Multi-channel synchronous detection method and system for electrical performance of automobile lamp wire harness

Abstract

The invention discloses a multi-channel synchronous detection method and system for the electrical performance of an automobile lamp wire harness, and relates to the field of multi-channel synchronous detection of the electrical performance, and the method comprises the steps: constructing an electrical performance detection list of the automobile lamp wire harness; establishing a hierarchical level of electrical performance detection, having a Boolean rule with reserved skipping, and dynamically regulating and controlling the type based on the rule to detect whether skipping exists; constructing a software switch configuration matrix, and dynamically regulating and controlling an electrical performance detection sequence based on the matrix; connecting an interface for detecting the electrical performance of the automobile lamp wire harness with a hardware detection platform, and acquiring a performance detection result; and constructing a comprehensive evaluation neural network model, outputting a reason label and probability of type detection skipping, and judging whether the skipped type is detected again or not. The method has the advantages that high safety and high reliability of electrical performance detection of the automobile lamp wire harness under the full-automatic condition are achieved, and the industrial pain point of result misalignment caused by misjudgment of traditional automatic detection is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of multi-channel synchronous testing of electrical performance, specifically to a method and system for multi-channel synchronous testing of the electrical performance of automotive lighting harnesses. Background Technology

[0002] As the nerve center of vehicle lighting, signaling, and intelligent driving assistance systems, the electrical performance reliability of automotive lighting harnesses directly affects driving safety and overall vehicle function. In the manufacturing quality control system, final inspection of wiring harnesses is a crucial step. Traditional testing methods mainly rely on manual operation using basic instruments such as multimeters, megohmmeters, and continuity testers to measure key parameters such as conductivity, resistance, insulation resistance, and withstand voltage point by point. This approach is labor-intensive, inefficient, and prone to misjudgments or omissions due to human fatigue or operational errors. To improve efficiency and consistency, automated testing technology has gradually become the industry standard. The current mainstream automated solution is an automated testing platform built on a "programmable power supply + multi-channel switch matrix + measuring instruments" system. This system controls the switch matrix to switch connection paths through pre-programmed test sequences, sequentially completing the measurements of preset items and automatically recording data and determining pass / fail status. This method achieves "one-time wiring, automatic completion of multiple tests," significantly increasing throughput while effectively reducing human error.

[0003] Furthermore, to meet the testing requirements of complex automotive lighting wiring harnesses (such as intelligent headlight wiring harnesses with CAN / LIN bus communication functions), integrated communication testing modules have begun to be embedded in the system. At the same time, some advanced systems introduce simple logical judgments. For example, when the current lead detection detects an open circuit, it automatically skips the subsequent load test of that circuit to avoid meaningless operations. This reflects basic programmed decision-making thinking. Overall, existing technologies have evolved from purely manual measurement to a testing mode centered on automation and programming, constituting the mainstream solution in current industrial production.

[0004] Despite significant advancements in existing automated testing technologies, their core logic remains essentially "automation executed according to a fixed script." When faced with complex and dynamic industrial environments, this reveals numerous inherent flaws, hindering the achievement of true intelligence and high reliability. For instance, while existing systems can execute tests sequentially, they lack the ability to dynamically assess the potential safety risks arising from the testing actions themselves. For example, when a short circuit is detected in a loop, the system may still mechanically execute subsequent withstand voltage tests or high-current load tests. This can easily lead to overheating, arcing, or even fire at the short circuit point, posing a serious threat to products and equipment. The system's "skip" logic is often... The current technology is based on a preset and simplistic approach, failing to provide refined and tiered risk avoidance based on real-time detected fault modes and their severity. Secondly, existing technologies lack the ability to self-reflect on and diagnose the reliability of test results, only able to output "pass / fail" criteria without being able to determine the credibility of these criteria themselves. When test results are abnormal due to system-specific issues such as poor contact of testing fixtures, interface oxidation, fluctuations in excitation sources, or electromagnetic interference, the system will indiscriminately classify them as "product failures," resulting in a large number of "false positives." This not only leads to the incorrect scrapping of qualified products but also masks the need for maintenance of the testing equipment itself, making it difficult to trace the root cause and optimize the process. Summary of the Invention

[0005] To address the aforementioned technical problems, this paper provides a multi-channel synchronous testing method and system for the electrical performance of automotive lighting wiring harnesses. This technical solution solves the problems mentioned in the background art, such as the lack of ability to dynamically assess the safety risks that may be caused by the testing action itself, the fact that the "skip" logic of the system is usually preset and simple, and the inability to carry out refined and hierarchical risk avoidance based on the fault modes and their severity detected in real time; and the lack of self-reflection and diagnostic ability on the reliability of the test results, which can only output "pass / fail" criteria, but cannot determine whether the criteria themselves are reliable, leading to the wrong scrapping of qualified products, and further masking the maintenance problems of the testing equipment itself, making it difficult to trace the root cause and optimize the process.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A multi-channel synchronous testing method for the electrical performance of automotive lighting wiring harnesses includes: Obtain the types and testing methods required for testing the electrical performance of automotive lighting wiring harnesses, set testing priorities according to the nature of the testing, and construct a list of testing methods for the electrical performance of automotive lighting wiring harnesses. Based on the detection priority, a hierarchical Boolean rule for electrical performance detection with reservation skipping is constructed, and the type detection is dynamically adjusted based on the rule to determine whether to skip. Based on the detection priority, the hierarchical level of electrical performance testing, and the Boolean rule of reserved skipping, a software switch configuration matrix is ​​constructed, and the order of electrical performance testing is dynamically adjusted based on this matrix; The hardware testing platform with integrated testing methods connects the interface for testing the electrical performance of automotive headlight wiring harnesses to the hardware testing platform, starts the control program of the software switch configuration matrix, and collects the performance test results. Construct a comprehensive evaluation neural network model, combine performance detection results, output the reason label and probability of skipping type detection, and determine whether to re-detect the skipped type.

[0007] Preferably, the step of constructing a hierarchical electrical performance detection hierarchy based on detection priority, with reserved skipping Boolean rules, and dynamically adjusting whether to skip type detection based on these rules specifically includes: Based on the testing priority, the failure of the preliminary test is regarded as the bottom event. The safety accident that occurs when subsequent tests are performed under the condition of the failure of the preliminary test is defined as the risk scenario, and this scenario is regarded as the top event. Based on historical or experimental data, assign objective failure probability values ​​to all bottom events; Construct a fault tree model, calculate the probability of occurrence of the top event from the probability of the bottom event upwards, and define the probability of occurrence as the safety risk coefficient from the failure of the pre-detection to the subsequent detection; For all combinations of detection types that need to be evaluated, the fault tree model is repeatedly executed, and the model output is used to construct a two-dimensional safety risk coefficient matrix. In this matrix, the rows represent the preliminary detections, the columns represent the subsequent detections, and the element values ​​in the matrix are the safety risk coefficients. Based on the safety risk coefficient, test experiments were conducted, and based on the experimental results, the safety risk coefficient was divided into three Boolean rule intervals. Among them, the first-level Boolean rule interval control instruction is 0, which means that its subsequent detection is not skipped; the second-level Boolean rule interval control instruction is 1, which means that the subsequent detection is executed conditionally, and the system will take a derating test and output a warning instruction while executing; the third-level Boolean rule interval control instruction is 2, which means that its subsequent detection is skipped and an alert instruction is output. Based on the two-dimensional safety risk coefficient matrix and combined with the three-level Boolean rule interval, a hierarchical Boolean rule with reserved skipping for electrical performance detection is constructed, and the type detection is dynamically adjusted based on the rule to determine whether to skip.

[0008] Preferably, the step of constructing a software switch configuration matrix based on detection priority, hierarchical levels of electrical performance testing, and Boolean rules for reserved skipping, and dynamically adjusting the electrical performance testing order based on this matrix specifically includes: Based on the detection priority, hierarchical structure of electrical performance testing, and Boolean rules for reserved skipping, a software switch configuration matrix is ​​constructed. This matrix is ​​a three-dimensional control structure, with its three dimensions as follows: Pre-test status: derived from real-time electrical performance test results, including pass and fail; Current pending detections: derived from the detection priority sequence; Target hardware channel: Sourced from the list of electrical performance tests for automotive headlight wiring harnesses; Each element of this matrix is ​​a control tuple whose content is dynamically generated by query hierarchy, Boolean rules with reserved skips, and security risk coefficient matrix; The tuple defines: switch on / off instructions, derating test parameters, and marking instructions. The switch on / off instructions are composed of the numbers 0 and 1, where 0 means off and 1 means on. During the detection process, the system executes the following procedures in sequence to achieve dynamic control, specifically including: Once a type of detection is completed, the system uses its result and the next detection to be performed as an index to query the software switch configuration matrix and obtain the corresponding control tuple. If the control tuple instruction is instruction 2, the system will skip the detection, record the alert, and index the next detection item. If the control tuple instruction is instruction 1, the system calls the derating test parameters, configures the corresponding excitation source and measurement range, executes the test, and records the data marked with warning. If the control tuple instruction is instruction 0, the current detection item will be executed normally; Repeat this process until all items in the detection sequence have been processed.

[0009] Preferably, the construction of the comprehensive evaluation neural network model, combining performance detection results, outputting the reason label and probability of type detection skipping, and determining whether to re-detect the skipped type specifically includes: A comprehensive evaluation neural network model is constructed. The input features of the model include: pre-detection state, safety risk coefficient, Boolean rule control instruction type, derating test parameters, and deviation values ​​of key indicators in the electrical performance test results. The model's output layer is set to a dual-output structure, consisting of the label of the reason for the detection skip and the confidence probability of the corresponding reason; The cause labels include at least: product failure, detection system error, poor tooling contact, and unstable excitation source; Based on historical data from automotive lighting harness testing, the corresponding historical electrical performance test results are collected through the automotive lighting harness electrical performance testing process. Based on historical electrical performance test results, combined with manually labeled skipping reasons, a model training sample set was constructed and divided into a training set, a validation set, and a test set. The sample set includes the pre-detection status, security risk coefficient, instruction type, detection results, and manually labeled skip reason tags under different detection scenarios; The model training sample set is preprocessed, and the preprocessing includes at least: one-hot encoding of labels, outlier removal, data normalization, and missing value imputation. For the classification task of cause labels, the cross-entropy loss function is used to calculate the difference between the model's predicted label and the manually labeled real label; With the goal of minimizing the cross-entropy loss function, the Adam optimizer is used to train the model based on the training set, and the trained model is defined as a comprehensive evaluation neural network model. Based on the comprehensive evaluation neural network model, when the model gives a high confidence probability for the detection system error or tooling contact failure label, the system can automatically trigger the instruction to re-detect the skipped detection type and highlight the unreliability of the detection result in the report; Based on the test set, the stability of the neural network model is comprehensively evaluated using performance metrics such as label accuracy, precision, recall, F1-Score, mean confidence probability, and retest trigger accuracy.

[0010] Furthermore, this solution proposes a multi-channel synchronous testing system for the electrical performance of automotive lighting harnesses, used to implement the aforementioned multi-channel synchronous testing method for the electrical performance of automotive lighting harnesses, including: The software configuration module is used to obtain the types and methods required for electrical performance testing of automotive lighting wiring harnesses, set testing priorities according to the testing nature, and construct an electrical performance testing list for automotive lighting wiring harnesses; based on the testing priorities, construct hierarchical levels for electrical performance testing and Boolean rules with reservation and skip options, and dynamically adjust whether to skip type testing based on these rules; based on the testing priorities and hierarchical levels for electrical performance testing and Boolean rules with reservation and skip options, construct a software switch configuration matrix, and dynamically adjust the electrical performance testing order based on this matrix. The hardware testing module is used to integrate the hardware testing platform for testing methods. It connects the interface for testing the electrical performance of automotive headlight wiring harnesses to the hardware testing platform, starts the control program of the software switch configuration matrix, and collects the performance testing results. The comprehensive evaluation module is used to construct a comprehensive evaluation neural network model, combine the performance detection results, output the reason label and probability of skipping type detection, and determine whether to re-detect the skipped type.

[0011] Preferably, the software configuration module includes: The detection list unit is used to obtain the types and detection methods required for the electrical performance testing of automotive lighting wiring harnesses, set detection priorities according to the nature of the tests, and construct an automotive lighting wiring harness electrical performance testing list. Boolean rule unit, which is used to construct hierarchical Boolean rules for electrical performance detection with reservation skipping based on detection priority, and dynamically adjust the type detection to skip based on the rule; The detection sequence unit is used to construct a software switch configuration matrix based on the detection priority, the hierarchical level of electrical performance detection, and Boolean rules for reserved skipping, and to dynamically adjust the electrical performance detection sequence based on the matrix.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention proposes a multi-channel synchronous testing scheme for the electrical performance of automotive lighting harnesses. Through hardware interface reuse and a software switch configuration matrix, it achieves multi-channel synchronous testing and dynamic control of the testing sequence, significantly improving testing efficiency and overcoming the limitations of traditional single-channel, fixed-sequence testing. Furthermore, by employing hierarchical Boolean rules for selective skipping and combining them with a fault tree model to quantify safety risks, it avoids safety accidents caused by blind testing and reduces data waste through derating testing. It also avoids the traditional single-mode of "continue if qualified, terminate if unqualified," increasing testing flexibility. Secondly, by introducing a neural network model into the evaluation of testing results, it enables intelligent tracing of skipped causes and automatic re-inspection, solving the pain points of difficulty in defining skipped causes and reliance on manual judgment for re-inspection in traditional testing. This improves the reliability of testing results, achieving high safety and high reliability in the fully automated testing of automotive lighting harnesses, effectively addressing the industry pain point of inaccurate results due to misjudgment in traditional automated testing. Attached Figure Description

[0013] Figure 1 This is a flowchart of a multi-channel synchronous detection method for the electrical performance of automotive lighting harnesses according to the present invention; Figure 2 The present invention provides a flowchart for constructing a hierarchical electrical performance detection system with reserved skipping Boolean rules, and dynamically adjusting the type detection skipping process based on these rules. Figure 3 The flowchart for constructing a comprehensive evaluation neural network model for this invention, outputting the reason label and probability of skipping type detection, and determining whether to re-detect the skipped type. Detailed Implementation

[0014] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0015] Reference Figure 1 As shown, a multi-channel synchronous testing method for the electrical performance of automotive lighting wiring harnesses includes: Obtain the types and testing methods required for testing the electrical performance of automotive lighting wiring harnesses, set testing priorities according to the nature of the testing, and construct a list of testing methods for the electrical performance of automotive lighting wiring harnesses. Based on the detection priority, a hierarchical Boolean rule for electrical performance detection with reservation skipping is constructed, and the type detection is dynamically adjusted based on the rule to determine whether to skip. Based on the detection priority, the hierarchical level of electrical performance testing, and the Boolean rule of reserved skipping, a software switch configuration matrix is ​​constructed, and the order of electrical performance testing is dynamically adjusted based on this matrix; The hardware testing platform with integrated testing methods connects the interface for testing the electrical performance of automotive headlight wiring harnesses to the hardware testing platform, starts the control program of the software switch configuration matrix, and collects the performance test results. Construct a comprehensive evaluation neural network model, combine performance detection results, output the reason label and probability of skipping type detection, and determine whether to re-detect the skipped type.

[0016] To achieve efficient, safe, and automated testing of the electrical performance of automotive lighting wiring harnesses, it is essential to first define the fundamental boundaries of the testing. Therefore, the first step is to identify all necessary electrical performance testing types and corresponding methods, prioritizing them based on the core principle of post-destructive testing, and constructing a standardized testing list. This serves as the foundation for all subsequent testing processes, preventing missed, incorrect, or disorganized testing. After clarifying the testing foundation, it is necessary to address safety risk management during the testing process. Therefore, the second step, based on testing priorities, uses a fault tree model to quantify the safety risks of failed pre-tests to subsequent tests, constructing hierarchical, selective skipping Boolean rules. This avoids safety incidents by skipping tests in high-risk scenarios and retains valid testing data in medium-risk scenarios through derating tests, breaking the traditional single-mode of "continue if qualified, terminate if unqualified," balancing safety and data integrity. To transform abstract testing rules into automatically executable operations, the third step involves constructing a software switch configuration matrix. This matrix serves as the core of software control, integrating the pre-test status, currently pending tests, and target hardware channels. The system dynamically generates control tuples based on key information, enabling precise switching of detection channels and intelligent control of the detection sequence. This addresses industry pain points such as channel switching errors and instruction execution lag in multi-channel synchronous detection. Secondly, after the software control logic is determined, it needs to be implemented in actual hardware. Therefore, the fourth step is to build an integrated hardware detection platform, reserving hardware interfaces to adapt to different wire harness interfaces. After self-testing to ensure equipment stability, the tested wire harness is connected and the software control program is started. The detection is executed according to matrix instructions, and data such as detection type, measured values, and pass / fail status are collected simultaneously to form a complete list of detection results, completing the implementation from "rule design" to "physical detection". Finally, to address the pain points of difficulty in identifying the causes of detection skips and reliance on manual re-inspection, the fifth step is to build a comprehensive evaluation neural network model. Based on the collected performance detection results, it intelligently identifies the root causes of detection skips and outputs confidence probabilities. It automatically triggers re-inspection for repairable problems such as detection system errors and poor tooling contact, and clearly marks irreversible problems such as product failures, forming a closed loop of "detection execution - result evaluation - problem repair", which greatly improves the reliability of detection results and ensures the accuracy of fully automated detection.

[0017] The process of obtaining the types and methods of testing required for automotive lighting wiring harness electrical performance testing, setting testing priorities based on the nature of the testing, and constructing an automotive lighting wiring harness electrical performance testing list specifically includes: The required test types for obtaining the electrical performance of the automotive lighting wiring harness to be tested include, but are not limited to: continuity, short circuit, insulation, withstand voltage, resistance, voltage drop, load, and communication. Obtain the detection method corresponding to the detection type. Detection types with the same detection method can use the same line interface. Based on the principle of destructive testing as the last resort, and combined with expert scoring, the priority of testing the electrical performance of automotive headlight wiring harnesses is set. Construct a list of electrical performance tests for automotive lighting wiring harnesses, comprising a set of data including the test types, their corresponding test methods, their qualification standards, and their test priorities.

[0018] To explain, to perform electrical performance testing on automotive lighting harnesses, it's essential to first understand which types of electrical performance indicators need to be tested, what the pass / fail standards are for each type of indicator, and what testing methods meet those standards. Therefore, pre-setting according to industry standards is necessary. Specifically, since testing methods are essentially determined by excitation signals and measured physical quantities—for example, continuity, resistance, and insulation testing can all be categorized as resistance measurements, but their excitation voltages and measurement ranges differ—they can reuse the same digital multimeter hardware. Testing can be achieved simply by configuring different excitation sources and measurement circuits. Therefore, for testing types with the same testing method, the same circuit interface can be used. Testing can be achieved simply by controlling the excitation source and measurement circuit of the testing type through software, thereby improving testing efficiency. Secondly, in multi-channel synchronous testing of automotive lighting harness electrical performance... When conducting tests, the priority of performance testing needs to be considered. The main reason is that some test types (such as withstand voltage tests) are destructive tests due to the high voltage stress applied. If they are performed first, any hidden defects in the wiring harness will cause it to be destroyed, rendering all subsequent test results invalid. Therefore, this solution prioritizes the electrical performance tests of automotive headlight wiring harnesses based on the principle of destructive testing last, combined with expert scoring. Preferred implementation cases include: forming an expert group, using the principle of destructive testing last as the core evaluation criterion, and for each test type, expert group members independently score according to this core evaluation criterion (for example, using a 10-point scale, the lower the score, the higher the priority under this criterion). The weighted average score of each test type is calculated, and the test type with the lower the destructive risk score (i.e., the lower the risk) has the higher the priority. Finally, all test types are sorted according to this logic to generate the test priority.

[0019] Reference Figure 2 As shown, the construction of a hierarchical electrical performance detection system with reserved skipping Boolean rules, and the dynamic adjustment of type detection based on these rules to determine whether to skip, specifically includes: Based on the testing priority, the failure of the preliminary test is regarded as the bottom event. The safety accident that occurs when subsequent tests are performed under the condition of the failure of the preliminary test is defined as the risk scenario, and this scenario is regarded as the top event. Based on historical or experimental data, assign objective failure probability values ​​to all bottom events; Construct a fault tree model, calculate the probability of occurrence of the top event from the probability of the bottom event upwards, and define the probability of occurrence as the safety risk coefficient from the failure of the pre-detection to the subsequent detection; For all combinations of detection types that need to be evaluated, the fault tree model is repeatedly executed, and the model output is used to construct a two-dimensional safety risk coefficient matrix. In this matrix, the rows represent the preliminary detections, the columns represent the subsequent detections, and the element values ​​in the matrix are the safety risk coefficients. Based on the safety risk coefficient, test experiments were conducted, and based on the experimental results, the safety risk coefficient was divided into three Boolean rule intervals. Among them, the first-level Boolean rule interval control instruction is 0, which means that its subsequent detection is not skipped; the second-level Boolean rule interval control instruction is 1, which means that the subsequent detection is executed conditionally, and the system will take a derating test and output a warning instruction while executing; the third-level Boolean rule interval control instruction is 2, which means that its subsequent detection is skipped and an alert instruction is output. Based on the two-dimensional safety risk coefficient matrix and combined with the three-level Boolean rule interval, a hierarchical Boolean rule with reserved skipping for electrical performance detection is constructed, and the type detection is dynamically adjusted based on the rule to determine whether to skip.

[0020] This can be explained by the fact that in the multi-channel synchronous testing of the electrical performance of automotive headlight wiring harnesses, the system needs to autonomously complete the full performance test. Therefore, while determining the testing priority, it is also necessary to consider whether a failure in a preceding test will pose a safety risk to subsequent tests. For example, if a short circuit is detected in a preceding test, during subsequent tests of the load, the applied operating current will cause the short circuit point to experience a sudden surge of current far exceeding the design value, leading to rapid local overheating. This could potentially cause the wiring harness insulation to melt, smoke, or even catch fire, resulting in a safety risk of increased circuit temperature. Therefore, this solution constructs a fault tree model to link the preceding tests... The risk of safety accidents arising from subsequent tests due to non-compliance is quantified by establishing a safety risk coefficient. Based on this coefficient, a three-level Boolean rule interval is set. During actual testing, this interval guides the tiered and selective skipping of automotive lighting wiring harness electrical performance type tests, avoiding significant safety risks and achieving autonomous multi-channel synchronous safety testing. The construction of the fault tree model is illustrated with a typical preferred case as follows: For the combination of "pre-test A is a short-circuit test, and subsequent test B is a load test," the risk scenario is defined as "the wiring harness overheating caused by performing a load test under short-circuit failure conditions." Taking "fire" as the top event, and tracing downwards, its direct cause (intermediate event) is "a large current continuously flowing through the short circuit point" and "local heat accumulation that cannot dissipate." These two events are related by an AND gate. The "large current continuously flowing through the short circuit point" is constructed by the AND gate between the bottom event "short circuit test failure" (i.e., detection A fails) and "performing load test" (i.e., the inherent property of detection B). Objective probability values ​​are assigned to the bottom events "short circuit test failure" (based on its historical failure rate) and "latent damage to the wiring harness" (based on the probability of material defects). The probability of the top event occurring is then calculated using AND gate logic (probability multiplication). This probability value is the safety risk coefficient of the test combination (short circuit test, load test), and it is filled into the corresponding position of the safety risk coefficient matrix (row = short circuit test, column = load test). The test experiment specifically includes: since the safety risk coefficient is a probability value with a value range of [0,1], at the beginning of the experiment, this range is first divided into three equal ranges. For each test combination (A,B), under the condition that test A is artificially set to be unqualified, test B is executed, and physical quantities that can directly reflect the safety status (such as 'maximum temperature of the wiring harness', 'peak current of the circuit', or 'smoke generation of the insulation layer') are monitored. The objective criteria for defining the interval are: The upper bound of the first-level interval (instruction 0): corresponds to the highest risk factor where all observable physical quantities do not exceed their safety design specifications; The upper bound of the second-level interval (instruction 1): corresponds to the highest risk coefficient of reversible physical anomalies detected by the system (such as excessive temperature but without causing permanent damage); The third level (instruction 2): Any risk level that would trigger an irreversible safety event (such as smoke, fire, or insulation breakdown) is classified into this level. Based on the above experimental observations, the initial equal interval boundaries are dynamically adjusted and calibrated to define three specific intervals for control decisions. Based on this scheme, a two-dimensional safety risk coefficient matrix for controlling whether to skip detection type and a Boolean rule logic control matrix with marking instructions can be obtained. This provides important control rules for subsequent multi-channel synchronous detection of automotive lighting harness electrical performance. The derating test refers to implementing it by systematically reducing test stress or shortening test duration. Its derating gradient follows the principle of proportional or step derating. To implement the derating operation, the system pre-stores or calculates a derating parameter lookup table in real time. The generation logic of this table is as follows: using the safety risk coefficient as the input index, the output is the derating parameters for different detection types. The tuple includes at least: current execution ratio, voltage execution ratio, and time execution ratio. The assignment rule for the derating parameter tuple is: the higher the safety risk coefficient, the greater the reduction of the actual test parameters relative to the standard test parameters, and therefore the lower the value of each execution ratio (for example, the value of each execution ratio = 1 - safety risk coefficient). The standard test parameters are the initially set test parameters, and the execution ratio refers to the ratio of the actual parameters to the standard parameters. When the system executes instruction 1, it queries this table to obtain the derating parameter tuple based on the currently calculated safety risk coefficient, and configures the excitation source and measurement range accordingly.

[0021] The process of constructing a software switch configuration matrix based on detection priority, hierarchical electrical performance testing, and Boolean rules for reserved skipping, and dynamically adjusting the electrical performance testing order based on this matrix, specifically includes: Based on the detection priority, hierarchical structure of electrical performance testing, and Boolean rules for reserved skipping, a software switch configuration matrix is ​​constructed. This matrix is ​​a three-dimensional control structure, with its three dimensions as follows: Pre-test status: derived from real-time electrical performance test results, including pass and fail; Current pending detections: derived from the detection priority sequence; Target hardware channel: Sourced from the list of electrical performance tests for automotive headlight wiring harnesses; Each element of this matrix is ​​a control tuple whose content is dynamically generated by query hierarchy, Boolean rules with reserved skips, and security risk coefficient matrix; The tuple defines: switch on / off instructions, derating test parameters, and marking instructions. The switch on / off instructions are composed of the numbers 0 and 1, where 0 means off and 1 means on. During the detection process, the system executes the following procedures in sequence to achieve dynamic control, specifically including: Once a type of detection is completed, the system uses its result and the next detection to be performed as an index to query the software switch configuration matrix and obtain the corresponding control tuple. If the control tuple instruction is instruction 2, the system will skip the detection, record the alert, and index the next detection item. If the control tuple instruction is instruction 1, the system calls the derating test parameters, configures the corresponding excitation source and measurement range, executes the test, and records the data marked with warning. If the control tuple instruction is instruction 0, the current detection item will be executed normally; Repeat this process until all items in the detection sequence have been processed.

[0022] This can be explained by the fact that after obtaining the detection priority and hierarchical structure of electrical performance testing, as well as the Boolean rules for reserved skipping, a software controller is needed to comprehensively regulate the electrical performance testing sequence. This software controller, according to input instructions, adjusts the opening and closing of hardware connections to achieve the autonomous testing sequence of the automotive headlight wiring harness's electrical performance. The three-dimensional software switch configuration matrix is ​​the core control carrier of this software controller. The index order of the three-dimensional software switch configuration matrix is ​​"previous detection state → current pending detection → target hardware channel," where the current pending detection and the target hardware channel have a one-to-one correspondence, meaning one pending detection item matches one target hardware channel. This ensures accurate switching of detection channels, solving the core problem of "how to accurately locate the control logic," and avoiding problems such as incorrect channel switching and disordered detection sequence during multi-channel testing. The control tuple is dynamically generated by the system in real time, with the generation trigger condition being the completion of each item. During the testing task, the generation logic uses the current pre-test status, the next test type to be executed, and the corresponding target hardware channel as indexes to query the hierarchical Boolean rules and safety risk coefficient matrix that retain skipped items, extract the corresponding on / off instructions, derating test parameters, and marking instructions, and combine them to form a control tuple. When the safety risk coefficient matrix is ​​updated, the control tuple is automatically updated synchronously. The real-time dynamic generation mechanism of the control tuple ensures that the testing process can adapt to changes in real-time testing results. For example, when the pre-test status changes from "qualified" to "unqualified", the tuple will be updated synchronously, thereby adjusting the subsequent testing strategy to adapt to the dynamic requirements of wiring harness testing. The explicit composition of the derating test parameters provides a specific basis for the execution of instruction 1, avoiding the blindness of derating operations. Through this core control carrier and dynamic execution process, fully automated multi-channel synchronous testing of automotive lighting wiring harness electrical performance is achieved, ensuring the stability of testing between different performance types.

[0023] The hardware testing platform of the integrated testing method connects the interface for testing the electrical performance of automotive headlight wiring harnesses to the hardware testing platform, starts the control program of the software switch configuration matrix, and collects the performance test results, specifically including: Based on the detection method, the hardware devices are integrated into a comprehensive hardware detection platform, which reserves hardware interfaces for detection of different electrical performance types. Before connecting the automotive headlight wiring harness, start the hardware testing platform and execute the platform's self-test process to ensure that the equipment itself is in normal condition. The self-test process refers to the rapid continuity and short-circuit checks of the test channels inside the platform by controlling the switch matrix without connecting it to the test harness. After the platform self-test passes, connect the car headlight wiring harness interface to the hardware interface of the hardware testing platform, and start the electrical performance test by loading the corresponding test list and software switch configuration matrix. Obtain the electrical performance test results of the automotive headlight wiring harness and generate an electrical performance test result list, wherein the list includes: test type, measured value, pass status and corresponding marking instruction record.

[0024] Explained as follows: After setting up the software control program, it is necessary to build a hardware environment for testing the electrical performance of automotive lighting harnesses. By integrating the hardware devices for various testing methods into a comprehensive hardware testing platform, all performance types can be tested through this platform. Since testing types with the same testing method can use the same line interface, the number of hardware devices required in the actual testing process is not large, and the corresponding hardware interfaces are not very complex. Therefore, after building the comprehensive hardware testing platform, this solution ensures the stability of the platform's functions through a platform self-test process. When the platform is stable, it loads the corresponding test list and software switch configuration matrix, starts electrical performance testing, and generates an electrical performance test result list, thereby achieving multi-channel synchronous testing of the electrical performance of automotive lighting harnesses.

[0025] Reference Figure 3 As shown, the construction of the comprehensive evaluation neural network model, which outputs the reason label and probability of skipping type detection, and determines whether to re-detect the skipped type, specifically includes: A comprehensive evaluation neural network model is constructed. The input features of the model include: pre-detection state, safety risk coefficient, Boolean rule control instruction type, derating test parameters, and deviation values ​​of key indicators in the electrical performance test results. The model's output layer is set to a dual-output structure, consisting of the label of the reason for the detection skip and the confidence probability of the corresponding reason; The cause labels include at least: product failure, detection system error, poor tooling contact, and unstable excitation source; Based on historical data from automotive lighting harness testing, the corresponding historical electrical performance test results are collected through the automotive lighting harness electrical performance testing process. Based on historical electrical performance test results, combined with manually labeled skipping reasons, a model training sample set was constructed and divided into a training set, a validation set, and a test set. The sample set includes the pre-detection status, security risk coefficient, instruction type, detection results, and manually labeled skip reason tags under different detection scenarios; The model training sample set is preprocessed, and the preprocessing includes at least: one-hot encoding of labels, outlier removal, data normalization, and missing value imputation. For the classification task of cause labels, the cross-entropy loss function is used to calculate the difference between the model's predicted label and the manually labeled real label; With the goal of minimizing the cross-entropy loss function, the Adam optimizer is used to train the model based on the training set, and the trained model is defined as a comprehensive evaluation neural network model. Based on the comprehensive evaluation neural network model, when the model gives a high confidence probability for the detection system error or tooling contact failure label, the system can automatically trigger the instruction to re-detect the skipped detection type and highlight the unreliability of the detection result in the report; Based on the test set, the stability of the neural network model is comprehensively evaluated using performance metrics such as label accuracy, precision, recall, F1-Score, mean confidence probability, and retest trigger accuracy.

[0026] The explanation is that the automotive lighting wiring harness electrical performance testing process can effectively achieve fully automated multi-channel synchronous testing of automotive lighting wiring harness electrical performance, ensuring stable testing across different performance types. However, this process may be affected by various factors such as product failure, testing system errors, poor tooling contact, and unstable excitation sources. These factors can lead to inaccurate electrical performance test results during actual testing, increasing safety risks and causing subsequent types of testing to be skipped, thus affecting the final evaluation of the multi-channel synchronous testing of automotive lighting wiring harness electrical performance. Therefore, this solution leverages the powerful reasoning capabilities of a neural network model, using all test results and process data as input. The deviation value of key indicators in the electrical performance test results is the core quantitative basis for the model to achieve accurate root cause analysis. Specifically, this deviation value includes the difference or ratio between the measured values ​​and standard values ​​of indicators such as voltage drop, loop resistance, and test current. It directly reflects the degree of deviation between the test data and theoretical physical rules. By utilizing the correlation between various performance characteristics, such as the clear physical rules between resistance, voltage, and current, the reliability of each type of test result can be evaluated, thereby determining whether retesting is necessary. To ensure the integrity and reliability of the system's testing, the manual labeling logic is as follows: When historical records show that a "non-conforming" wiring harness is confirmed to be a qualified product after review or process analysis, the labeling is further subdivided based on the testing equipment's troubleshooting results. Specifically, if the issue is due to instrument precision deviation or program configuration errors, it is labeled as a testing system error; if it is due to poor contact of the tooling probe or interface oxidation, it is labeled as poor tooling contact; if it is due to fluctuations in the load power supply or test current output, it is labeled as an unstable excitation source. When historical records confirm that the product does indeed have a fault (such as poor wiring harness terminal crimping or conductor damage), it is labeled as a product fault. Here, "product fault" refers to a failure in electrical performance caused by defects in the automotive headlight wiring harness itself, which has been confirmed through comprehensive analysis by the neural network model, after ruling out errors in the testing system itself. The "high confidence probability" refers to the situation where the confidence probability value corresponding to the cause label output by the model exceeds a preset threshold. This preset threshold can be determined by fitting a validation set using a cross-validation method to adapt to the accuracy requirements of different testing scenarios. As a preferred implementation case, the core parameters of the comprehensive evaluation neural network model are set as shown in Table 1. Table 1. Core parameters for comprehensive evaluation of neural network models

[0027] This case study uses historical electrical performance testing results as a foundation, combined with manually labeled skip reasons, to construct a model training sample set. This set is divided into training, validation, and test sets in an 8:1:1 ratio. The model employs a three-layer fully connected MLP structure with 128, 64, and 32 neurons respectively, forming a progressively refined feature extraction architecture. The hidden layers use the ReLU activation function to accelerate convergence and mitigate gradient vanishing. The output layer uses the Softmax activation function to generate a normalized probability distribution for four categories. Based on this output, two key decision pieces of information are obtained: the category with the highest probability is selected as the skip reason label, and this highest probability value is used as the confidence probability of the corresponding reason. Training employs the cross-entropy loss function, with differentiated category weights set within the loss function. Product malfunction is assigned the highest weight of 1.5, and unstable stimulus sources are assigned a relatively high weight of 1.2. The aim is to maximize the recall rate of these two high-risk issues, ensuring that safety hazards and systemic equipment problems are not missed. The detection system error and tooling contact failure categories maintain a baseline weight of 1.0 to avoid the model focusing too much on occasional problems and affecting the accuracy of key defect discrimination. The Adam optimizer (learning rate 0.001) is used for parameter updates. The training is set to 100 epochs with a batch size of 32. An early stopping mechanism (patience value = 10) is enabled during training. When the validation loss does not decrease for 10 consecutive epochs, the training is terminated early to avoid invalid iterations. A preset threshold of 0.85 is set after the model output layer. When the model's predicted probability for the "detection system error" or "tooling contact failure" category exceeds this threshold, the re-detection process is automatically triggered. This threshold effectively controls the frequency of false triggers while ensuring the reliability of re-detection, achieving a balance between automated detection and manual review. As a preferred implementation case, this solution evaluates the stability and accuracy of the neural network model using performance metrics such as label accuracy, precision, recall, F1-Score, mean confidence probability, and re-detection trigger accuracy. The specific performance metrics are defined in Table 2. Table 2 defines the performance indicators for comprehensively evaluating the stability and accuracy of neural network models.

[0028] Table 2 defines six core performance metrics for evaluating the stability and accuracy of a comprehensive neural network model. These metrics quantify the model's performance in the automotive headlight wiring harness fault classification task from different dimensions. Label accuracy, as a macro-level evaluation metric of the overall model performance, reflects the model's comprehensive ability to correctly classify all samples. This metric is simple and intuitive to calculate and is suitable for rapid evaluation of the model's basic performance. Precision and recall constitute a pair of complementary evaluation perspectives. Precision focuses on the proportion of results predicted by the model as belonging to a certain category that actually belong to that category, emphasizing the model's ability to avoid misclassifying other categories as that category. Recall, on the other hand, focuses on the proportion of samples that actually belong to that category that are correctly identified by the model, emphasizing the model's ability to avoid misclassifying that category as other categories. In industrial inspection scenarios, high precision can reduce invalid re-inspections caused by false positives, while high recall can avoid quality risks caused by missed detections. The F1-Score, as the harmonic mean of precision and recall, provides a single comprehensive measure of performance. The evaluation metrics are as follows: The mean confidence probability is calculated by statistically analyzing the average confidence score of correctly identified samples in each category. This verifies the reliability of the model's internal confidence mechanism and assesses the model's certainty in making correct judgments. A high mean score indicates high internal consistency in making correct judgments, enhancing the credibility of the output results. The re-inspection trigger accuracy is a specific evaluation metric designed for specific business logic, only counting samples requiring re-inspection due to detection system errors and tooling contact defects. This metric measures the accuracy of the model's decision on whether to initiate the re-inspection process. These six metrics together constitute a multi-level, multi-dimensional model evaluation system, covering both the general evaluation needs of classification tasks and customized designs for the specific requirements of automotive wiring harness inspection scenarios, providing comprehensive and objective performance data for model optimization and deployment. In practical implementation cases, the following standards can be used as a reference for evaluating the stability and accuracy of the system through these six performance indicators: Based on the test set, the test set is divided into 5 independent time periods according to production weeks, and each time period is tested independently using the pre-trained model parameters. If the results show that: label accuracy fluctuation ≤ 2.5%, four-class precision / recall fluctuation ≤ 3.0%, F1-Score mean fluctuation ≤ 2.8%, confidence probability mean fluctuation ≤ 2.0%, and re-examination trigger accuracy fluctuation ≤ 1.8%, and all indicators fluctuate less than 3%, and the product failure recall rate fluctuates less than 2.0% in each time period, and the absolute value does not fall below the safety threshold of 95%, then the model is considered to have temporal stability. The fluctuation value is calculated as (maximum value - minimum value) / average value × 100%. Add 5% Gaussian noise to the core features of the test set (voltage drop, loop resistance, safety risk coefficient, etc.) to simulate random fluctuations in industrial field data. If the test results show that: the tag accuracy decreases by ≤1.5%, the four-class precision / recall decreases by ≤2.0%, the mean F1-Score decreases by ≤1.8%, the mean confidence probability decreases by ≤1.2%, and the re-detection trigger accuracy decreases by ≤1.0%, and all indicators decrease by <2%, then the model's data perturbation stability is indicated. With a fixed model structure and hyperparameters, the model is trained 5 times using 5 different random seeds (100, 200, 300, 400, 500). If the results show that the standard deviation of label accuracy is ≤0.8%, the standard deviation of precision / recall for the four classes is ≤0.9%, the standard deviation of the mean of F1-Score is ≤0.7%, the standard deviation of the mean of confidence probability is ≤0.6%, the standard deviation of retest trigger accuracy is ≤0.5%, and the standard deviation of all indicators is <1%, then the model's training randomness is stable. It is understood that the above values ​​(e.g., 3.0%, 2.0%, 1.0%) and the safety baseline (95%) are recommended values ​​in the preferred embodiment. In actual deployment, they can be appropriately adjusted according to the quality system requirements, detection accuracy targets and risk tolerance of the specific production line. The stability and accuracy of the neural network model are comprehensively reflected and evaluated through the model's time stability, data perturbation stability and training randomness stability. It is important to note that during model training and evaluation, the training set data is used to iteratively update model parameters based on the cross-entropy loss function and the Adam optimizer, learning the mapping relationship between input features and skip reason labels. The validation set is used to monitor the training process and optimize the model state. Specifically, after each training round, the validation set is used to evaluate the model's loss value and initial performance. If the validation loss does not decrease for 10 consecutive rounds, an early stopping mechanism is triggered to terminate training, preventing overfitting and determining the optimal model parameters. The test set is used to complete performance evaluation and stability verification. This involves calculating six core metrics (label accuracy, precision, recall, F1-Score, mean confidence probability, and re-examination trigger accuracy) to objectively evaluate the model's classification performance. Based on the test set, a three-dimensional verification (time stability, data perturbation stability, and training randomness stability) is conducted to verify the model's performance consistency under different operating conditions. The evaluation results of the test set are the final basis for determining whether the model meets industrial deployment requirements.

[0029] Furthermore, based on the same inventive concept as the aforementioned multi-channel synchronous testing method for the electrical performance of automotive lighting harnesses, this solution proposes a multi-channel synchronous testing system for the electrical performance of automotive lighting harnesses, comprising: The software configuration module is used to obtain the types and methods required for electrical performance testing of automotive lighting wiring harnesses, set testing priorities according to the testing nature, and construct an electrical performance testing list for automotive lighting wiring harnesses; based on the testing priorities, construct hierarchical levels for electrical performance testing and Boolean rules with reservation and skip options, and dynamically adjust whether to skip type testing based on these rules; based on the testing priorities and hierarchical levels for electrical performance testing and Boolean rules with reservation and skip options, construct a software switch configuration matrix, and dynamically adjust the electrical performance testing order based on this matrix. The hardware testing module is used to integrate the hardware testing platform for testing methods. It connects the interface for testing the electrical performance of automotive headlight wiring harnesses to the hardware testing platform, starts the control program of the software switch configuration matrix, and collects the performance testing results. The comprehensive evaluation module is used to construct a comprehensive evaluation neural network model, combine the performance detection results, output the reason label and probability of skipping type detection, and determine whether to re-detect the skipped type. The software configuration module includes: The detection list unit is used to obtain the types and detection methods required for the electrical performance testing of automotive lighting wiring harnesses, set detection priorities according to the nature of the tests, and construct an automotive lighting wiring harness electrical performance testing list. Boolean rule unit, which is used to construct hierarchical Boolean rules for electrical performance detection with reservation skipping based on detection priority, and dynamically adjust the type detection to skip based on the rule; The detection sequence unit is used to construct a software switch configuration matrix based on the detection priority, the hierarchical level of electrical performance detection, and Boolean rules for reserved skipping, and to dynamically adjust the electrical performance detection sequence based on the matrix.

[0030] In summary, the advantages of this invention are: it achieves high safety and high reliability in the electrical performance testing of automotive lighting harnesses under fully automated conditions, effectively solving the industry pain point of inaccurate results due to misjudgment in traditional automated testing.

[0031] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for multi-channel synchronous detection of electrical performance of an automobile lamp wiring harness, characterized in that, The application relates to a method for detecting the electrical performance of an automobile lamp wire harness. The method comprises the following steps: acquiring the types and detection methods required for detecting the electrical performance of an automobile lamp wire harness, setting the detection priority according to the detection nature, and constructing a list of the electrical performance detection of the automobile lamp wire harness; constructing a hierarchical electrical performance detection and a Boolean rule with reserved skipping according to the detection priority, and dynamically regulating whether the type detection is skipped based on the rule; constructing a software switch configuration matrix according to the detection priority, the hierarchical electrical performance detection and the Boolean rule with reserved skipping, and dynamically regulating the electrical performance detection sequence based on the matrix; integrating a hardware detection platform of the detection method, connecting the interface of the electrical performance detection of the automobile lamp wire harness with the hardware detection platform, starting the control program of the software switch configuration matrix, and collecting the performance detection result; 2. The method for detecting the electrical performance of a wire harness of a vehicle lamp according to claim 1, characterized in that, constructing a comprehensive evaluation neural network model, combining the performance detection result, outputting the reason label and probability of the type detection skipping, and judging whether the skipped type is re-detected. The acquisition of the types and detection methods required for detecting the electrical performance of an automobile lamp wire harness, the setting of the detection priority according to the detection nature, and the construction of the list of the electrical performance detection of the automobile lamp wire harness specifically comprise the following steps: acquiring the required detection types of the electrical performance of the automobile lamp wire harness to be detected, wherein the detection types include but are not limited to conduction, short circuit, insulation, voltage resistance, resistance, voltage drop, load and communication; acquiring the detection methods corresponding to the detection types, and using the same line interface for the detection types with the same detection method; setting the priority of the electrical performance detection of the automobile lamp wire harness based on the principle of destructive detection and the expert scoring method; 3. The method of claim 2, wherein the method further comprises: constructing the list of the electrical performance detection of the automobile lamp wire harness, which comprises a group of data of the detection types, the detection methods corresponding to the detection types, the qualified standards and the detection priorities. The construction of the hierarchical electrical performance detection and the Boolean rule with reserved skipping according to the detection priority, and the dynamic regulation of whether the type detection is skipped based on the rule specifically comprise the following steps: according to the detection priority, taking the unqualified pre-detection as a bottom event, defining the safety accident caused by the execution of subsequent detection under the condition of the unqualified pre-detection as a risk scenario, and taking the scenario as a top event; based on historical data or experimental data, giving an objective failure probability value to all bottom events; constructing a fault tree model, calculating the occurrence probability of the top event from the bottom event probability, and defining the occurrence probability value as a safety risk coefficient from the unqualified pre-detection to the subsequent detection; repeating the execution of the fault tree model for all detection type combinations to be evaluated, and constructing a two-dimensional safety risk coefficient matrix from the model output result; wherein the rows of the safety risk coefficient matrix are pre-detection, the columns are subsequent detection, and the element values in the matrix are safety risk coefficients; according to the safety risk coefficient, carrying out a testability experiment, and dividing the safety risk coefficient into three Boolean rule intervals according to the experimental result. The first layer Boolean rule interval control instruction is 0, indicating that the subsequent detection is not skipped; the second layer Boolean rule interval control instruction is 1, indicating that the subsequent detection is executed conditionally, the system will adopt the reduced test, and a warning mark instruction is output during the execution; and the third layer Boolean rule interval control instruction is 2, indicating that the subsequent detection is skipped, and an alert mark instruction is output. According to the two-dimensional safety risk coefficient matrix and in combination with the three-layer Boolean rule interval, a layered and reserved-skip Boolean rule of the electrical performance detection is constructed, and whether the type detection is skipped is dynamically regulated based on the rule.

4. The method of claim 3, wherein the method further comprises: The software switch configuration matrix is constructed according to the detection priority and the layered and reserved-skip Boolean rule of the electrical performance detection, and the electrical performance detection sequence is dynamically regulated based on the matrix, and the specific process includes the following steps. According to the detection priority and the layered and reserved-skip Boolean rule of the electrical performance detection, a software switch configuration matrix is constructed, and the matrix is a three-dimensional control structure, and the three dimensions are respectively: The pre-detection state is derived from the real-time detection result of the electrical performance type, including qualified and unqualified; The current detection to be executed is derived from the detection priority sequence; The target hardware channel is derived from the automobile lamp harness electrical performance detection list; Each element of the matrix is a control tuple, and the content is dynamically generated by querying the layered and reserved-skip Boolean rule and the safety risk coefficient matrix; In the tuple, the switch on-off instruction, the reduced test parameter and the mark instruction are defined, and the switch on-off instruction is composed of numbers 0 and 1, wherein 0 is off and 1 is on. In the detection process, the system executes the following processes in sequence to realize dynamic regulation, which specifically includes the following steps. After a type detection is completed, the system queries the software switch configuration matrix according to the result and the next detection to be executed as an index, and obtains the corresponding control tuple; If the control tuple instruction is instruction 2, the system directly skips the detection, records alert, and indexes the next detection item; If the control tuple instruction is instruction 1, the system calls the reduced test parameter, configures the corresponding excitation source and measurement range, then executes the detection, and records the warning mark data; If the control tuple instruction is instruction 0, the current detection item is executed normally; The process is repeated until all items in the detection sequence are executed.

5. The method of claim 4, wherein the method further comprises: The hardware detection platform of the integrated detection mode connects the interface of the automobile lamp harness electrical performance detection with the hardware detection platform, starts the control program of the software switch configuration matrix, and collects the performance detection result, which specifically includes the following steps. Based on the hardware device of the detection mode, the hardware device is integrated into a comprehensive hardware detection platform, and the platform reserves hardware interfaces for different electrical performance type detections; Before the automobile lamp harness is connected, the hardware detection platform is started and a platform self-checking process is executed to ensure that the state of the device itself is normal; The self-checking process refers to that the control switch matrix checks the test channel in the platform for quick conduction and short circuit under the condition that the measured harness is not connected; The self-checking process refers to that the control switch matrix checks the test channel in the platform for quick conduction and short circuit under the condition that the measured harness is not connected; After the platform self-check is passed, the automobile lamp wire harness interface is connected with the hardware interface of the hardware detection platform, the corresponding detection list and the software switch configuration matrix are loaded, and the electrical performance detection is started; An electrical performance detection result list of the automobile lamp wire harness is obtained, and the list includes detection types, measured values, qualified states, and corresponding marker instruction records.

6. The method of claim 5, wherein the method further comprises: The comprehensive evaluation neural network model is constructed, the performance detection result is combined, the reason label and the probability of type detection skipping are output, and it is judged whether the skipped type is re-detected, and specifically includes: The comprehensive evaluation neural network model is constructed, and the input features of the model include pre-detection states, safety risk coefficients, Boolean rule control instruction types, de-rating test parameters, and key indicator deviation values in the electrical performance detection result; The output layer of the model is set to a double-output structure, which is the reason label of detection skipping and the confidence probability of the corresponding reason; The reason label at least includes product failure, detection system error, poor tool contact, and unstable excitation source; Based on the historical data of automobile lamp wire harness detection, the corresponding historical electrical performance detection result is collected through the automobile lamp wire harness electrical performance detection process; Based on the historical electrical performance detection result, the model training sample set is constructed combined with the manually labeled skipping reason label, and is divided into a training set, a validation set, and a test set; The sample set includes pre-detection states, safety risk coefficients, instruction types, detection results, and manually labeled skipping reason labels in different detection scenarios; The model training sample set is preprocessed, and the preprocessing at least includes one-hot encoding of labels, outlier removal, data normalization, and missing value filling; For the classification task of the reason label, a cross-entropy loss function is used to calculate the difference between the model prediction label and the manually labeled true label; The Adam optimizer is used to train the training set based on the minimization of the cross-entropy loss function, and the trained model is defined as the comprehensive evaluation neural network model; Based on the comprehensive evaluation neural network model, when the model gives a high confidence probability to the detection system error or the tool contact poor label, the system can automatically trigger the instruction to re-detect the skipped detection type, and the report highlights that the detection result is unreliable; Based on the test set, the stability of the comprehensive evaluation neural network model is evaluated through label accuracy, precision, recall, F1-Score, confidence probability mean, and recheck trigger accuracy performance indicators.

7. A multi-channel synchronous detection system for electrical performance of an automotive light wiring harness, characterized in that, The automobile lamp wire harness electrical performance multi-channel synchronous detection method is used to implement any one of claims 1-6, comprising: The software configuration module is used to acquire types and detection manners required by the automobile lamp wiring harness electrical performance detection, set detection priorities according to detection properties, and construct an automobile lamp wiring harness electrical performance detection list; according to the detection priorities, construct a layered electrical performance detection, a reserved skipping Boolean rule, and dynamically control whether the type detection is skipped based on the rule; according to the detection priorities and the layered electrical performance detection, the reserved skipping Boolean rule, construct a software switch configuration matrix, and dynamically control the electrical performance detection sequence based on the matrix; The hardware detection module is used to integrate a hardware detection platform of the detection manner, connect the interface of the automobile lamp wiring harness electrical performance detection with the hardware detection platform, start a control program of the software switch configuration matrix, and collect performance detection results; The comprehensive evaluation module is used to construct a comprehensive evaluation neural network model, combine the performance detection results, output a reason label and a probability of the type detection skipping, and judge whether the skipped type is re-detected.

8. The multi-channel synchronous detection system for electrical performance of an automobile lamp wiring harness according to claim 7, wherein, The software configuration module comprises: A detection list unit is used to acquire types and detection manners required by the automobile lamp wiring harness electrical performance detection, set detection priorities according to detection properties, and construct an automobile lamp wiring harness electrical performance detection list; A Boolean rule unit is used to construct a layered electrical performance detection, a reserved skipping Boolean rule according to the detection priorities, and dynamically control whether the type detection is skipped based on the rule; A detection sequence unit is used to construct a software switch configuration matrix according to the detection priorities and the layered electrical performance detection, the reserved skipping Boolean rule, and dynamically control the electrical performance detection sequence based on the matrix.

Images