiving ambient light and common-mode interference signals that are not affected by raindrops and converted by the reference receiving channel). These two signals together characterize the intensity of raindrop obstruction and the level of environmental interference at the location of the detection unit.

[0030] The main detection unit and each slave detection unit independently acquire raw rainfall signals: First, the infrared emitting component of each detection unit illuminates a 940nm infrared LED with constant current drive (50mA~150mA), and emits detection light onto the outer surface of the windshield after being focused by a collimating lens; Second, the signal receiving channel receives the light signal returned after being scattered by raindrops at a sampling frequency of ≥10Hz, and converts it into a rainfall signal channel voltage Vs. At the same time, the reference receiving channel receives the background light signal that is not modulated by raindrops at the same sampling frequency, and converts it into a reference channel voltage Vr; Then, each detection unit temporarily stores the acquired Vs and Vr as raw rainfall signals in its local cache. For the main detection unit, its main control unit directly reads its own raw rainfall signal through an internal analog-to-digital converter; for each slave detection unit, the raw rainfall signal it collects is uploaded to the main detection unit's Bluetooth 5.0 BLE main communication unit via a Bluetooth 5.0 BLE slave communication unit at a period of 100ms and encrypted with AES-128. After receiving the data, the main detection unit aligns the raw rainfall signals of all detection units according to the node ID and timestamp, thereby completing the acquisition of the raw rainfall signal for the entire array.

[0031] Step A2: Obtain the compensation coefficient corresponding to each detection unit. The compensation coefficient is used to compensate the original rainfall signal collected by the corresponding detection unit to eliminate the baseline drift of each detection unit during the acquisition process.

[0032] In this embodiment of the application, obtaining the compensation coefficient corresponding to each detection unit includes: Step A201: Obtain the deployment status of the main detection unit and at least one slave detection unit.

[0033] Specifically, deployment status refers to the information set indicating whether each detection unit has completed physical installation and is ready for normal operation, including at least one of the following: adsorption status (whether the detection unit is firmly adsorbed onto the inner surface of the vehicle's windshield via a magnetic structure and the gap between the annular sealing lip and the glass is ≤0.1mm), power status (whether the battery power is sufficient and within the normal power supply range), and hardware self-test status (whether the infrared LED, detector, and communication module are working properly). The main detection unit is the core node in the rain detection unit array, responsible for initiating network formation, coordinating calibration, fusing data, and generating wiper control commands. The slave detection units are auxiliary nodes in the array, responsible for collecting rainfall signals at their location and reporting them to the main detection unit, without making independent control decisions.

[0034] The main detection unit obtains its own and all slave detection units' deployment status through a combination of internal self-testing and external queries. First, the main detection unit executes an internal self-test procedure: it reads the battery voltage and remaining charge (SOC) output by its own power management unit, obtains the current magnetic attraction force value (which should be within the adjustable range of 5N~25N) through the detection attraction force feedback circuit, and verifies whether the working registers of the dual-channel differential rain detection unit, Bluetooth communication unit, and main control unit are normal. If all self-test items pass, the main detection unit's own deployment status is marked as "deployed"; otherwise, it is marked as "abnormal" and a fault code is recorded.

[0035] Secondly, the main detection unit broadcasts a "status query command" via its Bluetooth 5.0 BLE main communication unit. This command includes the device identifier and request type of the main detection unit, with a transmission period of 100ms and a duration of 5 seconds. All powered-on slave detection units listen to the broadcast channel. Upon receiving the status query command, each slave detection unit performs a self-test process similar to that of the main detection unit. This includes detecting its own adsorption force (using the built-in adsorption status monitoring circuit to determine whether the gap exceeds the standard based on the stability of the signal variance), measuring the battery SOC, and verifying the detection and communication hardware. Then, the self-test results, along with its own node ID, hardware version, and location type (front windshield / side windshield / rear windshield), are encapsulated into a response data frame and encrypted using AES-128 before being sent back to the main detection unit.

[0036] The master detection unit receives responses from all slave detection units within 5 seconds. If a slave detection unit fails to respond for three consecutive broadcast cycles, its deployment status is determined to be "unresponsive." If it responds but its self-test results contain any anomalies (e.g., adsorption force < 8N, SOC ≤ 18%, hardware self-test failure), its deployment status is marked as "abnormal." The master detection unit summarizes its own and all slave detection unit deployment statuses to generate a status table. Each row in the table records the detection unit ID, adsorption status (adsorbed / not adsorbed / gap too large), power status (normal / low battery / failed), and hardware status (normal / faulty). The output result "all deployment statuses are deployed" is true only when the master detection unit's own status is "deployed" and all paired slave detection units are also "deployed" (i.e., no anomalies, no unresponsiveness). Otherwise, the master detection unit displays specific abnormal information (e.g., "Left front windshield rain sensor adsorption is loose, please reinstall") on the vehicle's infotainment app via the vehicle domain controller and waits for intervention or automatic retry.

[0037] As an example, a vehicle is equipped with one main detection unit and two slave detection units (left front windshield and right front windshield). After powering on, the main detection unit performs a self-test: battery SOC is 85%, adsorption force is 15N (normal), the detection unit self-test passes, and its deployment status is marked as "deployed". Subsequently, the main detection unit broadcasts a status query command. The left front windshield slave detection unit replies: SOC 90%, adsorption force 12N, hardware normal, location type is front windshield; the right front windshield slave detection unit replies: SOC 20% (SOC≤18% alarm, 20% normal), adsorption force 10N, hardware normal. The main detection unit summarizes and determines that all units are in a "deployed state" (no abnormalities), outputting "All deployment statuses are deployed" as true, thus allowing subsequent network requests to continue. If the right front windshield slave detection unit replies that its adsorption force is only 3N (less than 8N), the main detection unit determines its deployment status as "abnormal", outputs false, and displays a message in the vehicle's app: "Right front windshield rain sensor adsorption force is insufficient, please reinstall."

[0038] Step A202: When all deployment statuses are in the deployed state, the main detection unit broadcasts a network request and establishes a target communication network with at least one slave detection unit that responds to the network request.

[0039] Specifically, a network request refers to a data frame broadcast by the master detection unit when initiating network establishment. This frame contains the device identifier (node ​​ID) of the master detection unit, the communication protocol version number, the network name, and encryption parameters, and is used to notify the slave detection units within range to prepare to join the network. The target communication network refers to the Bluetooth 5.0 BLE star topology wireless communication link established between the master detection unit and at least one slave detection unit. The master detection unit, as the central node, is responsible for network synchronization, data scheduling, and encryption management. Each slave detection unit, as an external node, communicates directly with the master detection unit only, and the nodes are not interconnected.

[0040] First, the main detection unit constructs a network request frame via its Bluetooth 5.0 BLE master communication unit. This frame's data structure includes at least one of the following: a frame header (identifying a network request), the main detection unit's node ID (a unique identifier), the communication protocol version number (e.g., V1.0), the network name (a pre-shared service set identifier), the initialization vector and key index used for AES-128 encryption, and a broadcast duration parameter (e.g., 30 seconds). The main detection unit periodically transmits this network request frame on the Bluetooth broadcast channel at 100ms intervals, with the transmission power set to 0dBm, ensuring an in-vehicle coverage distance ≤12m.

[0041] Secondly, all powered-on and properly deployed slave detection units continuously listen to the broadcast channel via Bluetooth 5.0 BLE. Upon receiving a network request frame, the slave detection unit first verifies whether the communication protocol version number in the frame is compatible with its own. If incompatible, it ignores the frame; if compatible, it extracts the node ID and network name of the master detection unit and generates a response frame. The response frame includes: the slave detection unit's node ID (a factory-preset unique identifier), hardware version number, current remaining battery power (SOC, accuracy ±1%), deployment location type (pre-configured as front windshield, side windshield, or rear windshield), and a randomly generated session token. The slave detection unit encrypts the response frame using the AES-128 encryption algorithm and then unicasts it to the master detection unit via the Bluetooth data channel.

[0042] Upon receiving a response from a slave detection unit, the master detection unit decrypts and verifies the response: checking if the node ID is within a preset whitelist, if the hardware version is compatible, and if the location type is valid. If the verification passes, the master detection unit assigns the slave detection unit a short internal network address (e.g., an integer from 1 to 8) and records its node information (ID, address, location type, SOC). Simultaneously, the master detection unit sends a "join confirmation" command to the slave detection unit, which includes the assigned short address and a network synchronization timestamp.

[0043] Upon receiving confirmation from the slave detection unit, the short address is saved and the system enters network synchronization mode, completing the unicast link establishment with the master detection unit. The master detection unit continuously broadcasts network requests and receives responses until the preset maximum number of slave detection units (N≤8) is reached or the broadcast times out (e.g., 30 seconds). Ultimately, the master detection unit and all successfully responding slave detection units together form a star-topology target communication network centered on the master detection unit. If a slave detection unit fails to respond after the timeout, the master detection unit marks it as "network failure" and prompts the user to check the sensor via the vehicle domain controller.

[0044] Step A203: The main detection unit controls the target communication network to enter the calibration state.

[0045] Specifically, calibration state refers to a special working mode that all detection units within the target communication network enter synchronously. In this mode, each detection unit stops normal rainfall detection and wiper control output, and instead collects reference signals under rainless conditions (i.e., the rainfall signal channel voltage Vs0 and the reference channel voltage Vr0 under rainless conditions) so that the compensation coefficient can be calculated later to eliminate baseline drift. After entering calibration state, each detection unit synchronously collects reference signals according to the sampling duration (e.g., 10 consecutive seconds) and sampling frequency (e.g., 10Hz) specified by the main detection unit. During the collection period, it does not respond to changes in rainfall or trigger wiper action.

[0046] First, the main control unit of the main detection unit generates an "enter calibration state" control command, which includes the following parameters: calibration identifier (fixed field), calibration duration (e.g., 10 seconds), sampling frequency (e.g., 10Hz), and synchronization timestamp (based on the main detection unit's local clock, with an accuracy requirement of ≤1ms). The main detection unit, through its Bluetooth 5.0 BLE master communication unit, encrypts the above command using AES-128 and sends it to each slave detection unit within the target communication network via broadcast or unicast per node.

[0047] To ensure all nodes enter calibration mode synchronously, the master detection unit adopts a strategy of "broadcasting a warning first, then unicasting confirmation": First, a "calibration warning" message is broadcast, informing all slave detection units that they are about to enter calibration mode and the calibration start time (absolute timestamp); upon receiving the warning, each slave detection unit prepares its local timer; when the calibration start time arrives, the master detection unit simultaneously sends a "calibration start" unicast command to all slave detection units, and each slave detection unit immediately enters calibration mode upon receiving it. After entering calibration mode, the master detection unit itself also synchronously switches to calibration mode and stops routine rainfall detection. In calibration mode, all detection units continuously acquire reference signals according to the sampling frequency (10Hz) specified in the command, temporarily storing the rainfall signal channel voltage Vs0 and reference channel voltage Vr0 acquired each time in a local buffer without performing differential processing or feature extraction.

[0048] At the end of the calibration duration (e.g., after 10 seconds), the main detection unit sends a "calibration end" command again. All detection units exit the calibration state, save the collected reference signal data, and prepare for the subsequent compensation coefficient calculation stage. Throughout the process, the main detection unit monitors the calibration status synchronization of each slave detection unit in real time through the target communication network. If a slave detection unit fails to respond to the "calibration start" command within the specified time, the main detection unit marks it as a calibration failure and prompts the user to check this node on the vehicle's APP.

[0049] Step A204: Obtain reference signals acquired by the main detection unit and at least one slave detection unit in the calibration state, and calculate the compensation coefficient of each detection unit based on the reference signals.

[0050] Specifically, the reference signal refers to the raw signal data collected by each detection unit in a rainless environment under calibrated conditions. This includes the rainfall signal channel voltage Vs0 (the voltage value received by the signal receiving channel after the infrared emitted light is reflected or scattered by the windshield) and the reference channel voltage Vr0 (the ambient light and common-mode interference voltage value collected by the reference receiving channel). These two sets of data are collected synchronously under rainless, unobstructed, and well-adhesive conditions, serving as the benchmark for subsequent compensation calculations. The compensation coefficient is a scalar value (denoted as k) used to differentially compensate the raw rainfall signal of each detection unit, and its definition formula is Vdiff=Vs k×Vr, where Vs is the voltage of the rainfall signal channel and Vr is the voltage of the reference channel. The goal of the compensation coefficient is to make the differential output Vdiff approach zero in the rainless state, thereby eliminating baseline drift caused by ambient light, temperature drift, device differences and changes in adsorption state. Each detection unit has its own independent compensation coefficient, which is calculated and stored separately.

[0051] First, the master detection unit sends an "upload reference signal" command to each slave detection unit via the target communication network. This command includes a requested data frame format (requiring each detection unit to report multiple sets of raw Vs0 and Vr0 data collected during the calibration duration). Upon receiving the command, each slave detection unit encapsulates its locally cached reference signal (e.g., 100 sets of data sampled at 10Hz within 10 seconds) into a data frame, encrypts it using AES-128, and uploads it to the master detection unit. Simultaneously, the master detection unit's own main control unit directly reads the self-calibration reference signal data from its internal cache. After receiving the reference signals from all slave detection units, the master detection unit calculates the compensation coefficient for each detection unit (including itself). The calculation process is as follows: For a given detection unit, assuming it has collected n sets of reference signals (Vs0(i), Vr0(i)), where i=1,2,…,n, the master detection unit first filters the n sets of data, removing obvious outliers (e.g., Vr0=0 or Vs0 exceeding the normal range of 0~0.5V), and then calculates the average value of the remaining valid data. and According to the differential compensation principle, the following should be satisfied under calibration conditions: Therefore, the compensation coefficient .like If the value is very close to zero (less than the preset threshold of 0.01V), the reference channel is considered abnormal. In this case, k is set to the default value of 1.0, and the detection unit is marked as needing maintenance. After calculating the k value, the master detection unit associates it with the node ID of the detection unit and stores it in the master detection unit's local non-volatile memory. At the same time, it sends the calculated k value to the corresponding slave detection unit through the target communication network for use by the slave detection unit in subsequent normal operation mode (the slave detection unit can also save its own k value locally to achieve distributed compensation). After the master detection unit completes the compensation coefficient calculation for all detection units, the entire calibration process is completed, and the system can then exit the calibration state and enter the normal rainfall detection mode.

[0052] By acquiring the deployment status of the main detection unit and at least one slave detection unit, the operational readiness of each detection unit can be clearly identified, providing a prerequisite for subsequent networking and calibration work. When all units are in the deployed state, the main detection unit broadcasts a networking request and establishes a target communication network with at least one slave detection unit responding to the request, achieving coordinated operation among the detection units and ensuring efficient transmission and interaction of rainfall data. By controlling the target communication network to enter calibration mode through the main detection unit, the working benchmarks of each detection unit can be unified, reducing system errors between different detection units. By acquiring reference signals collected by the main detection unit and at least one slave detection unit in calibration mode, and calculating the compensation coefficient for each detection unit based on these reference signals, the calculation of the compensation coefficient becomes more targeted and accurate, further improving the compensation effect of the original rainfall signal and ensuring the accuracy of rainfall feature extraction.

[0053] Step A3: For each detection unit, the original rainfall signal is compensated according to the compensation coefficient to obtain the compensated rainfall signal, and the corresponding rainfall features are extracted from the compensated rainfall signal.

[0054] Specifically, compensating for rainfall signals refers to adjusting the rainfall signal channel voltage Vs in the original rainfall signal and the reference channel voltage Vr according to the formula Vdiff=Vs. The differential voltage signal obtained after k×Vr calculation has eliminated common-mode interference such as ambient light, temperature drift, and adsorption gaps, and truly reflects the degree of obstruction of the optical path by raindrops; the rainfall characteristics refer to the quantization parameters extracted from the compensated rainfall signal Vdiff, including at least one of the following: peak value P (characterizing raindrop size and rainfall intensity, i.e., the maximum amplitude of the Vdiff waveform in one sampling period), variance Var (characterizing the uniformity of raindrop distribution and the degree of signal fluctuation, i.e., the statistical variance of the Vdiff sample values), and frequency F (characterizing the raindrop impact frequency, i.e., the number of zero crossings or pulse counts of the Vdiff waveform per unit time).

[0055] First, each detection unit's dual-channel differential rainfall detection unit continuously acquires raw rainfall signals at a fixed sampling frequency (≥10Hz) to obtain the current rainfall signal channel voltage Vs and reference channel voltage Vr. Then, the detection unit reads the locally stored compensation coefficient k (k for the main detection unit is calculated and stored by itself, while k for the slave detection units is sent and stored by the main detection unit after calibration), according to the formula Vdiff=Vs. The differential voltage is calculated in real time using k×Vr to obtain the compensated rainfall signal Vdiff. This calculation process is completed by the differential processing circuit inside the detection unit or the hardware multiplier in the main control unit, with a calculation delay of ≤5ms. After obtaining Vdiff, the detection unit further extracts rainfall characteristics from the signal: the peak value P is extracted by recording the maximum amplitude of Vdiff within a sampling window (e.g., 100ms, corresponding to 10 sampling points), which is the peak value of that window; the variance Var is calculated by statistically analyzing the variance of all Vdiff sample values ​​within the same window, i.e. Where n is the number of sampling points within the window, The mean is used; the frequency F is extracted by detecting the number of times the Vdiff waveform crosses a preset threshold (e.g., 0.02V), and the number of crossings per unit time is the frequency. These characteristic parameters constitute the rainfall characteristics of this detection unit within the current sampling period. For the main detection unit, its main control unit directly reads the rainfall characteristics it extracts; for each slave detection unit, the extracted rainfall characteristics (along with node ID and timestamp) are uploaded to the main detection unit via the target communication network after being encrypted using AES-128.

[0056] By acquiring the raw rainfall signal collected by each detection unit, basic data support is provided for subsequent rainfall feature extraction, ensuring the authenticity and integrity of the data source. By acquiring the compensation coefficient corresponding to each detection unit, the baseline drift problem that occurs during the acquisition process of different detection units can be specifically solved, avoiding the interference of baseline drift on the accuracy of rainfall detection. By compensating the raw rainfall signal according to the compensation coefficient for each detection unit to obtain a compensated rainfall signal, and extracting the corresponding rainfall features from the compensated rainfall signal, the deviation of the original signal is effectively corrected, improving the accuracy and reliability of rainfall features, and providing accurate rainfall data basis for subsequent wiper control.

[0057] Step S102: Perform validity testing on the rainfall characteristics collected by each detection unit to obtain the valid rainfall characteristics corresponding to each detection unit.

[0058] In this embodiment, rainfall features refer to the parameters extracted by each detection unit from the compensated rainfall signal: peak value P (characterizing raindrop intensity), variance Var (characterizing signal fluctuation), and frequency F (characterizing raindrop impact frequency); validity detection refers to identifying and removing invalid rainfall features caused by gap anomalies, occlusion anomalies, or fault anomalies through preset judgment rules, while retaining features that truly reflect the rainfall situation; valid rainfall features refer to the credible rainfall features remaining after anomaly removal, which are used for subsequent fusion calculations.

[0059] The main detection unit performs validity checks on the rainfall characteristics of all detection units (including itself and slave detection units) within the current cycle. The main detection unit sequentially performs anomaly detection and removal: First, it calculates the average of the signal variances of all detection units as the baseline variance. Detection units with variances greater than 2.5 times the baseline variance are identified as having intermittent anomalies and their rainfall characteristics are removed. Then, among the remaining characteristics, it identifies detection units with a peak value of 0V and calculates the average of the peak values ​​of other detection units. If this average is ≥0.05V, the unit is identified as having an occlusion anomaly and its rainfall characteristics are removed. Finally, it checks whether each slave detection unit has experienced data loss or data exceeding limits (e.g., peak value <0 or >2.0V) for five consecutive cycles. If so, it is identified as having a fault anomaly and its rainfall characteristics are removed. After removal, the remaining rainfall characteristics are the valid rainfall characteristics. If the number of valid detection units after removal is 0, it enters a safe mode and uses the default intermittent wiping strategy.

[0060] Step S103: Fuse the effective rainfall features corresponding to each detection unit to obtain fused rainfall features.

[0061] In this embodiment, effective rainfall features refer to the rainfall features from each detection unit that are retained after effectiveness detection. Each effective rainfall feature includes a peak value Pi and a variance Varfusion (frequency Fi). Fusion rainfall features refer to the comprehensive parameters obtained by weighting and fusing multiple effective rainfall features according to preset weights. For example, they include fusion peak value Pfusion (characterizing the overall rainfall intensity of the entire vehicle) and fusion variance Varfusion (characterizing the uniformity of raindrop distribution throughout the vehicle), which are used for subsequent rainfall level determination.

[0062] Specifically, the main detection unit performs weighted fusion calculations based on the location type of each valid detection unit (main detection unit, front windshield secondary detection unit, side / rear windshield secondary detection unit) according to a preset weight allocation rule. First, the main detection unit acquires the rainfall characteristics of all valid detection units for the current period, including at least one of the following: the peak value Pi and variance Vari for each unit, and the location type of each unit. Then, the weight value of each unit is determined according to the following weight allocation rule: weight Wm = 0.45 for the main detection unit (located in the core area of ​​the front windshield); weight Wf = 0.25 for each front windshield secondary detection unit (located on both sides of the front windshield); and weight Ws = 0.1 for each side or rear windshield secondary detection unit. The sum of the weights of all valid detection units should equal 1. If the sum of weights is less than 1 due to abnormal removal, the remaining weights are normalized. Next, the main detection unit calculates the fused peak value according to the fusion formula: Pfusion=Wm×Pm+∑(Wf×Pf)+∑(Ws×Ps), which is the peak value of the main detection unit multiplied by its weight, plus the peak values ​​of all front / rear windshield secondary detection units multiplied by their weights, plus the peak values ​​of all side / rear windshield secondary detection units multiplied by their weights. Similarly, the fused variance is calculated: Varfusion=Wm×Varm+∑(Wf×Varf)+∑(Ws×Vars). The calculated Pfusion and Varfusion are the fused rainfall characteristics. If the number of effective detection units is only 1 (e.g., only the main detection unit is effective), then the peak value and variance of that unit are directly used as the fused rainfall characteristics, and no further weighting is performed ("If the number of effective nodes is ≤1, then only the main node data is used").

[0063] Step S104: Control the vehicle's windshield wipers to perform wiping actions corresponding to the rainfall characteristics.

[0064] In this embodiment of the application, controlling the windshield wipers of the vehicle to perform the wiping action corresponding to the fused rainfall characteristics includes: generating a control command for the windshield wipers based on the fused rainfall characteristics; and sending the control command to the windshield wipers to perform the wiping action corresponding to the fused rainfall characteristics.

[0065] Specifically, the control commands for the windshield wipers are generated based on the fused rainfall characteristics, including: determining the rainfall level that matches the fused rainfall characteristics based on the mapping relationship between preset rainfall characteristics and preset rainfall levels; obtaining the corresponding wiping parameters from the preset windshield wiper control strategy based on the rainfall level; and generating control commands for the windshield wipers using the wiping parameters.

[0066] The fused rainfall characteristics refer to the parameters obtained after anomaly removal and weighted fusion: fused peak value Pfusion (characterizing the overall rainfall intensity) and fused variance Varfusion (characterizing the uniformity of raindrop distribution); the mapping relationship between preset rainfall characteristics and preset rainfall levels refers to a pre-stored lookup table that divides the range of values ​​for fused peak value and fused variance into multiple levels; the rainfall level is an integer used to quantify the current rainfall amount; the preset wiper control strategy refers to predefined wiper action rules that assign corresponding wiping parameters to each rainfall level; the wiping parameters are the parameters in the specific control command, including at least the wiping mode and wiping frequency; the control command is a binary command frame sent by the main detection unit to the wiper, containing the target wiping mode and target wiping frequency; the wiper includes a vehicle domain controller and a wiper actuator, the vehicle domain controller receives the command and drives the actuator to complete the wiping action.

[0067] For example, the mapping relationship between preset rainfall characteristics and preset rainfall levels, and the corresponding wiper control commands are shown in the table below:

[0068] First, the main detection unit acquires the fused rainfall characteristics calculated for the current period, namely the fused peak value Pfusion and the fused variance Varfusion. Then, the main detection unit queries the internally stored mapping table between preset rainfall characteristics and preset rainfall levels, comparing Pfusion and Varfusion with the value ranges of each level in the table to determine the matching rainfall level. Next, based on this rainfall level, the main detection unit retrieves the corresponding wiping parameters from the preset wiper control strategy: according to the mapping table, level 0 corresponds to no wiping (no parameters), level 1 corresponds to intermittent wiping with a wiping frequency of 3 seconds / wipe, level 2 corresponds to low-speed wiping with a frequency of 35 wipes / minute, level 3 corresponds to medium-speed wiping with a frequency of 50 wipes / minute, level 4 corresponds to high-speed wiping with a frequency of 70 wipes / minute, and level 5 corresponds to high-speed continuous wiping (without interruptions). Then, the main detection unit uses the acquired wiping parameters to generate control commands for the windshield wipers. These commands encapsulate the target wiping mode (e.g., intermittent wiping) and target wiping frequency (e.g., 3 seconds / wipe). The command format is a fixed-length binary frame including a checksum. Finally, the main detection unit sends the control commands to the windshield wipers via the controller area network bus or local interconnect network bus. Upon receiving the commands, the vehicle domain controller in the windshield wipers parses the wiping parameters and drives the wiper actuators to perform wiping actions according to the specified mode and frequency. If the rainfall level is 0 (no rain), a stop wiping command is generated, and the wipers stop operating.

[0069] Specifically, after generating control commands, the updated rainfall characteristics collected by each detection unit are acquired; when the updated rainfall characteristics collected by each detection unit meet the sleep conditions, the main detection unit and at least one slave detection unit are controlled to enter sleep mode; or, when the updated rainfall characteristics collected by any detection unit meet the wake-up conditions, each detection unit is woken up through any detection unit to resume data acquisition.

[0070] Updated rainfall characteristics refer to the rainfall characteristics re-collected and extracted by each detection unit in the next one or more sampling periods after the control command is generated, including at least one of the following: peak value P, variance Var, and frequency F of each detection unit; the dormancy condition refers to the system having no effective rainfall for 60 consecutive seconds, i.e., the fused peak value Pfusion < 0.05V and the fused variance Varfusion < 0.01 calculated based on the updated rainfall characteristics of all detection units. The presence of rainless conditions indicates that the rainless state has lasted for a sufficiently long time. Sleep mode refers to an ultra-low power state entered by the detection unit. In this mode, the main detection unit and all slave detection units stop data acquisition, communication, and wiper control, retaining only the wake-up monitoring circuit, with power consumption ≤8μA. The wake-up condition is that the peak value P in the updated rainfall characteristics acquired by any detection unit is ≥0.05V, indicating that raindrops have fallen on the detection area of ​​that detection unit. "Wake up each detection unit" means that when a detection unit that meets the wake-up condition automatically recovers from sleep mode, it sends a wake-up signal to the main detection unit through the target communication network. The main detection unit then synchronously wakes up all other detection units, restoring the entire array to normal rainfall acquisition and control processes, with a wake-up delay ≤10ms.

[0071] This embodiment provides a windshield wiper control method. Figure 3 This is a flowchart of a windshield wiper control method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S201: Obtain rainfall characteristics collected by each detection unit in the rainfall detection unit array, wherein the rainfall detection unit array is deployed on the vehicle and includes a main detection unit and at least one slave detection unit.

[0072] Step S202: Perform validity testing on the rainfall characteristics collected by each detection unit to obtain the valid rainfall characteristics corresponding to each detection unit.

[0073] In this embodiment of the application, the validity of the rainfall characteristics collected by each detection unit is detected to obtain the valid rainfall characteristics corresponding to each detection unit, including: Step B1: Obtain the abnormal factors that affect the effectiveness of rainfall detection in the detection unit, and obtain at least one judgment index corresponding to each abnormal factor.

[0074] Specifically, abnormal factors include: gap abnormalities, occlusion abnormalities, and fault abnormalities; the judgment indicators include the signal variance of the rainfall signal, the signal peak value of the rainfall signal, and the communication status of the detection unit. The signal variance is the judgment indicator corresponding to the gap abnormality type, the signal peak value is the judgment indicator corresponding to the occlusion abnormality type, and the communication status is the judgment indicator corresponding to the fault abnormality type.

[0075] Anomalies refer to interference types that may distort the rainfall characteristics collected by the detection unit and fail to accurately reflect the actual rainfall conditions. Anomalies include at least one of the following: gap anomalies, occlusion anomalies, and fault anomalies. Judgment indicators refer to one or more quantifiable and measurable parameters used to identify the presence of each anomaly. Gap anomalies refer to stray reflection interference caused by loose adsorption of the detection unit or excessive gaps, manifested as an abnormally increased variance of the rainfall signal. The corresponding judgment indicator is the signal variance of the rainfall signal (i.e., the statistical variance value of the compensated rainfall signal Vdiff within a sampling window, denoted as Var, characterizing the degree of signal fluctuation). Occlusion anomaly refers to the detection area of ​​the detection unit being partially obstructed by objects such as leaves, snow, and stains, resulting in an abnormally low or zero peak value of the rainfall signal. The corresponding judgment index is the peak value of the rainfall signal (i.e., the difference between the maximum and minimum amplitude of the compensated rainfall signal Vdiff within a sampling window, denoted as P, representing the raindrop intensity). Fault anomaly refers to a hardware failure or communication link interruption in the detection unit itself, resulting in the inability to upload data normally. The corresponding judgment index is the communication status of the detection unit (i.e., whether the main detection unit can receive data frames from the detection unit within the expected period, including data loss or data exceeding limits).

[0076] First, the main control unit of the main detection unit pre-stores anomaly factor-judgment index mapping table in its internal firmware. This table clearly defines the judgment index types (signal variance, signal peak value, communication status) corresponding to the three types of anomalies (gap anomaly, occlusion anomaly, and fault anomaly) and the quantization thresholds used in subsequent steps. The main detection unit directly reads this mapping table, thus completing the preparation of information to obtain the anomalies and at least one judgment index corresponding to each anomaly factor. Second, the main detection unit acquires the specific values ​​of each judgment index of each detection unit (including itself and all slave detection units) in real time within the current cycle for subsequent anomaly identification. The main detection unit temporarily stores these real-time acquired judgment index data (Var, P, and communication status of each detection unit) in its local memory and associates them with the aforementioned mapping table to provide input for anomaly rejection.

[0077] Step B2: Identify abnormal features in the rainfall characteristics based on at least one judgment index, and remove the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics.

[0078] It should be noted that the elimination operations corresponding to abnormal factors (such as gap abnormalities, occlusion abnormalities, and fault abnormalities) can be used individually or in combination according to actual needs. When used in combination, each elimination operation is executed in a preset order, and the remaining rainfall characteristics after the previous elimination are used as the input for the next elimination, finally obtaining the effective rainfall characteristics. The combination order is not limited to the gap, occlusion, and fault order shown in Case 4, and other orders such as occlusion, gap, and fault can also be used.

[0079] By identifying anomalous factors affecting the effectiveness of rainfall detection by the detection unit and obtaining at least one judgment index corresponding to each anomalous factor, the direction and standards for anomalous detection are clarified, providing a basis for accurately identifying anomalous rainfall characteristics. By identifying anomalous features in rainfall characteristics based on at least one judgment index and removing anomalous features from the rainfall characteristics to obtain effective rainfall characteristics, invalid and erroneous rainfall data interference can be eliminated, ensuring that the subsequently fused rainfall characteristics can truly reflect the actual rainfall situation, improving the rationality and accuracy of wiper control, and avoiding abnormal wiper operation due to anomalous data.

[0080] By clearly defining anomalous factors, including gap anomalies, occlusion anomalies, and malfunction anomalies, the system comprehensively covers the main anomaly types that the detection unit may encounter during rainfall detection, avoiding the omission of key anomalies. Furthermore, by specifying judgment indicators, including the signal variance of the rainfall signal, the signal peak value of the rainfall signal, and the communication status of the detection unit, and corresponding to gap anomalies, occlusion anomalies, and malfunction anomalies respectively, each anomaly type has a clear judgment basis. This achieves accurate classification and identification of anomaly features, improves the targeting and efficiency of anomaly detection, provides more detailed support for the subsequent acquisition of effective rainfall features, and further ensures the reliability of rainfall data.

[0081] Specifically, identifying anomalous features in rainfall characteristics based on at least one criterion and removing these anomalous features from the rainfall characteristics to obtain valid rainfall characteristics includes at least the following three cases: Scenario 1: When the anomalous factor is intermittent anomaly, calculate the average variance based on the signal variance in each rainfall feature to obtain the baseline variance; identify the first anomalous feature in the rainfall features whose signal variance is greater than the baseline variance, and remove the first anomalous feature from the rainfall features to obtain the effective rainfall features.

[0082] Specifically, the baseline variance is the arithmetic mean of the signal variances of all participating detection units, used as a benchmark to determine whether the variance of a certain detection unit is abnormal, representing the typical level of the signal variance of a normal detection unit; the first abnormal feature refers to those rainfall feature data whose signal variance is greater than the baseline variance, and the detection units corresponding to these features are considered to have gap anomalies and need to be removed; the effective rainfall feature refers to the rainfall feature retained after anomaly removal and identified as truly reflecting the rainfall situation.

[0083] First, the main detection unit acquires the rainfall characteristics of all detection units (including the main detection unit itself and all slave detection units) within the current sampling period, and extracts the signal variance value Vari (i represents the detection unit number) for each detection unit. Assume there are M detection units participating in this calculation (M equals the number of the main detection unit plus the number of currently active slave detection units). Second, the main control unit of the main detection unit calculates the rainfall characteristics according to the formula... The arithmetic mean of the variances of all detection units is calculated and used as the baseline variance. Then, the main detection unit compares the Vari of each detection unit with the baseline variance Varavg. Based on the quantization threshold, if the Vari of a detection unit is ≥ 2.5 × Varavg (i.e., greater than 2.5 times the baseline variance), the detection unit is determined to have gap anomalies, and its corresponding rainfall features (including peak value P, variance Var, and frequency F) are marked as the first anomalous feature. Finally, the main detection unit removes all rainfall features marked as first anomalous features from the rainfall feature set of the current period. The remaining rainfall features are the valid rainfall features, used for subsequent weighted fusion calculations. If the number of valid detection units is insufficient after removal, the system can still continue to operate (e.g., only the main detection unit data is retained), but a corresponding alarm will be triggered.

[0084] By calculating the average variance of the signal variance in each rainfall feature when the anomaly is intermittent anomaly, a baseline variance is obtained, providing an objective and unified reference standard for judging intermittent anomalies and avoiding errors caused by subjective judgment. By identifying the first anomaly feature in the rainfall features whose signal variance is greater than the baseline variance and removing the first anomaly feature from the rainfall features, the effective rainfall features are obtained. This can accurately screen out invalid rainfall data caused by intermittent anomalies, eliminate the impact of intermittent anomalies on the accuracy of rainfall detection, and ensure that the effective rainfall features can truly reflect the actual rainfall, providing accurate data support for wiper control.

[0085] Scenario 2: When the abnormal factor is occlusion, identify the second abnormal feature in the rainfall feature whose signal peak value is the first preset value; for other detection units besides the detection unit corresponding to the candidate abnormal feature, calculate the average value of the signal peak value in their corresponding rainfall feature to obtain the benchmark peak value; detect whether the benchmark peak value is greater than or equal to the second preset value; when the benchmark peak value is greater than or equal to the second preset value, remove the second abnormal feature from the rainfall feature to obtain the effective rainfall feature.

[0086] Specifically, the first preset value refers to the peak threshold used to determine whether an occlusion anomaly has occurred, preset to 0V (i.e., the signal peak value equals 0); the second abnormal feature refers to the rainfall feature data whose signal peak value equals the first preset value. The detection units corresponding to these features are suspected of having occlusion anomalies, but further verification is needed to confirm whether other detection units have rain; the baseline peak value refers to the value obtained by calculating the arithmetic mean of the peak values ​​of other normal detection units after removing candidate abnormal detection units, used to determine whether the current environment is indeed rainy (if there is rain and the peak value of a certain node is 0, then the node is confirmed to be occluded); the second preset value refers to the threshold used to determine whether the baseline peak value has reached the state of rain, preset to 0.05V (i.e., when the average peak value of other detection units is ≥0.05V, it is considered that the environment is rainy, and the node with a peak value of 0 at this time is indeed an occlusion anomaly).

[0087] First, the main detection unit acquires the rainfall characteristics of all detection units within the current sampling period and extracts the signal peak value Pi for each detection unit. The main detection unit checks each Pi against a first preset value (i.e., Pi = 0V), marking the rainfall characteristics of all detection units that satisfy Pi = 0V as second anomaly features; these detection units are then considered candidate anomaly detection units. Second, for each candidate anomaly detection unit, the main detection unit calculates the average peak value of all other detection units, obtaining the baseline peak value Pavg_others. The specific formula is: M represents the total number of detection units. Then, the main detection unit checks whether the baseline peak value is greater than or equal to a second preset value (i.e., Pavg_others ≥ 0.05V). If the condition is met, it means that all other detection units except for this candidate node have detected effective rainfall (peak value ≥ 0.05V), while the peak value of this candidate node is 0. Therefore, it is determined that this candidate node does indeed have an occlusion anomaly, and its corresponding second anomaly feature is removed from the rainfall features. If the condition is not met (i.e., baseline peak value < 0.05V), it means that there is no rain in other areas, and the peak value of this candidate node being 0 may be due to global lack of rain rather than occlusion; therefore, this feature is not removed. The main detection unit performs the above judgment on all candidate anomaly detection units one by one, and finally retains the rainfall features that have not been removed as effective rainfall features.

[0088] By identifying a second abnormal feature in rainfall characteristics with a signal peak value equal to a first preset value when the abnormal factor is occlusion, the rainfall data corresponding to the detection unit that may have occlusion can be quickly located, improving the efficiency of occlusion anomaly identification. By calculating the average value of the signal peak value in the rainfall characteristics of other detection units besides the detection unit corresponding to the candidate abnormal feature, a baseline peak value is obtained, providing a reference for judging the authenticity of occlusion anomalies. By detecting whether the baseline peak value is greater than or equal to the second preset value, it is possible to distinguish whether the signal peak anomaly is caused by occlusion of a single detection unit or by low overall rainfall, avoiding misjudgment. By removing the second abnormal feature from the rainfall characteristics when the baseline peak value is greater than or equal to the second preset value, the valid rainfall characteristics can be obtained, which can accurately exclude invalid data caused by occlusion, ensuring the accuracy of the valid rainfall characteristics and ensuring that the wiper control action matches the actual rainfall.

[0089] Scenario 3: When the abnormal factor is a fault, identify the third abnormal feature in the rainfall characteristics that has missing or excessive data in a continuous preset number period, remove the third abnormal feature from the rainfall characteristics, and obtain the effective rainfall characteristics.

[0090] Specifically, "continuous preset quantity period" refers to the counting threshold preset by the main detection unit, such as a preset of 5 periods (i.e., 5 consecutive 100ms sampling periods); "data missing" means that the main detection unit does not receive any data frame uploaded by a certain detection unit within a certain sampling period, or the received data frame is incomplete (such as missing node ID, timestamp, or rainfall characteristic parameters); "data exceeding limits" means that the rainfall characteristic data uploaded by the detection unit exceeds the normal working range, such as the voltage range corresponding to peak value P<0V or P>50mm / h (exceeding the range), or unreasonable values ​​such as negative variance Var; "third abnormal feature" refers to the rainfall characteristic data corresponding to the detection unit that meets the fault abnormality judgment conditions, and these features need to be removed from the valid dataset.

[0091] First, the master detection unit maintains a continuous anomaly counter for each slave detection unit, initially set to 0. During each sampling period (100ms), the master detection unit receives data frames uploaded by each slave detection unit via the target communication network. For each slave detection unit, the master detection unit checks whether a complete data frame was successfully received during that period, and whether the rainfall characteristics (peak value P, variance Var, frequency F) in the received data are within a preset valid range (e.g., 0 ≤ P ≤ 2.0V, Var ≥ 0). If the master detection unit does not receive a data frame from a slave detection unit, or if the received data contains a negative P value or exceeds 2.0V, it determines that the detection unit has experienced a "data missing or data exceeding limit" event during that period, and increments the continuous anomaly counter for that detection unit by 1; if the data is normal during that period, the continuous anomaly counter is reset to zero. Then, at the end of each period, the master detection unit checks whether the continuous anomaly counter for each slave detection unit has reached a preset number of consecutive periods (e.g., preset to 5). When the continuous anomaly counter of a slave detection unit reaches ≥5, the master detection unit determines that the detection unit is faulty and marks its corresponding rainfall feature (including all historical and current period data) as a "third anomaly feature". The master detection unit removes this third anomaly feature from the rainfall feature set of the current period and no longer participates in subsequent fusion calculations. At the same time, the master detection unit marks the detection unit as a "failed node", stops receiving its data, and prompts the user on the vehicle's infotainment app via the vehicle domain controller: "Location F rain sensor communication abnormal or faulty, please check". For faults in the master detection unit itself (such as failure of its own detection unit), it identifies the fault through a self-test program and directly enters a fail-safe mode (e.g., relying only on slave detection unit data or using the default wiper speed).

[0092] By identifying a third abnormal feature in rainfall characteristics—specifically, a third abnormal feature—where data is missing or exceeds limits for consecutive preset number periods when the abnormal factor is a fault, it is possible to accurately determine whether the detection unit is malfunctioning, thus avoiding misjudging occasional data anomalies as faults. By removing the third abnormal feature from the rainfall characteristics to obtain valid rainfall characteristics, invalid and erroneous data from the faulty detection unit can be eliminated, preventing faulty data from affecting the rainfall fusion results and ensuring the accuracy of subsequent wiper control commands. It can also indirectly indicate potential faults in the detection unit, facilitating timely maintenance.

[0093] Scenario 4: When the abnormal factors include gap anomaly type, shading anomaly type, and fault anomaly type, calculate the average variance based on the signal variance in each rainfall feature to obtain the baseline variance; identify the first abnormal feature in the rainfall features whose signal variance is greater than the baseline variance, and remove the first abnormal feature from the rainfall features to obtain the first remaining feature; identify the second abnormal feature in the first remaining feature whose signal peak value is a first preset value; for other detection units besides the detection units corresponding to the candidate abnormal features, calculate the average value of the signal peak values ​​in the corresponding first remaining features to obtain the baseline peak value; detect whether the baseline peak value is greater than or equal to the second preset value; when the baseline peak value is greater than or equal to the second preset value, remove the second abnormal feature from the first remaining feature to obtain the second remaining feature; identify the third abnormal feature in the second remaining feature that has data missing or data exceeding the limit for a consecutive preset number of periods, remove the third abnormal feature from the second remaining feature to obtain the effective rainfall feature.

[0094] Specifically, this scenario involves a comprehensive process for handling three types of anomalies: gap anomalies, occlusion anomalies, and fault anomalies. The first remaining feature refers to the rainfall feature retained after gap anomaly removal; the second remaining feature refers to the rainfall feature retained after further occlusion anomaly removal; and the effective rainfall feature is the final rainfall feature used for subsequent fusion calculations after three levels of removal. The meanings of other terms (such as signal variance, baseline variance, first anomaly feature, signal peak value, first preset value, second anomaly feature, baseline peak value, second preset value, continuous preset number of periods, missing data, data exceeding limits, and third anomaly feature) are the same as those defined in scenarios one, two, and three above.

[0095] The three-level rejection process is performed sequentially, with the remaining features after each level serving as input for the next level. Level 1 (Gap Anomaly Rejection): The main detection unit calculates the average of the signal variances of all current detection units as the baseline variance. Detection units with variances greater than 2.5 times the baseline variance are marked as first anomaly features and rejected, resulting in the first remaining feature. Level 2 (Obstruction Anomaly Rejection): In the first remaining feature, detection units with a signal peak value of 0V are identified as candidate anomaly features. For each candidate, the average of the peak values ​​of all other detection units (excluding itself) is calculated as the baseline peak value. If this baseline peak value is ≥0.05V, the candidate is confirmed as an obstruction anomaly, and its corresponding second anomaly feature is removed from the first remaining feature, resulting in the second remaining feature. Level 3 (Fault Anomaly Rejection): In the second remaining feature, each slave detection unit is checked for data loss or exceeding limits (e.g., peak value <0 or >2.0V) for five consecutive cycles. If so, it is marked as a third anomaly feature and rejected. The final remaining feature is the effective rainfall feature. If the number of effective detection units after rejection is ≤1, only the main detection unit data is retained.

[0096] Figure 4 This is a schematic diagram of the validity detection and weighted fusion process of rainfall features in an embodiment of the present invention. The process includes: firstly, performing three levels of anomaly removal on the original rainfall features collected by the main / auxiliary nodes, namely, gap anomaly removal (removing node data with variance ≥ 2.5 times the average variance), occlusion anomaly removal (removing node data with a peak value of 0 and the average peak value of other nodes ≥ 0.05V), and fault anomaly removal (removing node data with no data for 5 consecutive cycles / out of range), to obtain valid node data; then, performing weighted fusion calculation according to preset weights, where the weight of the main node is 0.45, the weight of the front windshield auxiliary node is 0.25 / each, and the weight of the side and rear windshield auxiliary nodes is 0.1 / each, finally generating the fusion peak value P_fusion and the fusion variance Var_fusion, which serve as the core input parameters for wiper control decisions.

[0097] By combining and using the anomaly removal steps corresponding to gap anomalies, occlusion anomalies, and malfunction anomalies, invalid data corresponding to gap anomalies is first removed, then erroneous data caused by occlusion anomalies is precisely eliminated, and finally the interference caused by malfunction anomalies is eliminated. This achieves a comprehensive and orderly investigation of various anomalies, maximizes the accuracy and reliability of effective rainfall characteristics, avoids rainfall data deviations caused by incomplete removal of a single anomaly, provides strong data support for the precise control of windshield wipers, and significantly improves the rationality and stability of windshield wiper control.

[0098] Step S203: Fuse the effective rainfall features corresponding to each detection unit to obtain fused rainfall features.

[0099] Step S204: Control the vehicle's windshield wipers to perform wiping actions that integrate the rainfall characteristics.

[0100] This embodiment provides a windshield wiper control method. Figure 5 This is a flowchart of a windshield wiper control method according to an embodiment of the present invention, such as... Figure 5 As shown, the process includes the following steps: Step S301: Obtain rainfall characteristics collected by each detection unit in the rainfall detection unit array, wherein the rainfall detection unit array is deployed on the vehicle and includes a main detection unit and at least one slave detection unit.

[0101] In this embodiment, each detection unit in the rain detection unit array is attached to the vehicle via a magnetic structure. The magnetic structure includes a magnetic base, a magnetic circuit assembly, a sealing assembly, and a buffer pad. The magnetic circuit assembly is disposed inside the magnetic base and is used to adjust the magnetic attraction force. The sealing assembly is disposed at the edge of the magnetic base and is used to form a seal with the vehicle. The buffer pad is disposed on the magnetic base and is used to conform to the vehicle to form a buffer.

[0102] Specifically, the magnetic base can be a flexible base injection molded from medical-grade liquid silicone and neodymium iron boron N42 permanent magnets, with a contact layer roughness Ra≤0.6μm, used to provide basic adsorption force and avoid scratching the glass; the magnetic circuit assembly can be a built-in rotary magnetic flux adjustment ring, which changes the magnetic circuit conduction area by rotation, so that the adsorption force can be continuously adjusted in the range of 5N~25N; the sealing assembly can be an annular sealing lip integrally formed on the edge of the magnetic base, with a thickness of 0.6mm~1.2mm and a width of 2mm~4mm, used to form a line contact seal with the glass and eliminate adsorption gaps (gap ≤0.1mm); the buffer pad can be a TPU (thermoplastic polyurethane elastomer) pad with a thickness of 0.3mm set at the four corners of the magnetic base, used to improve the flatness of the fit and cushion and prevent scratches during installation.

[0103] Each detection unit is magnetically attached to the inner surface of the vehicle's windshield. First, the magnetic base serves as the overall support, its main body being medical-grade liquid silicone with a Shore hardness of 30-50, embedded with neodymium iron boron (N42) permanent magnets. The silicone material ensures flexible contact with the glass, preventing scratches. Second, the magnetic circuit assembly is located inside the magnetic base, specifically a rotatable magnetic flux adjustment ring. This ring surrounds the permanent magnets, allowing users to manually rotate it to change the effective cross-sectional area through which the magnetic lines of force pass, thereby continuously adjusting the attraction force of the detection unit on the glass, ranging from 5N to 25N, suitable for automotive glass with thicknesses from 3mm to 6mm. Third, the sealing component is located at the edge of the magnetic base and is integrally molded with the base, forming an annular sealing lip with a width of 2mm~4mm and a thickness of 0.6mm~1.2mm. When the detection unit is attracted to the glass, the sealing lip is deformed under pressure, forming a tight line contact seal with the glass surface, isolating the detection area from the external environment, preventing moisture and dust from entering, and controlling the adsorption gap to within 0.1mm, eliminating stray reflection interference caused by the gap. Fourth, the buffer pads are located at the four corners of the magnetic base and are made of 0.3mm thick TPU material. During installation, they first contact the glass, acting as a buffer to prevent scratches from installation impacts, and also helping to improve the flatness of the fit between the base and the glass. The above four components work together to achieve reliable adsorption that is non-destructive, adjustable, sealed, and scratch-resistant.

[0104] Figure 6This is a schematic diagram of the magnetic flexible sealing and adsorption layered structure of the detection unit in this embodiment of the invention. The structure includes, from top to bottom, a sensor housing, a steel boron N42 permanent magnet, a rotary magnetic flux adjustment ring, a medical-grade liquid silicone base with a Shore hardness of 30-50, a 0.6-1.2mm thick annular sealing lip, and a 0.3mm thick TPU scratch-resistant buffer pad, which is finally adsorbed onto the inner surface of the car glass. The rotary magnetic flux adjustment ring can achieve a continuously adjustable adsorption force of 5-25N, the annular sealing lip can control the adsorption gap to ≤0.1mm, and the TPU buffer pad provides scratch protection during installation. This structure realizes the adjustable and sealed installation of the detection unit, eliminates stray reflection interference, and ensures the accuracy of rainfall detection.

[0105] Each detection unit in the rain detection unit array is magnetically attached to the vehicle, eliminating the need for drilling, welding, or other destructive installation methods. This simplifies the installation process, reduces installation difficulty, and minimizes damage to the vehicle. The magnetic structure, including the magnetic base, provides a stable mounting surface for the magnetic circuit assembly, sealing assembly, and buffer pad, ensuring the overall stability of the magnetic structure. The magnetic circuit assembly, located inside the magnetic base, adjusts the magnetic force according to the vehicle's surface material and installation requirements, ensuring a firm attachment of the detection unit and preventing it from falling off or shifting due to vehicle vibrations during travel, thus guaranteeing the stability of the detection operation. The sealing assembly, located at the edge of the magnetic base, forms a seal with the vehicle.

[0106] It adopts a magnetic flexible sealing adsorption structure, which can achieve non-destructive and quick installation without disassembling the interior and wiring. The adsorption force can be flexibly adjusted to adapt to car glass of different thicknesses. It will not fall off when the vehicle is traveling at high speed and will not scratch the car glass.

[0107] Step S302: Perform validity testing on the rainfall characteristics collected by each detection unit to obtain the valid rainfall characteristics corresponding to each detection unit.

[0108] Step S303: Fuse the effective rainfall features corresponding to each detection unit to obtain fused rainfall features.

[0109] Step S304: Control the vehicle's windshield wipers to perform wiping actions that integrate the rainfall characteristics.

[0110] After generating control commands, the main detection unit continues to acquire updated rainfall features collected by each detection unit at 100ms intervals. The main detection unit performs the same three-level anomaly removal and weighted fusion calculation as described above on the updated rainfall features of all detection units within each cycle, obtaining the updated fusion peak value Pfusion and fusion variance Varfusion. Then, the main detection unit determines whether the sleep condition is met: it maintains a rainless timer, initially set to 0. If Pfusion < 0.05V and Varfusion < 0.01V in the current cycle... If the rainless timer reaches 60 seconds (i.e., 600 consecutive rainless cycles), the main detection unit determines that the sleep condition is met. At this time, the main detection unit broadcasts a "sleep command" to all slave detection units via the target communication network. This command includes a synchronization timestamp for immediate sleep. Upon receiving the command, all slave detection units enter deep sleep mode synchronously with the main detection unit: shutting down the infrared transmitting component, signal receiving channel, differential processing circuit, and most communication modules, retaining only an ultra-low power wake-up monitoring circuit (power consumption ≤ 8μA). The main detection unit itself also enters the same sleep mode. In sleep mode, the wake-up monitoring circuit of each detection unit continuously monitors rainfall changes in the detection area. When the wake-up monitoring circuit of any detection unit (whether main or slave) detects a peak value P ≥ 0.05V in the collected updated rainfall characteristics, the detection unit immediately wakes up from sleep mode and restores all functions. Then, the detection unit sends a "wake-up signal" to the main detection unit via the target communication network. Upon receiving the wake-up signal, the master detection unit immediately broadcasts a "wake-up command" to all slave detection units, simultaneously waking up all other detection units. All detection units complete the wake-up process within 10ms, resuming normal rainfall synchronization, data upload, anomaly removal, weighted fusion, and wiper control procedures. If the master detection unit is not yet awake at the time of wake-up, the awakened slave detection unit first sends a signal to trigger the master detection unit to wake up, and then the master detection unit wakes up the remaining nodes.

[0111] This embodiment also provides a windshield wiper control device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0112] This embodiment provides a windshield wiper control device, such as... Figure 7 As shown, it includes: The acquisition module 71 is used to acquire the rainfall characteristics collected by each detection unit in the rainfall detection unit array, wherein the rainfall detection unit array is deployed on the vehicle and includes a main detection unit and at least one slave detection unit. The detection module 72 is used to detect the validity of the rainfall characteristics collected by each detection unit and obtain the valid rainfall characteristics corresponding to each detection unit. The fusion module 73 is used to fuse the effective rainfall features corresponding to each detection unit to obtain fused rainfall features; The control module 74 is used to control the vehicle's windshield wipers to perform wiping actions that integrate rainfall characteristics.

[0113] In this embodiment of the application, the acquisition module 71 is specifically used to acquire the original rainfall signal collected by each detection unit; acquire the compensation coefficient corresponding to each detection unit, wherein the compensation coefficient is used to compensate the original rainfall signal collected by the corresponding detection unit to eliminate the baseline drift of each detection unit during the acquisition process; for each detection unit, the original rainfall signal is compensated according to the compensation coefficient to obtain a compensated rainfall signal, and the corresponding rainfall features are extracted from the compensated rainfall signal.

[0114] In this embodiment, the acquisition module 71 is specifically used to acquire the deployment status of the main detection unit and at least one slave detection unit; when both deployment statuses are in the deployed state, the main detection unit broadcasts a network request and establishes a target communication network with at least one slave detection unit that responds to the network request; the main detection unit controls the target communication network to enter the calibration state; the main detection unit acquires the reference signal collected by the main detection unit and at least one slave detection unit in the calibration state, and calculates the compensation coefficient of each detection unit based on the reference signal.

[0115] In this embodiment of the application, the detection module 72 is specifically used to acquire abnormal factors that affect the effectiveness of rainfall detection by the detection unit, and acquire at least one judgment index corresponding to each abnormal factor; identify abnormal features in the rainfall features based on at least one judgment index, and remove the abnormal features from the rainfall features to obtain effective rainfall features.

[0116] In this embodiment of the application, the abnormal factors include: gap abnormality, occlusion abnormality, and fault abnormality; the judgment indicators include the signal variance of the rainfall signal, the signal peak value of the rainfall signal, and the communication status of the detection unit. The signal variance is the judgment indicator corresponding to the gap abnormality type, the signal peak value is the judgment indicator corresponding to the occlusion abnormality type, and the communication status is the judgment indicator corresponding to the fault abnormality type.

[0117] In this embodiment of the application, the detection module 72 is specifically used to calculate the average variance based on the signal variance in each rainfall feature when the abnormal factor is intermittent abnormality, to obtain the baseline variance; identify the first abnormal feature in the rainfall feature whose signal variance is greater than the baseline variance, and remove the first abnormal feature from the rainfall feature to obtain the effective rainfall feature.

[0118] In this embodiment, the detection module 72 is specifically used to identify a second abnormal feature in the rainfall features whose signal peak value is a first preset value when the abnormal factor is occlusion abnormality; for other detection units besides the detection unit corresponding to the candidate abnormal feature, calculate the average value of the signal peak value in their corresponding rainfall features to obtain a reference peak value; detect whether the reference peak value is greater than or equal to the second preset value; when the reference peak value is greater than or equal to the second preset value, remove the second abnormal feature from the rainfall features to obtain the effective rainfall features.

[0119] In this embodiment of the application, the detection module 72 is specifically used to identify a third abnormal feature in the rainfall features that has missing or excessive data for a continuous preset number of cycles when the abnormal factor is a fault abnormality, and remove the third abnormal feature from the rainfall features to obtain the effective rainfall features.

[0120] In this embodiment, the detection module 72 is specifically configured to: calculate the average variance based on the signal variance of each rainfall feature to obtain a baseline variance when the abnormal factors include gap anomaly type, shading anomaly type, and fault anomaly type; identify a first abnormal feature in the rainfall feature whose signal variance is greater than the baseline variance, and remove the first abnormal feature from the rainfall feature to obtain a first remaining feature; identify a second abnormal feature in the first remaining feature whose signal peak value is a first preset value; calculate the average value of the signal peak values ​​in the corresponding first remaining feature for other detection units besides the detection units corresponding to the candidate abnormal features to obtain a baseline peak value; detect whether the baseline peak value is greater than or equal to a second preset value; when the baseline peak value is greater than or equal to the second preset value, remove the second abnormal feature from the first remaining feature to obtain a second remaining feature; identify a third abnormal feature in the second remaining feature that has data missing or data exceeding limits for a consecutive preset number of periods, remove the third abnormal feature from the second remaining feature to obtain an effective rainfall feature.

[0121] In this embodiment, each detection unit in the rain detection unit array is attached to the vehicle via a magnetic structure. The magnetic structure includes a magnetic base, a magnetic circuit assembly, a sealing assembly, and a buffer pad. The magnetic circuit assembly is disposed inside the magnetic base and is used to adjust the magnetic attraction force. The sealing assembly is disposed at the edge of the magnetic base and is used to form a seal with the vehicle. The buffer pad is disposed on the magnetic base and is used to conform to the vehicle to form a buffer.

[0122] Please see Figure 8 , Figure 8 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 8 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system).

[0123] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0124] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0125] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0126] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0127] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0128] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0129] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A windshield wiper control method, characterized in that, The method includes: The rainfall characteristics collected by each detection unit in the rainfall detection unit array are acquired, wherein the rainfall detection unit array is deployed on a vehicle and includes a main detection unit and at least one slave detection unit; The validity of the rainfall characteristics collected by each detection unit is tested to obtain the valid rainfall characteristics corresponding to each detection unit; The effective rainfall features corresponding to each detection unit are fused to obtain the fused rainfall features; The vehicle's windshield wipers are controlled to perform the wiping action corresponding to the fused rainfall characteristics.

2. The method according to claim 1, characterized in that, The acquisition of rainfall features collected by each detection unit in the rainfall detection unit array includes: Acquire the raw rainfall signal collected by each of the detection units; Obtain the compensation coefficient corresponding to each detection unit, wherein the compensation coefficient is used to compensate the original rainfall signal collected by the corresponding detection unit in order to eliminate the baseline drift of each detection unit during the acquisition process; For each detection unit, the original rainfall signal is compensated according to the compensation coefficient to obtain a compensated rainfall signal, and the corresponding rainfall features are extracted from the compensated rainfall signal.

3. The method according to claim 2, characterized in that, The step of obtaining the compensation coefficient corresponding to each detection unit includes: Obtain the deployment status of the main detection unit and the at least one slave detection unit; When all the deployment states are in the deployed state, the main detection unit broadcasts a networking request and establishes a target communication network with the at least one slave detection unit that responds to the networking request. The main detection unit controls the target communication network to enter a calibration state. The reference signals collected by the main detection unit and the at least one slave detection unit in the calibration state are acquired, and the compensation coefficient of each detection unit is calculated based on the reference signals.

4. The method according to claim 1, characterized in that, The step of performing validity testing on the rainfall characteristics collected by each detection unit to obtain the valid rainfall characteristics corresponding to each detection unit includes: Identify the abnormal factors that affect the effectiveness of rainfall detection by the detection unit, and obtain at least one judgment index corresponding to each of the abnormal factors; Based on the at least one determination index, abnormal features in the rainfall characteristics are identified, and the abnormal features are removed from the rainfall characteristics to obtain the effective rainfall characteristics.

5. The method according to claim 4, characterized in that, The abnormal factors include: gap abnormalities, obstruction abnormalities, and fault abnormalities; The judgment indicators include the signal variance of the rainfall signal, the signal peak value of the rainfall signal, and the communication status of the detection unit. The signal variance is the judgment indicator corresponding to the gap anomaly type, the signal peak value is the judgment indicator corresponding to the occlusion anomaly type, and the communication status is the judgment indicator corresponding to the fault anomaly type.

6. The method according to claim 5, characterized in that, The step of identifying abnormal features in the rainfall characteristics based on the at least one judgment index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics includes: When the abnormal factor is an intermittent abnormality, the average variance is calculated based on the signal variance in each rainfall feature to obtain the baseline variance; Identify the first anomalous feature in the rainfall characteristics whose signal variance is greater than the baseline variance, and remove the first anomalous feature from the rainfall characteristics to obtain the effective rainfall characteristics.

7. The method according to claim 5, characterized in that, The step of identifying abnormal features in the rainfall characteristics based on the at least one judgment index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics includes: When the abnormal factor is an obstruction abnormality, a second abnormal feature in the rainfall characteristics is identified whose signal peak value is a first preset value; For detection units other than those corresponding to candidate anomaly features, calculate the average value of the signal peak value in their corresponding rainfall features to obtain the baseline peak value; Detect whether the reference peak value is greater than or equal to the second preset value; When the baseline peak value is greater than or equal to the second preset value, the second abnormal feature is removed from the rainfall feature to obtain the effective rainfall feature.

8. The method according to claim 5, characterized in that, The step of identifying abnormal features in the rainfall characteristics based on the at least one judgment index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics includes: When the abnormal factor is a fault, a third abnormal feature is identified in the rainfall feature that has data missing or data exceeding the limit for a continuous preset number of cycles. The third abnormal feature is removed from the rainfall feature to obtain the effective rainfall feature.

9. The method according to claim 5, characterized in that, The step of identifying abnormal features in the rainfall characteristics based on the at least one judgment index, and removing the abnormal features from the rainfall characteristics to obtain the effective rainfall characteristics includes: When the abnormal factors include gap abnormality type, shading abnormality type and fault abnormality type, the average variance is calculated based on the signal variance in each rainfall feature to obtain the baseline variance; Identify the first anomalous feature in the rainfall features whose signal variance is greater than the benchmark variance, and remove the first anomalous feature from the rainfall features to obtain the first remaining feature; Identify a second abnormal feature among the first remaining features whose signal peak value is a first preset value; For detection units other than those corresponding to candidate abnormal features, calculate the average value of the signal peaks in the corresponding first remaining features to obtain the baseline peak value; Detect whether the reference peak value is greater than or equal to the second preset value; When the benchmark peak value is greater than or equal to the second preset value, the second abnormal feature is removed from the first remaining feature to obtain the second remaining feature; Identify a third abnormal feature in the second remaining features that has data missing or data exceeding limits for a continuous preset number of periods, and remove the third abnormal feature from the second remaining features to obtain the effective rainfall feature.

10. The method according to claim 1, characterized in that, Each detection unit in the rainfall detection unit array is magnetically attached to the vehicle via a magnetic structure. The magnetic structure includes a magnetic base, a magnetic circuit assembly, a sealing assembly, and a buffer pad. The magnetic circuit assembly is located inside the magnetic base and is used to adjust the magnetic attraction force. The sealing assembly is located at the edge of the magnetic base and is used to form a seal with the vehicle. The buffer pad is located on the magnetic base and is used to conform to the vehicle to form a buffer.

11. A vehicle, characterized in that, The vehicle includes a controller and a windshield wiper. The controller includes a memory and a processor, which are communicatively connected. The memory stores computer instructions, and the processor executes the computer instructions to perform the method of any one of claims 1 to 10.

Images