A new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion

Abstract

This invention relates to a health monitoring system for a new energy vehicle chassis based on multi-source heterogeneous sensor fusion, belonging to the field of new energy vehicle manufacturing, and more specifically to the field of motor vehicle substructure. The system includes: a roller acquisition device for collecting various dynamometer-related data from a single-roller automotive chassis dynamometer; and a health monitoring device for intelligently analyzing the remaining mileage of the new energy vehicle's chassis shock absorbers before failure, using an intelligent chassis health monitoring model based on the multi-source heterogeneous sensor data and various dynamometer-related data. Through this invention, addressing the technical problem of the difficulty in predicting the timing of chassis shock absorber failures in new energy vehicles, this invention enables the use of different types of sensors to acquire various types of sensor data from the new energy vehicle chassis mounted on a single-roller automotive chassis dynamometer, and then employs an intelligent chassis health monitoring model to solve the aforementioned technical problem.

Description

Technical Field

[0001] The present invention relates to the manufacturing of new energy vehicles, and more specifically to the lower structure of motor vehicles, and particularly to a new energy vehicle chassis health monitoring system based on the fusion of multi-source heterogeneous sensors. Background Technology

[0002] New energy vehicles refer to automobiles that use unconventional vehicle fuels as their power source (or use conventional vehicle fuels but employ new onboard power devices), integrating advanced technologies in vehicle power control and drive, resulting in vehicles with advanced technical principles and new technologies and structures. New energy vehicles include four main types: hybrid electric vehicles (HEVs), battery electric vehicles (BEVs, including solar-powered vehicles), fuel cell electric vehicles (FCEVs), and other new energy vehicles (such as those using supercapacitors, flywheels, and other high-efficiency energy storage devices). Unconventional vehicle fuels refer to fuels other than gasoline and diesel. The health of the substructure of a new energy vehicle is related to the overall manufacturing quality and safety performance of the vehicle.

[0003] For example, Chinese invention patent publication CN114705455A discloses a load monitoring device for a new energy vehicle chassis, relating to the field of load monitoring devices. This new energy vehicle chassis load monitoring device includes a connecting base, one side of which is fixedly connected to the vehicle body. An on-board power supply is installed inside the vehicle body, and the on-board power supply is connected to a processor via wires. The processor is connected to a load monitoring component, an on-board display screen, and a data setting module via signal lines. The processor processes the information collected by sensors and cameras and displays it on the on-board display screen. The information is compared with parameter values ​​under normal load conditions. When the collected information is close to the parameter values ​​under normal load conditions, the display screen illuminates a yellow light, accompanied by a longer interval of beeping from the on-board speaker. When the collected information reaches the parameter values ​​under normal load conditions, the display screen illuminates a red light, accompanied by a shorter interval of beeping from the on-board speaker.

[0004] For example, Chinese invention patent publication CN118857767A discloses a method, device, electronic device, and storage medium for monitoring abnormal vehicle chassis conditions. The method includes monitoring abnormal vehicle chassis conditions based on the vertical acceleration state of the vehicle chassis. This monitoring includes: obtaining calculated values ​​of the vertical acceleration of the vehicle chassis based on model training; collecting measured values ​​of the vertical acceleration of the vehicle chassis under the current vehicle driving condition; determining whether the vehicle chassis is in an abnormal state based on the deviation between the measured and calculated values ​​of the vertical acceleration of the vehicle chassis; and identifying abnormal conditions of the vehicle chassis if the deviation exceeds a preset deviation threshold. This solution monitors the vertical acceleration state of the vehicle during normal driving and determines abnormal vehicle chassis conditions, enabling earlier and more accurate detection of such conditions.

[0005] Obviously, the aforementioned existing technologies only involve predicting vehicle chassis failures or analyzing abnormal vehicle chassis conditions based on single-sensor data. On the one hand, the prediction and analysis based on single-sensor data lacks the accuracy of the prediction and analysis results due to the lack of comprehensive and rich multi-dimensional sensor data. On the other hand, there is a lack of targeted prediction mechanisms for chassis failures of new energy vehicles. For example, it is impossible to reliably predict the remaining mileage before the chassis shock absorber of a new energy vehicle fails, making it difficult to predict in advance when the critical chassis component, namely the shock absorber, of a new energy vehicle will fail. Owners of new energy vehicles are prone to the predicament of sudden shock absorber failure during driving. Summary of the Invention

[0006] To address the technical problems in existing technologies, this invention provides a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion. This system utilizes various types of sensors, including mechanical sensors, visual sensors, and gyroscope sensors, to acquire different types of sensor data from a new energy vehicle chassis mounted on a single-roller automotive chassis dynamometer. This data is then input into a custom-designed intelligent chassis health monitoring model to perform intelligent analysis of the remaining mileage before the chassis shock absorbers fail. This completes the intelligent health monitoring of the new energy vehicle chassis based on multi-source heterogeneous sensor fusion, providing crucial driving reference data for owners of new energy vehicles.

[0007] According to the present invention, a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion is provided, the system comprising:

[0008] The roller acquisition device is used to collect the roller diameter, roller surface friction coefficient, roller weight, and roller type number of the single-roller automotive chassis dynamometer, so as to serve as various dynamometer correlation data of the single-roller automotive chassis dynamometer.

[0009] The first measuring device is used to measure the torque and power of the drive wheels of a new energy vehicle at a set test speed using a force sensing unit on the lever arm that connects the stator and the housing of a single-roller automotive chassis dynamometer.

[0010] The second measuring device is used to acquire overhead images of the chassis of new energy vehicles taken by a mobile vision sensor in overhead shooting mode.

[0011] The third measuring device is used to measure the body vibration amplitude of a new energy vehicle located above a single-roller automotive chassis dynamometer using a gyroscope sensing unit.

[0012] The health monitoring device is connected to the roller acquisition device, the first measuring device, the second measuring device, and the third measuring device, respectively. It is used to use the intelligent chassis health monitoring model to intelligently analyze the remaining mileage of the new energy vehicle before the chassis shock absorber fails, based on the set test vehicle speed, various dynamometer correlation data, the torque and power of the drive wheels of the new energy vehicle, the hue channel value, brightness channel value, saturation channel value, imaging depth value, coordinate value, and body vibration amplitude of each pixel in the overhead image of the chassis of the new energy vehicle.

[0013] Among them, the intelligent chassis health monitoring model is a convolutional neural network after each learning iteration, and the number of learning iterations is monotonically positively correlated with the roller diameter.

[0014] The intelligent chassis health monitoring model is based on a convolutional neural network structure consisting of an input layer, a convolutional layer, an activation layer, a pooling layer, and a fully connected layer. Data processing is achieved through hierarchical feature extraction.

[0015] According to a second aspect of the present invention, a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion is provided. The system includes a memory and multiple processors. The memory stores a computer program, which is configured to be executed by the multiple processors to complete the following steps:

[0016] Collect the roller diameter, roller surface friction coefficient, roller weight, and roller type number of the single-roller automotive chassis dynamometer to serve as various dynamometer-related data for the single-roller automotive chassis dynamometer;

[0017] The torque and power of the drive wheels of a new energy vehicle at a set test speed are measured using a force sensing unit on the lever arm that connects the stator and the housing of a single-roller automotive chassis dynamometer.

[0018] Acquire top-down images of the chassis of new energy vehicles taken by a mobile vision sensor in overhead shooting mode;

[0019] The amplitude of body vibration of a new energy vehicle located above a single-roller automotive chassis dynamometer was measured using a gyroscope sensing unit.

[0020] The intelligent chassis health monitoring model uses the set test vehicle speed, various dynamometer data, torque and power of the drive wheels of the new energy vehicle, hue channel values, brightness channel values, saturation channel values, imaging depth values ​​and coordinate values ​​of each pixel in the overhead image of the chassis of the new energy vehicle, as well as the body vibration amplitude of the new energy vehicle to intelligently analyze the remaining mileage of the chassis shock absorber of the new energy vehicle before it fails.

[0021] Among them, the intelligent chassis health monitoring model is a convolutional neural network after each learning iteration, and the number of learning iterations is monotonically positively correlated with the roller diameter.

[0022] The intelligent chassis health monitoring model is based on a convolutional neural network structure consisting of an input layer, a convolutional layer, an activation layer, a pooling layer, and a fully connected layer. Data processing is achieved through hierarchical feature extraction.

[0023] Therefore, it can be seen that the present invention has at least the following four prominent substantive features:

[0024] The first aspect involves employing a multi-source heterogeneous sensor fusion mechanism based on artificial intelligence to perform testing and intelligent evaluation of the chassis of new energy vehicles on a single-roller automotive chassis dynamometer. This mechanism integrates and utilizes data from various types of sensors, including mechanical sensors, visual sensors, and gyroscope sensors, to reliably predict the remaining mileage of the chassis shock absorber before it fails during the current test. By acquiring sensor data of different dimensions and types, it achieves intelligent health monitoring of the chassis performance of new energy vehicles, providing key reference data for the use of new energy vehicles for organizations and individuals who own them.

[0025] The second aspect: To achieve intelligent assessment of the remaining mileage of the chassis shock absorbers in new energy vehicles, a customized intelligent chassis health monitoring model was introduced. This intelligent chassis health monitoring model is a convolutional neural network that has undergone multiple learning iterations, and the number of learning iterations is monotonically positively correlated with the roller diameter. This allows for the construction of intelligent chassis health monitoring models with different structures for different vehicle testing environments. The convolutional neural network used consists of an input layer, a convolutional layer, an activation layer, a pooling layer, and a fully connected layer. Data processing is achieved through hierarchical feature extraction. The customized structural designs mentioned above ensure the stability and effectiveness of the intelligent assessment of the remaining mileage of the chassis shock absorbers in new energy vehicles.

[0026] Thirdly, to achieve intelligent assessment of the remaining mileage of the chassis shock absorbers in new energy vehicles, multiple basic data points are introduced, including data from various types of sensors such as mechanical sensors, visual sensors, and gyroscope sensors. Specifically, these basic data points include the set test speed, various dynamometer-related data from a single-roller automotive chassis dynamometer, the torque and power of the drive wheels of the new energy vehicle, the hue channel values, brightness channel values, saturation channel values, imaging depth of field values, and coordinate values ​​of each pixel in a top-down image of the chassis of the new energy vehicle, and the vehicle body vibration amplitude. The various dynamometer-related data from the single-roller automotive chassis dynamometer include the roller diameter, roller surface friction coefficient, roller weight, and roller type number. The use of these multiple basic data points further ensures the stability and effectiveness of the intelligent assessment of the remaining mileage of the chassis shock absorbers in new energy vehicles.

[0027] Fourthly, in each learning iteration of the convolutional neural network, the known mileage of a tested new energy vehicle from the test time to the time of chassis shock absorber failure is used as a single output of the convolutional neural network. The set test speed, various dynamometer-related data, the torque and power of the drive wheels of the new energy vehicle at the test time, the hue, brightness, saturation, depth of field, and coordinate values ​​of each pixel in the overhead image of the chassis at the test time, and the vehicle body vibration amplitude at the test time are used as multiple inputs to the convolutional neural network to complete the learning process, thus ensuring the learning effectiveness of each iteration of the convolutional neural network. Attached Figure Description

[0028] The embodiments of the present invention will now be described with reference to the accompanying drawings, wherein:

[0029] Figure 1 This is a schematic diagram of the technical process of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion according to the present invention.

[0030] Figure 2 This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the first embodiment of the present invention.

[0031] Figure 3 This is an internal structural diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the second embodiment of the present invention.

[0032] Figure 4This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the third embodiment of the present invention.

[0033] Figure 5 This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the fourth embodiment of the present invention.

[0034] Figure 6 This is an internal structural diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the fifth embodiment of the present invention.

[0035] Figure 7 This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the sixth embodiment of the present invention. Detailed Implementation

[0036] like Figure 1 The diagram illustrates a technical flow of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, according to the present invention. The present invention relates to the manufacturing of new energy vehicles, and more specifically to the field of motor vehicle substructures.

[0037] The specific technical process of this invention is as follows:

[0038] Technical Process A: To achieve intelligent assessment of the remaining mileage of the chassis shock absorbers of new energy vehicles, a customized intelligent chassis health monitoring model was introduced. The new energy vehicle was set on a single-roller automobile chassis dynamometer for testing.

[0039] Specifically, the customized structure of the intelligent chassis health monitoring model is mainly reflected in the following aspects:

[0040] First: The intelligent chassis health monitoring model is a convolutional neural network that has undergone various learning iterations. The convolutional neural network used consists of an input layer, a convolutional layer, an activation layer, a pooling layer, and a fully connected layer. Data processing is achieved through hierarchical feature extraction.

[0041] Secondly, the number of times the convolutional neural network is learned is monotonically and positively correlated with the roller diameter, thus enabling the construction of intelligent chassis health monitoring models with different structures for different vehicle testing environments, such as... Figure 1 As shown;

[0042] For example, the roller diameter of a single-roller automotive chassis dynamometer is 1500 mm, and the corresponding convolutional neural network is trained 800 times; the roller diameter of a single-roller automotive chassis dynamometer is 1800 mm, and the corresponding convolutional neural network is trained 900 times; the roller diameter of a single-roller automotive chassis dynamometer is 2100 mm, and the corresponding convolutional neural network is trained 1000 times; the roller diameter of a single-roller automotive chassis dynamometer is 2400 mm, and the corresponding convolutional neural network is trained 1100 times, and so on.

[0043] Furthermore, in each learning iteration of the convolutional neural network, the known mileage of a tested new energy vehicle from the test time to the time of chassis shock absorber failure is used as a single output of the convolutional neural network. The set test speed, various dynamometer correlation data, the torque and power of the drive wheels of the new energy vehicle at the test time, the hue channel values, brightness channel values, saturation channel values, imaging depth of field values ​​and coordinate values ​​of each pixel in the overhead image of the chassis of the new energy vehicle at the test time, and the body vibration amplitude of the new energy vehicle at the test time are used as multiple inputs of the convolutional neural network to complete this learning, thereby ensuring the learning effect of each learning iteration of the convolutional neural network.

[0044] In this way, the stability and effectiveness of the intelligent assessment of the remaining mileage of the chassis shock absorbers of new energy vehicles are ensured through the customized structural designs mentioned above.

[0045] Technical Process B: To achieve intelligent assessment of the remaining mileage of the chassis shock absorbers of new energy vehicles, multiple basic data sources, including mechanical sensors, visual sensors, and gyroscope sensors, are introduced to realize the fusion and use of multi-source heterogeneous sensor data.

[0046] Specifically, the multiple basic data include the set test vehicle speed, various dynamometer-related data of the single-roller automobile chassis dynamometer, the torque and power of the drive wheels of the new energy vehicle, the hue channel value, brightness channel value, saturation channel value, imaging depth of field value and coordinate value of each pixel in the overhead image of the chassis of the new energy vehicle, and the body vibration amplitude of the new energy vehicle.

[0047] More specifically, the various dynamometer-related data of the single-roller automotive chassis dynamometer include the roller diameter, roller surface friction coefficient, roller weight, and roller type number.

[0048] In this way, by using the above-mentioned basic data, the stability and effectiveness of the intelligent assessment of the remaining mileage of the chassis shock absorbers of new energy vehicles are further guaranteed.

[0049] Technical Process C: Utilizing the intelligent chassis health monitoring model with customized structural design based on Technical Process A, and based on multiple basic data from various types of sensors, including mechanical sensors, visual sensors, and gyroscope sensors, introduced in Technical Process B, the remaining mileage of the new energy vehicle chassis shock absorbers tested on a single-roller automotive chassis dynamometer is intelligently analyzed.

[0050] For example, a chassis shock absorber malfunction is caused by an oil leak or the damping of the chassis shock absorber being lower than the set damping threshold.

[0051] Technical Process D: The remaining mileage of the new energy vehicle chassis shock absorbers, which is intelligently analyzed in Technical Process C, is wirelessly transmitted to a remote vehicle test management server via a mobile communication link.

[0052] For example, the mobile communication link is based on time-division duplex communication mode or frequency-division duplex communication mode, and the remote vehicle test management server is a big data service network element, cloud computing service network element or blockchain service network element;

[0053] Therefore, through the coordinated operation of the above-mentioned technical processes, a multi-source heterogeneous sensor fusion mechanism based on artificial intelligence is used to perform testing and intelligent evaluation of the new energy vehicle chassis located on a single-roller automotive chassis dynamometer. The multi-source heterogeneous sensor fusion mechanism integrates and utilizes data from various types of sensors, including mechanical sensors, visual sensors, and gyroscope sensors, to reliably predict the remaining mileage before the chassis shock absorber fails during the current test. Thus, by acquiring sensor data of various dimensions and types, intelligent health monitoring of the new energy vehicle chassis performance is completed, providing key reference data for the use of new energy vehicles for units and individuals who own them.

[0054] The key points of this invention are: the hardware construction of a multi-sensor testing environment for new energy vehicle chassis located on a single-roller automotive chassis dynamometer; the fusion and use of various types of sensor data from mechanical sensors, visual sensors, and gyroscope sensors through a multi-source heterogeneous sensor fusion mechanism; and intelligent chassis health monitoring models with different customized structures corresponding to different testing environments.

[0055] The following will describe in detail an embodiment of the new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion.

[0056] First Embodiment

[0057] Figure 2 This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the first embodiment of the present invention.

[0058] like Figure 2 As shown, the new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion includes the following components:

[0059] The roller acquisition device is used to collect the roller diameter, roller surface friction coefficient, roller weight, and roller type number of the single-roller automotive chassis dynamometer, so as to serve as various dynamometer correlation data of the single-roller automotive chassis dynamometer.

[0060] For example, a roller acquisition device is used to acquire the roller diameter, roller surface friction coefficient, roller weight, and roller type number of a single-roller automotive chassis dynamometer, as various dynamometer-related data of the single-roller automotive chassis dynamometer. This includes using multiple different acquisition units to acquire the roller diameter, roller surface friction coefficient, roller weight, and roller type number of the single-roller automotive chassis dynamometer separately.

[0061] Specifically, dynamometers for new energy vehicle chassis are divided into single-roller and double-roller types:

[0062] A single-roller chassis dynamometer has one roller supporting each drive wheel. The roller diameter is relatively large (generally between 1500 and 2500 mm), with fewer support bearings and less mechanical loss of the test bench. The larger the roller diameter, the more the wheel rolls on the roller as if it were rolling on a flat road. The slip ratio between the tire and the roller is small and the rolling resistance is small, resulting in high testing accuracy. However, the manufacturing and installation costs are high, and it is generally used in manufacturing plants and research units.

[0063] A dual-roller chassis dynamometer has two rollers supporting each drive wheel. The roller diameter is small (generally between 180 and 500 mm). Compared with a single-roller chassis dynamometer, it has four additional support bearings and one coupling. During the testing process, its mechanical loss is greater. The smaller the roller diameter, the greater the difference between the contact between the wheel and the roller and that on a flat road. The slip rate between the tire and the roller increases, and the rolling resistance increases, so the testing accuracy is poor. The advantages are low equipment cost and ease of use. It is generally used in the automotive industry, repair industry, and automotive testing lines or stations.

[0064] The rotating rollers are equivalent to a continuously moving road surface, upon which the wheels of the tested new energy vehicles roll. The rollers simulate the road surface, and their surfaces can be smooth, knurled, grooved, or coated, depending on the application. The goal is to make the roller's adhesion as close as possible to the actual road conditions. Smooth rollers are currently the most widely used type. For double-roller smooth rollers, although the coefficient of adhesion is lower due to the increased tire-to-roller pressure, the adhesion between the wheel and the smooth roller can generate sufficient traction. Coated smooth rollers can increase adhesion and are a promising option. Knurled and grooved rollers are rarely used because the slippage rate cannot be kept constant during use.

[0065] The first measuring device is used to measure the torque and power of the drive wheels of a new energy vehicle at a set test speed using a force sensing unit on the lever arm that connects the stator and the housing of a single-roller automotive chassis dynamometer.

[0066] Specifically, the measurement of the torque and power of the drive wheels of a new energy vehicle at a set test speed using a force sensing unit on the lever arm connecting the stator and the housing of a single-roller automotive chassis dynamometer includes: performing analysis of the torque and power of the drive wheels of the new energy vehicle by measuring the reaction force of the drive wheels;

[0067] The second measuring device is used to acquire overhead images of the chassis of new energy vehicles taken by a mobile vision sensor in overhead shooting mode.

[0068] For example, acquiring an overhead image of the chassis of a new energy vehicle taken by a mobile vision sensor in overhead mode includes: the resolution of the mobile vision sensor is 2K resolution or 4K resolution;

[0069] The third measuring device is used to measure the body vibration amplitude of a new energy vehicle located above a single-roller automotive chassis dynamometer using a gyroscope sensing unit.

[0070] For example, measuring the body vibration amplitude of a new energy vehicle located above a single-roller automotive chassis dynamometer using a gyroscope sensing unit includes: the gyroscope sensing unit used is located on the body of the new energy vehicle;

[0071] The health monitoring device is connected to the roller acquisition device, the first measuring device, the second measuring device, and the third measuring device, respectively. It is used to use the intelligent chassis health monitoring model to intelligently analyze the remaining mileage of the new energy vehicle before the chassis shock absorber fails, based on the set test vehicle speed, various dynamometer correlation data, the torque and power of the drive wheels of the new energy vehicle, the hue channel value, brightness channel value, saturation channel value, imaging depth value, coordinate value, and body vibration amplitude of each pixel in the overhead image of the chassis of the new energy vehicle.

[0072] Specifically, a numerical simulation mode can be used to simulate and test the data processing process of the remaining mileage before the chassis shock absorber of the new energy vehicle fails, based on the set test vehicle speed, various dynamometer-related data, the torque and power of the drive wheels of the new energy vehicle, the hue channel value, brightness channel value, saturation channel value, imaging depth value and coordinate value of each pixel in the overhead image of the chassis of the new energy vehicle, and the body vibration amplitude of the new energy vehicle.

[0073] Among them, the intelligent chassis health monitoring model is a convolutional neural network after each learning iteration, and the number of learning iterations is monotonically positively correlated with the roller diameter.

[0074] For example, the roller diameter of a single-roller automotive chassis dynamometer is 1500 mm, and the corresponding convolutional neural network is trained 800 times; the roller diameter of a single-roller automotive chassis dynamometer is 1800 mm, and the corresponding convolutional neural network is trained 900 times; the roller diameter of a single-roller automotive chassis dynamometer is 2100 mm, and the corresponding convolutional neural network is trained 1000 times; the roller diameter of a single-roller automotive chassis dynamometer is 2400 mm, and the corresponding convolutional neural network is trained 1100 times, and so on.

[0075] The intelligent chassis health monitoring model is based on a convolutional neural network structure consisting of an input layer, a convolutional layer, an activation layer, a pooling layer, and a fully connected layer. Data processing is achieved through hierarchical feature extraction.

[0076] Among them, the chassis shock absorber malfunction is caused by oil leakage in the chassis shock absorber or the damping of the chassis shock absorber being lower than the set damping threshold.

[0077] Specifically, for the two faults of oil leakage in the chassis shock absorber or damping of the chassis shock absorber being lower than the set damping threshold, the time of occurrence of the first fault is taken as the time of failure of the chassis shock absorber.

[0078] The method of measuring the torque and power of the drive wheels of a new energy vehicle at a set test speed using a force sensing unit on the lever arm connecting the stator and the housing of a single-drum automotive chassis dynamometer includes: the single-drum automotive chassis dynamometer includes a housing, a stator, a rotor, a loading device, and a single drum; the loading device applies a braking torque to the rotor driving the single drum through the stator; at the same time, the stator receives the reaction torque of the rotor; the reaction torque is measured by the force sensing unit and converted into the torque and power of the drive wheels of the new energy vehicle.

[0079] In each learning iteration of the convolutional neural network, the known mileage of a tested new energy vehicle from the test time to the time when the chassis shock absorber malfunctioned is used as a single output of the convolutional neural network. The set test speed, various dynamometer correlation data, the torque and power of the drive wheels of the new energy vehicle at the test time, the hue channel values, brightness channel values, saturation channel values, imaging depth of field values ​​and coordinate values ​​of each pixel in the overhead image of the chassis of the new energy vehicle at the test time, and the body vibration amplitude of the new energy vehicle at the test time are used as multiple inputs of the convolutional neural network to complete this learning iteration.

[0080] The acquisition of the overhead chassis image of the new energy vehicle taken by the mobile vision sensor in overhead mode includes: the overhead chassis image only includes the chassis of the new energy vehicle, and the mobile vision sensor is located on the side of the single-roller automotive chassis dynamometer and includes a robotic arm, positioning equipment, drive equipment, micro-control equipment and ultra-high-definition camera equipment.

[0081] Second Embodiment

[0082] Figure 3 This is an internal structural diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the second embodiment of the present invention.

[0083] like Figure 3 As shown, compared to Figure 2 The new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion also includes:

[0084] The weight sensing device is connected to the roller acquisition device to measure the roller weight of the single-roller automobile chassis dynamometer and send the measured roller weight of the single-roller automobile chassis dynamometer to the roller acquisition device.

[0085] For example, the weight sensing device and the roller acquisition device share the same IIC configuration interface and the same power supply;

[0086] The weight sensing device, connected to the roller acquisition device, is used to measure the roller weight of the single-roller automotive chassis dynamometer and send the measured roller weight of the single-roller automotive chassis dynamometer to the roller acquisition device. The weight sensing device is located below the single-roller automotive chassis dynamometer and only contacts the single-roller automotive chassis dynamometer during measurement.

[0087] Third Embodiment

[0088] Figure 4 This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the third embodiment of the present invention.

[0089] like Figure 4 As shown, compared to Figure 3 The new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion also includes:

[0090] A visual analyzer, connected to a roller acquisition device, is used to analyze the roller type of a single-roller automotive chassis dynamometer using a visual analysis mechanism, and to parse the roller type number of the single-roller automotive chassis dynamometer based on the roller type obtained from the analysis.

[0091] Specifically, the roller type number of the single-roller automotive chassis dynamometer can be numerically represented by the ASCII code value of the string corresponding to the roller type of the single-roller automotive chassis dynamometer.

[0092] The visualization analyzer, connected to the roller acquisition device, is used to analyze the roller type of the single-roller automotive chassis dynamometer using a visual analysis mechanism, and to deduce the roller type number of the single-roller automotive chassis dynamometer based on the roller type obtained from the analysis. The visualization analyzer is set on the side of the roller of the single-roller automotive chassis dynamometer.

[0093] Fourth embodiment

[0094] Figure 5 This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the fourth embodiment of the present invention.

[0095] like Figure 5 As shown, compared to Figure 2 The new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion also includes:

[0096] The content transmission device, connected to the health monitoring device, is used to receive the remaining mileage of the chassis shock absorber of the new energy vehicle before the chassis shock absorber fails, and wirelessly transmit the remaining mileage of the new energy vehicle before the chassis shock absorber fails to the remote vehicle test management server via a mobile communication link.

[0097] For example, the remote vehicle test management server can be a big data service network element, a cloud computing service network element, or a blockchain service network element;

[0098] The content transmission device, connected to the health monitoring device, is used to receive the remaining mileage of the new energy vehicle before the chassis shock absorber fails, and wirelessly transmit the remaining mileage of the new energy vehicle before the chassis shock absorber fails to the remote vehicle test management server via a mobile communication link. The mobile communication link is based on time division duplex communication mode or frequency division duplex communication mode.

[0099] Fifth Embodiment

[0100] Figure 6 This is an internal structural diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the fifth embodiment of the present invention.

[0101] like Figure 6 As shown, compared to Figure 2 The new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion also includes:

[0102] The on-site display device is connected to the health monitoring device to receive the remaining mileage of the new energy vehicle before the chassis shock absorber fails, and displays the remaining mileage of the new energy vehicle before the chassis shock absorber fails.

[0103] For example, an LCD screen can be used to implement the field display device, which is used to receive the remaining mileage before the chassis shock absorber of the new energy vehicle fails, and to display the remaining mileage before the chassis shock absorber of the new energy vehicle fails.

[0104] Among them, the on-site display device is connected to the health monitoring device and is used to receive the remaining mileage before the chassis shock absorber of the new energy vehicle fails, and to display the remaining mileage before the chassis shock absorber of the new energy vehicle fails. The on-site display device is located on the side of the single-roller automobile chassis dynamometer.

[0105] Next, various embodiments of the present invention will be further described.

[0106] Optionally, within the above embodiments, in the new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion:

[0107] The roller diameter, roller surface friction coefficient, roller weight, and roller type number of the single-roller automotive chassis dynamometer are collected as various dynamometer-related data for the single-roller automotive chassis dynamometer. These data include: the roller type is a smooth roller, a knurled roller, a grooved roller, or a coated roller, and different types have different roller type numbers.

[0108] Among them, the data collected for the single-roller automotive chassis dynamometer, including the roller diameter, roller surface friction coefficient, roller weight, and roller type number, also include the roller diameter of the single-roller automotive chassis dynamometer being between 1500 mm and 2500 mm.

[0109] Among them, the intelligent chassis health monitoring model is a convolutional neural network after each learning iteration, and the number of learning iterations is monotonically positively correlated with the roller diameter, including: using a conversion formula to represent the numerical conversion relationship between the number of learning iterations and the roller diameter.

[0110] For example, the MATLAB toolbox can be used to test and simulate the data processing process that uses the number of learning cycles to represent the numerical transformation relationship between the number of learning cycles and the roller diameter in a monotonically positive correlation.

[0111] The numerical conversion relationship between the number of learning cycles and the roller diameter, expressed by the cycle conversion formula, includes: in the cycle conversion formula, the roller diameter of the single-roller automotive chassis dynamometer is the input value of the cycle conversion formula;

[0112] Furthermore, the numerical conversion relationship between the number of learning iterations and the roller diameter, expressed by the conversion formula, also includes: in the conversion formula, the number of learning iterations that are positively correlated with the roller diameter of the single-roller automotive chassis dynamometer is the output value of the conversion formula.

[0113] And, optionally, in the above embodiments, in the new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion:

[0114] The intelligent chassis health monitoring model uses a set test speed, various dynamometer-related data, the torque and power of the new energy vehicle's drive wheels, the hue channel values, brightness channel values, saturation channel values, imaging depth values ​​and coordinate values ​​of each pixel in the overhead image of the new energy vehicle's chassis, and the body vibration amplitude of the new energy vehicle to intelligently analyze the remaining mileage of the new energy vehicle's chassis shock absorber before the failure. This includes inputting the set test speed, various dynamometer-related data, the torque and power of the new energy vehicle's drive wheels, the hue channel values, brightness channel values, saturation channel values, imaging depth values ​​and coordinate values ​​of each pixel in the overhead image of the new energy vehicle's chassis, and the body vibration amplitude of the new energy vehicle into the intelligent chassis health monitoring model in parallel.

[0115] Specifically, programmable logic devices can be selected as parallel control devices to input the set test vehicle speed, various dynamometer-related data, torque and power of the drive wheels of the new energy vehicle, hue channel values, brightness channel values, saturation channel values, imaging depth values ​​and coordinate values ​​of each pixel in the overhead image of the chassis of the new energy vehicle, as well as the body vibration amplitude of the new energy vehicle into the intelligent chassis health monitoring model in parallel.

[0116] The method of using an intelligent chassis health monitoring model to intelligently analyze the remaining mileage of the new energy vehicle before the chassis shock absorber malfunctions, based on the set test vehicle speed, various dynamometer-related data, the torque and power of the drive wheels of the new energy vehicle, the hue channel value, brightness channel value, saturation channel value, imaging depth value and coordinate value of each pixel in the overhead image of the chassis of the new energy vehicle, and the body vibration amplitude of the new energy vehicle, also includes: executing the intelligent chassis health monitoring model to obtain the remaining mileage of the new energy vehicle before the chassis shock absorber malfunctions, output by the intelligent chassis health monitoring model;

[0117] The process of inputting the test vehicle speed, various dynamometer-related data, torque and power of the drive wheels of the new energy vehicle, hue channel values, brightness channel values, saturation channel values, imaging depth values ​​and coordinate values ​​of each pixel in the overhead image of the chassis of the new energy vehicle, and body vibration amplitude of the new energy vehicle into the intelligent chassis health monitoring model in parallel includes: performing numerical normalization processing on the test vehicle speed, various dynamometer-related data, torque and power of the drive wheels of the new energy vehicle, hue channel values, brightness channel values, saturation channel values, imaging depth values ​​and coordinate values ​​of each pixel in the overhead image of the chassis of the new energy vehicle, and body vibration amplitude of the new energy vehicle before inputting them into the intelligent chassis health monitoring model in parallel;

[0118] And among them, executing the intelligent chassis health monitoring model to obtain the remaining mileage of the new energy vehicle before the chassis shock absorber fails, as output by the intelligent chassis health monitoring model, includes: the remaining mileage of the new energy vehicle before the chassis shock absorber fails, as output by the intelligent chassis health monitoring model, is a numerically normalized representation.

[0119] Sixth Embodiment

[0120] Figure 7 This is an internal structure diagram of a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion, as shown in the sixth embodiment of the present invention.

[0121] like Figure 7 As shown, the multimedia content automatic generation and publishing system includes a memory and multiple processors. The memory stores a computer program, which is configured to be executed by the multiple processors to complete the following steps:

[0122] Step 71: Collect the roller diameter, roller surface friction coefficient, roller weight, and roller type number of the single-roller automotive chassis dynamometer to serve as various dynamometer correlation data for the single-roller automotive chassis dynamometer.

[0123] For example, a roller acquisition device is used to acquire the roller diameter, roller surface friction coefficient, roller weight, and roller type number of a single-roller automotive chassis dynamometer, as various dynamometer-related data of the single-roller automotive chassis dynamometer. This includes using multiple different acquisition units to acquire the roller diameter, roller surface friction coefficient, roller weight, and roller type number of the single-roller automotive chassis dynamometer separately.

[0124] Specifically, dynamometers for new energy vehicle chassis are divided into single-roller and double-roller types:

[0125] A single-roller chassis dynamometer has one roller supporting each drive wheel. The roller diameter is relatively large (generally between 1500 and 2500 mm), with fewer support bearings and less mechanical loss of the test bench. The larger the roller diameter, the more the wheel rolls on the roller as if it were rolling on a flat road. The slip ratio between the tire and the roller is small and the rolling resistance is small, resulting in high testing accuracy. However, the manufacturing and installation costs are high, and it is generally used in manufacturing plants and research units.

[0126] A dual-roller chassis dynamometer has two rollers supporting each drive wheel. The roller diameter is small (generally between 180 and 500 mm). Compared with a single-roller chassis dynamometer, it has four additional support bearings and one coupling. During the testing process, its mechanical loss is greater. The smaller the roller diameter, the greater the difference between the contact between the wheel and the roller and that on a flat road. The slip rate between the tire and the roller increases, and the rolling resistance increases, so the testing accuracy is poor. The advantages are low equipment cost and ease of use. It is generally used in the automotive industry, repair industry, and automotive testing lines or stations.

[0127] The rotating rollers are equivalent to a continuously moving road surface, upon which the wheels of the tested new energy vehicles roll. The rollers simulate the road surface, and their surfaces can be smooth, knurled, grooved, or coated, depending on the application. The goal is to make the roller's adhesion as close as possible to the actual road conditions. Smooth rollers are currently the most widely used type. For double-roller smooth rollers, although the coefficient of adhesion is lower due to the increased tire-to-roller pressure, the adhesion between the wheel and the smooth roller can generate sufficient traction. Coated smooth rollers can increase adhesion and are a promising option. Knurled and grooved rollers are rarely used because the slippage rate cannot be kept constant during use.

[0128] Step 72: Measure the torque and power of the drive wheels of the new energy vehicle at the set test speed using the force sensing unit on the lever arm connecting the stator and the single-roller automobile chassis dynamometer housing.

[0129] Specifically, the measurement of the torque and power of the drive wheels of a new energy vehicle at a set test speed using a force sensing unit on the lever arm connecting the stator and the housing of a single-roller automotive chassis dynamometer includes: performing analysis of the torque and power of the drive wheels of the new energy vehicle by measuring the reaction force of the drive wheels;

[0130] Step 73: Acquire an overhead image of the chassis of the new energy vehicle taken by the mobile vision sensor in overhead mode;

[0131] For example, acquiring an overhead image of the chassis of a new energy vehicle taken by a mobile vision sensor in overhead mode includes: the resolution of the mobile vision sensor is 2K resolution or 4K resolution;

[0132] Step 74: Use a gyroscope sensing unit to measure the body vibration amplitude of the new energy vehicle located above the single-roller automotive chassis dynamometer;

[0133] For example, measuring the body vibration amplitude of a new energy vehicle located above a single-roller automotive chassis dynamometer using a gyroscope sensing unit includes: the gyroscope sensing unit used is located on the body of the new energy vehicle;

[0134] Step 75: Using the intelligent chassis health monitoring model, based on the set test speed, various dynamometer-related data, the torque and power of the drive wheels of the new energy vehicle, the hue channel value, brightness channel value, saturation channel value, imaging depth value and coordinate value of each pixel in the overhead image of the chassis of the new energy vehicle, and the body vibration amplitude of the new energy vehicle, the remaining mileage of the chassis shock absorber of the new energy vehicle before the failure occurs.

[0135] Specifically, a numerical simulation mode can be used to simulate and test the data processing process of the remaining mileage before the chassis shock absorber of the new energy vehicle fails, based on the set test vehicle speed, various dynamometer-related data, the torque and power of the drive wheels of the new energy vehicle, the hue channel value, brightness channel value, saturation channel value, imaging depth value and coordinate value of each pixel in the overhead image of the chassis of the new energy vehicle, and the body vibration amplitude of the new energy vehicle.

[0136] Among them, the intelligent chassis health monitoring model is a convolutional neural network after each learning iteration, and the number of learning iterations is monotonically positively correlated with the roller diameter.

[0137] For example, the roller diameter of a single-roller automotive chassis dynamometer is 1500 mm, and the corresponding convolutional neural network is trained 800 times; the roller diameter of a single-roller automotive chassis dynamometer is 1800 mm, and the corresponding convolutional neural network is trained 900 times; the roller diameter of a single-roller automotive chassis dynamometer is 2100 mm, and the corresponding convolutional neural network is trained 1000 times; the roller diameter of a single-roller automotive chassis dynamometer is 2400 mm, and the corresponding convolutional neural network is trained 1100 times, and so on.

[0138] The intelligent chassis health monitoring model is based on a convolutional neural network structure consisting of an input layer, a convolutional layer, an activation layer, a pooling layer, and a fully connected layer. Data processing is achieved through hierarchical feature extraction.

[0139] Among them, the chassis shock absorber malfunction is caused by oil leakage in the chassis shock absorber or the damping of the chassis shock absorber being lower than the set damping threshold.

[0140] Specifically, for the two faults of oil leakage in the chassis shock absorber or damping of the chassis shock absorber being lower than the set damping threshold, the time of occurrence of the first fault is taken as the time of failure of the chassis shock absorber.

[0141] The method of measuring the torque and power of the drive wheels of a new energy vehicle at a set test speed using a force sensing unit on the lever arm connecting the stator and the housing of a single-drum automotive chassis dynamometer includes: the single-drum automotive chassis dynamometer includes a housing, a stator, a rotor, a loading device, and a single drum; the loading device applies a braking torque to the rotor driving the single drum through the stator; at the same time, the stator receives the reaction torque of the rotor; the reaction torque is measured by the force sensing unit and converted into the torque and power of the drive wheels of the new energy vehicle.

[0142] In each learning iteration of the convolutional neural network, the known mileage of a tested new energy vehicle from the test time to the time when the chassis shock absorber malfunctioned is used as a single output of the convolutional neural network. The set test speed, various dynamometer correlation data, the torque and power of the drive wheels of the new energy vehicle at the test time, the hue channel values, brightness channel values, saturation channel values, imaging depth of field values ​​and coordinate values ​​of each pixel in the overhead image of the chassis of the new energy vehicle at the test time, and the body vibration amplitude of the new energy vehicle at the test time are used as multiple inputs of the convolutional neural network to complete this learning iteration.

[0143] The acquisition of the overhead chassis image of the new energy vehicle taken by the mobile vision sensor in overhead mode includes: the overhead chassis image only includes the chassis of the new energy vehicle, and the mobile vision sensor is located on the side of the single-roller automotive chassis dynamometer and includes a robotic arm, positioning equipment, drive equipment, micro-control equipment and ultra-high-definition camera equipment.

[0144] Furthermore, in a new energy vehicle chassis health monitoring system based on multi-source heterogeneous sensor fusion according to the present invention:

[0145] The method involves acquiring an overhead image of the chassis of a new energy vehicle taken by a mobile vision sensor in overhead mode. The overhead chassis image includes only the chassis of the new energy vehicle. The method involves identifying the chassis imaging area in the imaging image taken by the mobile vision sensor in overhead mode based on the reference contour pattern and / or grayscale value distribution range of the chassis of the new energy vehicle, and outputting the chassis imaging area as the overhead chassis image of the new energy vehicle.

[0146] In this image, each pixel of the overhead view of the chassis of the new energy vehicle has H channel values ​​(hue channel values), S channel values ​​(brightness channel values) and V channel values ​​(saturation channel values) in the HSV color space, and the coordinate values ​​of each pixel of the overhead view of the chassis of the new energy vehicle are the vertical coordinate values ​​and the horizontal coordinate values ​​of the pixel.

[0147] Specifically, each pixel has a value between 0 and 255 for any of the H, S, and V channels in the HSV color space.

[0148] While the invention has been described in considerable detail, it should be understood that those skilled in the art can modify its elements without departing from the spirit and scope of the invention. It is believed that the system of the invention and its associated advantages will be understood from the foregoing description, and it will be clear that various changes can be made to its form, structure, and component arrangement without departing from the scope and spirit of the invention or sacrificing all its substantial advantages, and since the forms described above are merely illustrative embodiments of the invention, no further substantial changes are provided. The claims are intended to cover and include these changes.

Claims

1. A new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion, characterized in that, The system comprises: A roller collecting device for collecting the roller diameter, roller surface friction coefficient, roller weight and roller type number of a single-roller automobile chassis dynamometer as various dynamometer-related data of the single-roller automobile chassis dynamometer; A first measuring device for measuring the torque and power of the drive wheel of the new energy vehicle at a set test vehicle speed by using a force sensing unit connected to the force arm of the stator and the single-roller automobile chassis dynamometer shell; A second measuring device for obtaining the overhead view image of the vehicle-mounted chassis of the new energy vehicle taken by the mobile visual sensor in the overhead view mode; A third measuring device for measuring the body jitter amplitude of the new energy vehicle located above the single-roller automobile chassis dynamometer by using a gyroscopic sensing unit; A health monitoring device connected to the roller collecting device, the first measuring device, the second measuring device and the third measuring device, respectively, for intelligently analyzing the remaining mileage of the new energy vehicle before the failure of the chassis shock absorber of the new energy vehicle by using an intelligent chassis health monitoring model according to the set test vehicle speed, the various dynamometer-related data, the torque and power of the drive wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead view image of the vehicle-mounted chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle; Wherein the intelligent chassis health monitoring model is a convolutional neural network after learning, and the number of learning is monotonically positively correlated with the roller diameter; Wherein the structure of the convolutional neural network based on the intelligent chassis health monitoring model comprises an input layer, a convolutional layer, an activation layer, a pooling layer and a fully connected layer, and data processing is realized through hierarchical feature extraction; Wherein the failure of the chassis shock absorber is oil leakage of the chassis shock absorber or the damping of the chassis shock absorber is lower than the set damping threshold; Wherein the torque and power of the drive wheel of the new energy vehicle at the set test vehicle speed are measured by using the force sensing unit connected to the force arm of the stator and the single-roller automobile chassis dynamometer shell, which comprises a shell, a stator, a rotor, a loading device and a single roller. The loading device applies a braking torque to the rotor of the single roller through the stator, and at the same time, the stator receives the reaction torque of the rotor, which is measured by the force sensing unit and converted into the torque and power of the drive wheel of the new energy vehicle.

2. The new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion according to claim 1, wherein: In each learning performed on the convolutional neural network, the known mileage of a certain new energy vehicle from the test time to the failure time of the chassis shock absorber is taken as the single output content of the convolutional neural network, the set test speed, the related data of each test power, the torque and power of the driving wheel of the certain new energy vehicle at the test time, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the vehicle-mounted chassis of the certain new energy vehicle at the test time, and the body jitter amplitude of the certain new energy vehicle at the test time are taken as the multiple input contents of the convolutional neural network, and the learning is completed this time; Wherein, the overhead chassis image of the vehicle-mounted chassis of the new energy vehicle captured by the mobile visual sensor in the overhead mode includes: the overhead chassis image only includes the vehicle-mounted chassis of the new energy vehicle, and the mobile visual sensor is located on the side of the single-roller automobile chassis dynamometer and includes a mechanical arm, a positioning device, a driving device, a micro-control device and a super-definition camera device.

3. The new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion of claim 2, wherein, The system further comprises: A weight sensing device connected with the roller collecting device, for measuring the weight of the roller of the single-roller automobile chassis dynamometer, and sending the measured weight of the roller of the single-roller automobile chassis dynamometer to the roller collecting device; Wherein, the weight sensing device connected with the roller collecting device, for measuring the weight of the roller of the single-roller automobile chassis dynamometer, and sending the measured weight of the roller of the single-roller automobile chassis dynamometer to the roller collecting device includes: the weight sensing device is arranged below the single-roller automobile chassis dynamometer and only contacts the single-roller automobile chassis dynamometer during measurement.

4. The new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion of claim 2, wherein, The system further comprises: A visual analysis instrument connected with the roller collecting device, for analyzing the roller type of the single-roller automobile chassis dynamometer by using a visual analysis mechanism, and analyzing the roller type number of the single-roller automobile chassis dynamometer based on the analyzed roller type of the single-roller automobile chassis dynamometer; Wherein, the visual analysis instrument connected with the roller collecting device, for analyzing the roller type of the single-roller automobile chassis dynamometer by using a visual analysis mechanism, and analyzing the roller type number of the single-roller automobile chassis dynamometer based on the analyzed roller type of the single-roller automobile chassis dynamometer includes: the visual analysis instrument is arranged on the side of the roller of the single-roller automobile chassis dynamometer.

5. The new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion of claim 2, wherein, The system further comprises: A content transmission device connected with the health monitoring device, for receiving the remaining mileage of the new energy vehicle before the failure of the chassis shock absorber, and wirelessly transmitting the remaining mileage of the new energy vehicle before the failure of the chassis shock absorber to the remote vehicle test management server through a mobile communication link; Wherein, the content transmission device connected with the health monitoring device, for receiving the remaining mileage of the new energy vehicle before the failure of the chassis shock absorber, and wirelessly transmitting the remaining mileage of the new energy vehicle before the failure of the chassis shock absorber to the remote vehicle test management server through a mobile communication link includes: the mobile communication link is based on a time division duplex communication mode or a frequency division duplex communication mode.

6. The new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion of claim 2, wherein, The system further comprises: The field display device is connected with the health monitoring device, and is used to receive the remaining mileage of the new energy vehicle before the chassis shock absorber fails and display the remaining mileage of the new energy vehicle before the chassis shock absorber fails on site. The field display device is connected with the health monitoring device, and is used to receive the remaining mileage of the new energy vehicle before the chassis shock absorber fails and display the remaining mileage of the new energy vehicle before the chassis shock absorber fails on site.

7. The new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion according to any one of claims 2-6, characterized in that: The collected roller diameter, roller surface friction coefficient, roller weight and roller type number of the single-roller automobile chassis dynamometer are used as the various dynamometer-related data of the single-roller automobile chassis dynamometer, including that the roller type is a smooth roller, a knurled roller, a grooved roller or a coated roller, and different types have different roller type numbers; The collected roller diameter, roller surface friction coefficient, roller weight and roller type number of the single-roller automobile chassis dynamometer are used as the various dynamometer-related data of the single-roller automobile chassis dynamometer, including that the roller diameter of the single-roller automobile chassis dynamometer is between 1500 mm and 2500 mm; The intelligent chassis health monitoring model is a convolutional neural network after multiple learning, and the number of learning is monotonically positively correlated with the roller diameter, including that a number conversion formula is used to represent the numerical conversion relationship between the number of learning and the monotonically positive correlation with the roller diameter; The number conversion formula is used to represent the numerical conversion relationship between the number of learning and the monotonically positive correlation with the roller diameter, including that the roller diameter of the single-roller automobile chassis dynamometer is an input value of the number conversion formula; The number conversion formula is used to represent the numerical conversion relationship between the number of learning and the monotonically positive correlation with the roller diameter, including that the number of learning positively correlated with the roller diameter of the single-roller automobile chassis dynamometer is an output value of the number conversion formula.

8. The new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion according to any one of claims 2-6, characterized in that: The remaining mileage of the new energy vehicle before the chassis shock absorber fails is intelligently analyzed by using the intelligent chassis health monitoring model according to the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle, including: the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle are input to the intelligent chassis health monitoring model in parallel; The remaining mileage of the new energy vehicle before the chassis shock absorber fails is intelligently analyzed by using the intelligent chassis health monitoring model according to the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle, including: the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle are input to the intelligent chassis health monitoring model in parallel; The remaining mileage of the new energy vehicle before the chassis shock absorber fails is intelligently analyzed by using the intelligent chassis health monitoring model according to the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle, including: the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle are input to the intelligent chassis health monitoring model in parallel; The remaining mileage of the new energy vehicle before the chassis shock absorber fails is intelligently analyzed by using the intelligent chassis health monitoring model according to the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle, including: the set test vehicle speed, the various dynamometer correlation data, the torque and power of the driving wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle are input to the intelligent chassis health monitoring model in parallel; 9. A new energy vehicle-mounted chassis health monitoring system based on multi-source heterogeneous sensor fusion, the system comprising a memory and a plurality of processors, the memory storing a computer program configured to be executed by the plurality of processors to complete the following steps: Collecting the roller diameter, the roller surface friction coefficient, the roller weight and the roller type number of the single-roller automobile chassis dynamometer as various dynamometer correlation data of the single-roller automobile chassis dynamometer; Adopt the force sensor unit connected to the force arm of the stator and the single roller chassis dynamometer shell to measure the torque and power of the drive wheel of the new energy vehicle at the set test speed; Obtain the overhead chassis image of the vehicle-mounted chassis of the new energy vehicle shot by the mobile visual sensor in the overhead mode; Adopt the gyro sensor unit to measure the body jitter amplitude of the new energy vehicle located above the single roller chassis dynamometer; Use the intelligent chassis health monitoring model to intelligently analyze the remaining mileage of the new energy vehicle before the chassis shock absorber fails according to the set test speed, various dynamometer related data, the torque and power of the drive wheel of the new energy vehicle, the hue channel value, the brightness channel value, the saturation channel value, the imaging depth value and the coordinate value of each pixel point of the overhead chassis image of the vehicle-mounted chassis of the new energy vehicle, and the body jitter amplitude of the new energy vehicle. wherein The intelligent chassis health monitoring model is a convolutional neural network after learning, and the number of learning is monotonically positively correlated with the diameter of the roller; The structure of the convolutional neural network based on which the intelligent chassis health monitoring model is composed of an input layer, a convolutional layer, an activation layer, a pooling layer and a fully connected layer, and realizes data processing through hierarchical feature extraction; The failure of the chassis shock absorber refers to oil leakage of the chassis shock absorber or the damping of the chassis shock absorber being lower than the set damping threshold; The torque and power of the drive wheel of the new energy vehicle at the set test speed measured by the force sensor unit connected to the force arm of the stator and the single roller chassis dynamometer shell includes that the single roller chassis dynamometer includes a shell, a stator, a rotor, a loading device and a single roller. The loading device applies a braking torque to the rotor driving the single roller through the stator, at the same time, the stator receives the reaction torque of the rotor, the reaction torque is measured by the force sensor unit and converted into the torque and power of the drive wheel of the new energy vehicle.

Images