CN104698399A - Automotive lamp fault detecting system and method thereof - Google Patents

Abstract

The invention relates to a car lamp fault detection and a method thereof, which belong to the field of automobiles. The system includes a signal acquisition unit to collect voltage and current signals of a car lamp circuit; a controller uses a genetically optimized BP network algorithm to analyze and process the data of the acquisition unit , the voltage and current signals of the monitoring circuit do not exceed the system threshold; the display screen, connected to the controller, displays the detection output result of the controller; the warning unit, connected to the controller, sends out a warning reminder after receiving the warning signal from the controller. The invention can monitor in real time whether the lights of the car are normally on, and promptly remind the driver of the problem of reduced brightness of the car lights, thereby reducing the safety hazards of driving the car. Simultaneously, in the automobile detection system of the present invention, the BP network algorithm after genetic optimization is applied to process the car light signal data collected by the acquisition unit, and the real-time optimal data can be optimized to the controller, avoiding the abnormal data transmission caused by accidental factors To the false alarm situation caused by the controller.

Description

A vehicle lamp fault detection system and method thereof

technical field

The invention relates to the field of automobiles, in particular to a vehicle lamp failure detection and a method thereof.

Background technique

There are various lights on the car, such as headlights, tail lights, turn signals and brake lights, but there is no indication in the car whether the lights are shining normally during driving. Lamp failure will bring great hidden dangers to the safe driving of the car.

In addition to the fault that the lights are not on, the brightness of the lights will also become dark due to circuit problems inside the car or the age of use, which will affect the normal use of the lights. And the bulbs of large and small lamps often burn out during use, causing the bulbs to burn out. The reasons are that the voltage regulator is faulty, the voltage is too high; there are multiple reasons such as a short circuit between the generator armature and the magnetic field coil.

Contents of the invention

In order to solve the problem in the prior art that the automobile does not prompt the state of the lamp being off, the present invention provides a fault detection system and method for the lamp.

The technical solution of the present invention is: a car lamp fault detection system, the system includes

The signal acquisition unit is connected to the controller to collect the voltage and current signals of the lamp circuit;

The controller receives the information of the signal acquisition unit, uses the genetic optimization BP network algorithm to analyze and process the data of the acquisition unit, and the voltage and current signals of the monitoring circuit do not exceed the system threshold;

The display screen is connected to the controller to display the detection output result of the controller;

The warning unit is connected with the controller, and sends out a warning reminder after receiving the warning signal from the controller.

The signal acquisition unit includes a voltage sensor, a current sensor and a brightness sensor. The sensors in the signal acquisition unit are evenly distributed. If the controller fails to receive the voltage signal data or the current signal data, it will send a warning reminder to the warning unit. The controller is provided with a database, and the database is used to save the information of the signal acquisition unit. The warning unit includes a buzzer and a warning light, and is located on the display screen.

A method for a vehicle lamp fault detection system, the method steps comprising:

Step 1. Collect the signal of the car light and the car light circuit;

Step 2, using the genetically optimized BP neural network model to preprocess the data collected in step 1;

Step 3, the network model analyzes the processed data;

Step 4, output the detection result.

The signals collected in the first step include voltage and current signals in the lamp circuit and brightness signals of the lamp.

The genetically optimized BP neural network model in the second step utilizes a genetically optimized BP neural network algorithm, which is divided into three parts: determining the structure of the BP neural network, optimizing weights and thresholds with the genetic algorithm, and training and predicting the BP neural network.

The genetic algorithm process in the genetically optimized BP neural network model includes:

A. Calculate fitness;

B. Selection of chromosomes for replication;

C. Crossover and mutation process;

D. Generate new groups;

E. Determine whether the termination condition is met;

F. End if the termination condition is met, and return to step A if not.

The present invention has the following positive effects: the car light detection system of the present invention can monitor in real time whether the car lights are normally on and remind the driver of the problem of the brightness reduction of the car lights in time, reducing the safety hazards of car driving. Simultaneously, in the automobile detection system of the present invention, the BP network algorithm after genetic optimization is applied to process the car light signal data collected by the acquisition unit, and the real-time optimal data can be optimized to the controller, avoiding the abnormal data transmission caused by accidental factors To the false alarm situation caused by the controller.

Description of drawings

Fig. 1 is the functional block diagram of the vehicle lamp detection system among the present invention;

Fig. 2 is the working flow diagram of the vehicle light detection system method among the present invention;

Fig. 3 is the flowchart of the BP neural network optimized by genetic algorithm in the present invention;

Fig. 4 is the flowchart of neural network algorithm among the present invention.

Detailed ways

Referring to the accompanying drawings, through the description of the embodiments, the specific embodiments of the present invention include the shape, structure, mutual position and connection relationship of each part, the function and working principle of each part, and the manufacturing process of the various components involved. And the method of operation and use, etc., are described in further detail to help those skilled in the art have a more complete, accurate and in-depth understanding of the inventive concepts and technical solutions of the present invention.

A car lamp fault detection system, as shown in Figure 1, the system includes a signal acquisition unit, a controller, a display screen and a warning unit, and the controller is connected to the signal collection unit, the display screen and the warning unit.

In the system, the signal acquisition unit is connected to the controller to collect the voltage and current signals of the lamp circuit. The signal acquisition unit includes a voltage sensor, a current sensor and a brightness sensor. There are multiple sensors for signal acquisition and evenly arranged. The brightness sensor is located next to the lamp to monitor the brightness of the lamp in real time and send the brightness signal to the controller for data analysis and processing. The voltage and current sensor is located in the system circuit of the lamp to monitor the voltage and current in the lamp in real time and The monitored data is sent to the controller to analyze whether the circuit is normal.

The controller receives the information from the signal acquisition unit and sends a signal to the display screen and the warning unit. The controller judges whether the signal of the collection unit is less than the normal threshold of the system, and sends a warning signal to the warning unit and the display screen once it is greater than the system threshold. If the controller cannot receive the signal from the acquisition unit, it may be that the wire inside the lamp circuit is blown or other reasons form an open circuit state, and the controller will also send a warning signal to the warning unit and the display screen. When the voltage and current in the lamp circuit are abnormal, it may be that the load of the lamp circuit is too large, which will cause the wire to heat up and cause the lamp to be burned or affect the normal operation of the engine. At the same time, the brightness monitoring of the headlights is equally important, especially at night or in foggy or dark weather, when the visibility is very low, the reduction in the brightness of the headlights due to excessive use of the headlights will affect the normal driving of the driver and affect the surrounding environment. The vehicle's judgment on the running direction of the vehicle. Therefore, once the detection of voltage and current coefficient in the lamp circuit is abnormal, the driver must repair it in time. In the system of the present invention, the genetically optimized BP network algorithm is used in the controller to analyze and process the data of the acquisition unit, and the training of the algorithm model is completed using the database data set in the controller to complete the data and historical data of the signal acquisition unit stored in the database, which is convenient For machine training, the voltage and current signals of the monitoring circuit in the controller do not exceed the system threshold; after using this algorithm to optimize the data of the acquisition unit, the real-time optimal data can be optimized to the controller, avoiding abnormal data transmission caused by accidental factors To the false alarm situation caused by the controller.

The display screen is connected to the controller to display the detection output result of the controller, which will clearly show which car light is not bright or the brightness is not enough, which is convenient for the driver to find and repair.

The warning unit is connected to the controller. The warning unit includes a buzzer and a warning light, which can be located on the display screen, and sends out a warning reminder after receiving a warning signal from the controller. If the controller fails to receive voltage signal data or current signal data, it will send a warning reminder to the warning unit.

A method for a car lamp fault detection system, as shown in Figure 2, the method steps comprising:

Step 1 of S01 , collecting the signal of the vehicle light and the signal of the vehicle light circuit. Step 1: The signal acquisition unit completes the acquisition signal task mainly including voltage sensor, current sensor and brightness sensor, collects the brightness signal of the car light and the voltage and current signal of the car light circuit, monitors the voltage and current in the car light in real time, and will monitor The received data is sent to the controller to analyze whether the circuit is normal.

Step 2 of S02: preprocessing the data collected in step 1 by using the genetically optimized BP neural network model, the genetically optimized BP neural network model is located inside the system controller, and processes and analyzes the data received by the controller.

The genetically optimized BP neural network model first preprocesses the collected data. The preprocessing is to normalize the data measured by the sensor. Normalization can speed up the convergence of the training network. The specific function of normalization is Generalizes the statistical distribution of uniform samples. Whether it is for modeling or calculation, first of all, the basic unit of measurement should be the same to facilitate the use of the following genetic BP neural network algorithm.

Step 3 of S03 , the network model analyzes the processed data. The genetically optimized BP neural network model is mainly to process and analyze the data using the genetically optimized BP neural network algorithm to select the optimal and most accurate data.

Genetic BP neural network algorithm is mainly divided into three parts: determining the structure of BP neural network; genetic algorithm optimization weight and threshold; BP neural network training and prediction. The process is shown in Figure 3 and Figure 4. First, determine the topology of the neural network, and then encode the weights and thresholds of the neural network to obtain the initial population. After processing the neural network algorithm, it enters the genetic algorithm processing part. In the genetic algorithm When the new group generated in cannot meet the termination condition, continue to run from the part of the neural network algorithm. If the termination condition is met, the decoding process is performed to obtain the optimal neural network weight and threshold.

The flow chart of the BP neural network in the genetically optimized BP neural network algorithm is shown in Figure 4. After encoding the weights and thresholds of the neural network to obtain the initial population, decode the weights and thresholds, and assign the weights and thresholds to the newly created The BP network uses the training samples to train the network, then uses the test samples to test the network, and finally performs the test error, and continues to enter the genetic algorithm process. Network training is a process of continuously correcting weights and thresholds. Through training, the output error of the network becomes smaller and smaller.

The learning algorithm of BP neural network is based on gradient descent, so it is prone to local minima, and has disadvantages such as slow convergence speed and difficulty in determining network parameters and training parameters. Genetic algorithm is a search algorithm that draws on natural selection and natural genetic mechanism in biology. It can find the optimal or quasi-optimal solution in a complex and huge search space, and has the advantages of simple algorithm, applicability, and strong robustness. , its application is currently very mature. Based on the advantages and disadvantages of BP neural network and genetic algorithm, the combination of the two makes their advantages and disadvantages complement each other, which has made great progress.

The BP neural network structure is a topological structure, which is determined according to the number of input/output parameters of the sample, so that the number of optimized parameters of the genetic algorithm can be determined, thereby determining the code length of the population individual. Because the genetic algorithm optimization parameters are the initial weights and thresholds of the BP neural network, as long as the network structure is known, the number of weights and thresholds is known. The weights and thresholds of the neural network are generally randomly initialized to random numbers in the interval [-0.5, 0.5]. This initialization parameter has a great influence on network training, but it cannot be obtained accurately. For the same initial weight value and threshold, The training results of the network are the same, and the genetic algorithm is introduced to optimize the best initial weights and thresholds, and then select the best data.

Genetic algorithm optimization of BP neural network is to use genetic algorithm to optimize the initial weight value and threshold of BP neural network, so that the optimized BP neural network can perform better sample prediction.

The elements of genetic algorithm in optimizing BP neural network by genetic algorithm include:

A． Calculate fitness. Calculate the fitness value: the individual fitness uses the network function error, that is, the individual with a large error has a small fitness, specifically expressed as the fitness is the reciprocal of the network error function. In order to make the residual error between the predicted value and the expected value as small as possible when the BP network predicts, the present invention selects the norm of the error matrix between the predicted value and the expected value of the predicted sample as the output of the objective function.

B. Selection of chromosomes for replication. Selection of chromosome replication: After the calculation of individual fitness is completed, individuals with high fitness are selected to pass on to the next generation, so that the weight is getting closer and closer to the optimal solution.

C. Crossover and mutation process. Crossover and mutation process: Using probability-based two-way random search technology, with a certain probability, randomly select two chromosomes from the parent population for crossover operation, when the new chromosome improves the quality of the current solution, it will receive the improved solution as the new current solution.

D． Generate new groups.

E. Judging whether the termination condition is met.

F. End if the termination condition is satisfied, and return to step A if it is not satisfied. In the present invention, the genetic algorithm is used to optimize the BP neural network, and step F returns to the BP neural network part.

After the data detected by each sensor is processed by the genetically optimized BP neural network model, the optimal and optimal data is selected and compared with the system threshold to determine whether the sensor detection data is less than the threshold and whether the sensor monitoring part is working normally. If not, Then the controller outputs the detection result.

Step 4 of S04, outputting the detection result.

After the genetically optimized BP neural network model in the controller analyzes the sensor data, the controller judges the data and outputs the detection results on the display screen, such as headlights, taillights, turn signals and brake lights are operating normally or The brightness reduction or circuit breakage will be displayed on the display screen. At the same time, if it is a lamp failure problem, the warning light and buzzer of the warning unit will also flash and make a sound to remind the driver that the lamp has a fault problem. , thereby increasing display security.

The present invention has been exemplarily described above in conjunction with the accompanying drawings. Obviously, the specific implementation of the present invention is not limited by the above methods, as long as various insubstantial improvements are adopted in the method concept and technical solutions of the present invention, or there is no improvement Directly applying the conception and technical solutions of the present invention to other occasions falls within the protection scope of the present invention.

Claims (10)

1. a light fault detection system, is characterized in that, this system comprises:

Signal gathering unit, connection control device, gathers the voltage and current signal of car light circuit;

Controller, the information of Received signal strength collecting unit, utilize genetic optimization BP network algorithm to the data analysis process of collecting unit, voltage, the current signal of observation circuit are no more than system thresholds;

Display screen, connection control device, the detection Output rusults of display controller;

Alarm unit, connection control device, sends warning and reminds after receiving the alarm signal of controller.

2. light fault detection system according to claim 1, is characterized in that, described signal gathering unit comprises voltage sensor, current sensor and luminance sensor.

3. light fault detection system according to claim 2, is characterized in that, the sensor in described signal gathering unit is uniformly distributed.

4. light fault detection system according to claim 1, is characterized in that, described controller does not receive voltage signal data or current signal data all can send warning prompting to alarm unit.

5. light fault detection system according to claim 1, is characterized in that, described controller is provided with database, and database is for preserving the information of signal gathering unit.

6. light fault detection system according to claim 1, is characterized in that, described alarm unit comprise hummer and or warning lamp, be positioned on display screen.

7. a light fault detection system method, is characterized in that, the method step comprises:

The signal of step one, collection car light and car light circuit;

Step 2, the BP neural network model of genetic optimization is utilized to carry out pre-service to the data that step one collects;

Data after step 3, network model analyzing and processing;

Step 4, output detections result.

8. light fault detection system method according to claim 7, is characterized in that, the signal of the collection in described step one comprises the luminance signal of voltage, current signal and car light in car light circuit.

9. light fault detection system method according to claim 7, it is characterized in that, genetic optimization BP neural network algorithm is make use of in genetic optimization BP neural network model in described step 2, this algorithm is divided into three parts: determine BP neural network structure, genetic algorithm optimization weights and threshold, BP neural metwork training and prediction.

10. light fault detection system method according to claim 7, is characterized in that, the genetic algorithm flow process in described genetic optimization BP neural network model comprises:

Calculate fitness;

Selective staining body copies;

Intersection, mutation process;

Produce new colony;

Judge whether to meet end condition;

Meet end condition then to terminate, do not meet and then return steps A.