New energy automobile automatic test method and system

Abstract

The invention discloses a new energy automobile automatic test method and system, and relates to the technical field of electric automobile detection. A vehicle is fixed to a rotating hub platform, bus connection and robot calibration are completed, a digital twin model is constructed, and initial parameters are configured; decomposing a working condition curve to extract core parameters, and generating a refined control logic table in combination with the virtual model; the lower computer cooperates with the robot and the rotating hub platform to execute test operation according to a protocol; optimal control parameters are fed back based on real-time data and the virtual model; synchronously acquiring various parameters of the vehicle, the robot and the rotating hub; fusing the data and performing simulation comparison to generate a comprehensive test report; adjusting parameters and repeating the test to form a complete closed loop. Through multi-device cooperation, dynamic adaptation and data fusion, the test precision and stability are improved, the vehicle type adaptation capability is enhanced, the scene coverage range is expanded, the test safety is guaranteed, the test efficiency is improved, function extension and flexible operation and maintenance are supported, and the diversified test requirements of the new energy vehicle are met.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle testing technology, and in particular to an automatic testing method and system for new energy vehicles. Background Technology

[0002] Traditional testing of new energy vehicles primarily relies on two methods: manual operation and mechanical devices. While manual testing can flexibly handle complex scenarios and unexpected situations, making it suitable for performance evaluation and special condition verification in the early stages of new vehicle development, its accuracy is easily affected by differences in operator skill levels. Prolonged continuous testing can lead to operator fatigue, making it difficult to guarantee the consistency and repeatability of test results. Furthermore, manual testing is costly and time-consuming, failing to meet the standardized testing requirements of mass-produced vehicles. Mechanical devices, having undergone several generations of development, have evolved from simple fixed structures into complex systems integrating sensing and control functions. However, they still have inherent drawbacks: their complex structure leads to cumbersome installation processes, poor adaptability to different vehicle models, and the mechanical transmission characteristics can affect control performance, potentially even damaging the vehicle's internal structure. As testing requirements become more refined, they are gradually being replaced by more advanced control methods.

[0003] The rise of automated testing technologies has brought new solutions to the industry, mainly including hardware-in-the-loop (HIL), model-in-the-loop (MIL), software-in-the-loop (Software-in-the-loop), and virtual simulation methods. Hardware-in-the-loop technology improves testing safety and efficiency by connecting the actual control unit to a virtual model; MIL and Software-in-the-loop technologies support early problem discovery and rapid iterative optimization, deeply integrating into modern development processes; virtual simulation technology, based on the digital twin concept, significantly shortens the R&D cycle, reduces the need for physical prototype production, and has a clear advantage in hazardous condition testing. Driving robot technology has also evolved from mechanical structure-based to intelligent control-based. Third-generation products integrate intelligent algorithms and autonomous learning capabilities, improving control accuracy based on model predictive control strategies and possessing the potential to quickly adapt to different vehicle models. However, existing automated testing technologies still have many shortcomings: In terms of speed tracking control, traditional PID control methods for electric vehicle hub-driving robots are prone to problems such as large speed fluctuations and poor system adaptability under dynamic conditions. Even with the use of intelligent algorithms such as particle swarm optimization to optimize parameters, the control accuracy under complex conditions is still difficult to meet the requirements of high-precision testing. The driving robot's adaptability to different vehicle models is insufficient. When applied to new vehicle models, parameter calibration and system adjustment are required, and the teaching process is cumbersome and time-consuming. The efficiency and adaptability of existing self-learning capabilities are limited. Intelligent driving test solutions lack testing capabilities in complex real-world environments. When facing extreme conditions, abnormal scenarios, and complex traffic environments, the test coverage and confidence level need to be improved. In terms of human-machine interaction and safety, the process of safety personnel taking over vehicle control in autonomous driving tests is not fast and accurate enough, and is prone to misoperation or delays. The response speed and reliability in emergency situations are insufficient. Although CAN bus testing technology has realized the automatic generation of test projects, the degree of automation in test execution, result analysis, and problem localization is limited, making it difficult to meet the needs of large-scale, high-efficiency testing.

[0004] Existing technologies also have limitations: some focus on autonomous driving control based on operating condition curves or segmented vehicle speed control, solving only the control problem of a single parameter and lacking multi-device collaboration and dynamic adaptive adjustment mechanisms; while some patents have achieved accurate calculation of accelerator pedal opening, they have not incorporated technologies such as digital twins and multi-source data fusion, and cannot cope with multi-parameter coupling problems under complex operating conditions. With the rapid development of the new energy vehicle industry, the market has placed higher demands on the control accuracy, vehicle model adaptability, scenario coverage, safety, and automation level of testing systems. There is an urgent need for a new type of automatic testing method and system that integrates technologies such as multi-device collaboration, dynamic adaptive adjustment, digital twin feedback, and full-dimensional error compensation to overcome many pain points of existing technologies and provide efficient, accurate, and reliable testing support for the research and development and production of new energy vehicles. Summary of the Invention

[0005] This invention proposes an automatic testing method and system for new energy vehicles to solve the problems mentioned in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an automatic testing method for new energy vehicles, comprising: The test preparation steps are as follows: fix the new energy vehicle on the rotating platform, complete the CAN bus connection between the vehicle and the test system, perform visual positioning and zero-point calibration on the accelerator pedal robot, gear shifting robot and vehicle function triggering robot, build a digital twin model based on vehicle parameters, import the preset test condition curves and complete the initial parameter configuration of the rotating platform. In the working condition analysis step, the host computer decomposes the working condition curve in multiple dimensions, extracts the core parameters of vehicle speed, acceleration, slope and road resistance, and combines the digital twin model to simulate the vehicle dynamic response, generating a refined control logic table containing robot actions and hub loading strategies at each time node. The multi-device collaborative control process involves the lower-level machine receiving control commands based on a custom dual communication protocol, collaboratively controlling the accelerator pedal robot to adjust the opening degree through voltage signals, the gear shifting robot to complete gear switching, the vehicle's functions triggering the robot to execute button and touch screen operations, and the rotating platform simulating road resistance and slope in real time. The dynamic adaptive adjustment steps are based on real-time vehicle operation data and feedback from the digital twin model to calculate the working condition adaptation coefficient and dynamically adjust the robot's execution accuracy and hub loading parameters. The full-dimensional data acquisition process involves simultaneously collecting vehicle speed, motor torque, remaining battery power, voltage, current operating parameters, robot execution accuracy data, and hub simulation parameters through multi-channel acquisition units. The data fusion and analysis step involves using a multi-source data fusion algorithm to process the collected data, combining it with a digital twin model for simulation comparison, identifying test anomalies, and generating a comprehensive test report that includes performance indicators, operating condition tracking errors, and parameter change trends. The test closed-loop optimization steps involve adjusting the operating condition curve parameters and control strategies based on the test report, and re-executing the test process until the preset test accuracy requirements are met.

[0007] Furthermore, it also includes a dynamic operating condition adaptation step. Based on the vehicle's operating status and environmental parameters, an operating condition adaptation coefficient calculation model is constructed. Real-time optimization of the control strategy is achieved through multi-parameter coupled calculation, expressed as follows: in For dynamic operating condition adaptation coefficient, , , , These are weighted coefficients, and their sum is 1. For real-time vehicle speed, The standard operating speed is [not specified]. For real-time acceleration, For standard acceleration under operating conditions, For real-time road resistance, For standard road resistance, Remaining battery power The battery is fully charged. To simulate the slope of the rotating hub, multi-parameter coupled calculation can integrate vehicle dynamics, energy status and environmental simulation parameters to dynamically adapt to the control requirements of different test conditions.

[0008] Furthermore, it also includes a multi-robot collaborative calibration step. The calibration process is initiated periodically before and during the test. The host computer issues calibration instructions, and the slave computer controls the accelerator pedal robot, the gear shifting robot, and the vehicle's infotainment system to trigger the robots to perform preset calibration actions. The actual position data of the robots is collected through the vision positioning system, compared with the standard position data to calculate the calibration deviation, generate calibration coefficients, and send them to each robot execution unit to correct the position error and response delay of the actuator. The calibration cycle can be dynamically set according to the test duration and the robot's operating status.

[0009] Furthermore, it also includes a real-time feedback step for digital twins. Based on the vehicle's physical parameters and test process data, the operating status of the digital twin model is dynamically updated. The dynamic response, energy consumption, and component wear of the vehicle under the current control parameters are simulated. The simulation results are compared with the actual vehicle test data in real time to calculate the deviation value. When the deviation value exceeds the preset threshold, the control parameter adjustment command is automatically triggered to correct the robot's execution actions and hub loading strategy.

[0010] Furthermore, it includes a multi-dimensional error compensation step. A comprehensive error compensation model is constructed to address sensor measurement errors, robot execution errors, and hub simulation errors during the testing process. The control commands are corrected using a nonlinear function, expressed as follows: in To control the amount of instruction compensation, to For error compensation weighting coefficients, For sensor displacement measurement error, To perform attitude angle deviation for the robot, Due to vehicle speed measurement error, For test duration, This refers to the actual ambient temperature. For standard test temperature, To compensate for the error in the load applied to the hub, nonlinear compensation is performed by considering multiple sources of error from various dimensions.

[0011] Furthermore, it also includes test safety monitoring steps, real-time monitoring of vehicle operating status, battery voltage and temperature, motor operating parameters, and test system communication status, setting multiple safety thresholds, and immediately triggering a safety protection mechanism when parameters are detected to exceed the safety range or communication is interrupted. The host computer issues an emergency stop command, the slave computer controls the robot to reset and the hub to stop loading, and at the same time cuts off the vehicle's power input, records abnormal data and the time point of the fault, generates a safety warning report and pushes it to the test terminal.

[0012] Furthermore, it also includes a test strategy adaptive optimization step. Based on multiple rounds of test data and report results, reinforcement learning algorithms are used to iteratively optimize the working condition analysis logic, multi-device collaborative control parameters, and dynamic adjustment strategies, and establish a mapping relationship between test scenarios and optimal control strategies. When a new test condition is introduced or the test vehicle is changed, the system can automatically call the optimal strategy and quickly adapt, shortening test preparation time, improving test efficiency and result consistency. The optimized strategy can be stored in the strategy library for subsequent use.

[0013] An automated testing system for new energy vehicles includes: The host computer system uses an interactive GUI interface developed based on MATLAB. It has built-in modules for working condition analysis, digital twin modeling, control strategy generation, and test management. It supports the import of various standard working condition curves and the editing of custom working conditions. The digital twin modeling module constructs a virtual model of the vehicle and the test environment. Combined with the working condition analysis module, it generates a refined control logic table. The test management module realizes test task creation, progress monitoring, and report generation. The lower-level system, with the PXIe controller as its core, is equipped with a multi-functional I / O module, a CAN / LIN interface module, a timer / counter module, and a built-in multi-device collaborative control module, a dynamic adjustment module, and a communication management module. It achieves high-speed communication with the upper-level computer and execution units through a custom JSON command protocol and a binary real-time data protocol, and coordinates the precise execution of commands by each execution device. The dynamic adjustment module optimizes control parameters based on real-time data feedback. The multi-robot execution unit includes an accelerator pedal robot, a gear shifting robot, and a vehicle function triggering robot. The accelerator pedal robot controls the opening of the accelerator pedal through voltage signals. The gear shifting robot supports automatic switching between column shifter models. The vehicle function triggering robot integrates a 4-DOF robotic arm and a vision positioning system to achieve automated operation of the vehicle screen and steering wheel buttons. Each robot has a built-in independent calibration module. The hub simulation unit consists of a hub body, a loading system, a slope simulation system, and a data acquisition submodule. It can simulate different road resistance, loading mass, and slope conditions. The data acquisition submodule collects hub rotation speed and loading force parameters in real time and feeds them back to the lower-level computer. The data acquisition unit integrates multi-channel sensors, a CAN bus acquisition card, and an environmental monitoring module. It supports the acquisition of vehicle speed, motor torque, remaining battery power, voltage, current operating parameters, robot execution data, and environmental temperature and humidity data. The built-in data preprocessing module performs filtering and noise reduction. The data processing unit adopts an architecture that combines edge computing and cloud storage. Real-time data fusion and preliminary analysis are achieved at the edge, while the cloud performs in-depth mining of multi-source data based on big data algorithms to generate a comprehensive test report that includes performance indicators, error analysis, and optimization suggestions. The report can be exported and historical data can be traced. The safety monitoring unit consists of a status monitoring module, an emergency protection module, and an early warning push module. It monitors vehicle and system operating parameters in real time, sets multi-level safety thresholds, and quickly executes shutdown and reset operations when the safety protection mechanism is triggered. The early warning push module pushes abnormal information through audible and visual alarms and terminal notifications.

[0014] Furthermore, it also includes a remote operation and maintenance module, which supports remote status monitoring, parameter configuration, and fault diagnosis of the test system via the network. Operation and maintenance personnel can remotely view the test progress, equipment operating status, and data acquisition status, and issue operating condition adjustment and system calibration instructions. When a system failure occurs, the remote operation and maintenance module automatically uploads fault codes and operating logs, provides fault diagnosis suggestions, and supports multi-user permission management.

[0015] Furthermore, it also includes an expansion interface module, which reserves access interfaces for wearable devices, wireless sensor networks, and communication interfaces for third-party testing equipment. It supports access to external devices such as dedicated battery management system testing equipment, motor efficiency testing instruments, and high-precision environmental simulation devices, and expands the functions of battery cycle life testing, motor performance testing, and extreme environment adaptability testing. The interface supports plug-and-play and automatic adaptation, and the system software supports flexible expansion of functional modules to meet the diverse adaptation needs of different testing scenarios.

[0016] Compared with existing technologies, the beneficial effects of this invention are: The automatic testing method and system for new energy vehicles of the present invention, through the integration of multiple technologies and innovative design, comprehensively solves the problems of insufficient control precision, poor vehicle adaptability, limited scenario coverage, poor safety, and low degree of automation in existing testing technologies. It achieves all-round improvement in testing precision, adaptability, scenario coverage, safety assurance, testing efficiency, and extended operation and maintenance, and has significant technical advantages and application value.

[0017] In terms of testing accuracy and stability, the system employs a multi-device collaborative control and dynamic adaptive adjustment strategy, combined with real-time feedback from a digital twin model, to dynamically optimize robot execution accuracy and hub loading parameters, effectively improving tracking lag and parameter mismatch. A multi-dimensional error compensation model provides nonlinear compensation for sensor measurement errors, robot execution errors, and hub simulation errors, significantly reducing the impact of system errors on test results and substantially improving control accuracy and data reliability. A layered architecture and distributed collaborative design ensure consistent action across all devices, while a custom dual communication protocol meets the different needs of non-real-time command interaction and high-frequency real-time data transmission, further guaranteeing the stability and accuracy of the testing process.

[0018] In terms of vehicle model adaptation and testing efficiency, the system integrates multi-robot collaborative calibration and controller parameter self-learning strategies, enabling it to quickly learn the characteristics of different vehicle models, reduce manual intervention, shorten the teaching time for new model adaptation, and significantly enhance the ability to quickly adapt to different vehicle models. The adaptive optimization steps of the testing strategy are based on multiple rounds of test data and report results. Through reinforcement learning algorithms, iterative optimization of the operating condition analysis logic, multi-device collaborative control parameters, and dynamic adjustment strategies is achieved, establishing a mapping relationship between test scenarios and the optimal control strategy. When testing new operating conditions or new vehicle models, the optimal strategy can be automatically invoked and quickly adapted, significantly shortening test preparation time and improving test efficiency and result consistency. The fully automated design achieves automation of all stages from test preparation, operating condition analysis, multi-device collaborative control, data acquisition, fusion analysis to closed-loop optimization, reducing manual intervention, shortening the test cycle, and meeting the needs of large-scale testing.

[0019] In terms of scenario coverage and safety assurance, the all-dimensional data acquisition unit supports the simultaneous acquisition of multiple types of data, including vehicle speed, motor torque, battery parameters, robot execution data, and hub simulation parameters. Combined with multi-source data fusion algorithms and digital twin model simulation comparison, it can accurately identify test anomalies, effectively expand the test scenario coverage, and improve the testing capabilities for extreme working conditions and abnormal scenarios. The safety monitoring unit monitors the vehicle's operating status, battery parameters, motor operating parameters, and system communication status in real time. It sets multiple safety thresholds and quickly executes operations such as shutdown, robot reset, and disconnection of vehicle power input when the safety protection mechanism is triggered. At the same time, it records abnormal data and fault time points and pushes early warning reports to comprehensively ensure the safety of personnel and equipment during the testing process.

[0020] In terms of operation and maintenance (O&M) and expansion, the modular design allows each functional module to work independently. Maintenance only requires replacing or repairing the faulty module, eliminating the need for system downtime and significantly improving system availability and maintenance efficiency. The expansion interface module provides access to various external devices, supporting connection to dedicated battery management system testing equipment, motor efficiency testing instruments, high-precision environmental simulation devices, etc. It can be expanded to include battery cycle life testing, motor performance testing, and extreme environment adaptability testing. The interfaces support plug-and-play and automatic adaptation, and the system software supports flexible expansion of functional modules to meet diverse testing scenarios and needs. The remote O&M module supports remote status monitoring, parameter configuration, and fault diagnosis of the test system. It automatically uploads fault codes and operation logs and provides diagnostic suggestions, shortening fault repair time and improving O&M efficiency. Multi-user access control ensures the security of remote operations. Attached Figure Description

[0021] Figure 1 This is a schematic block diagram of an automatic testing method for new energy vehicles proposed in this invention; Figure 2 This is a schematic block diagram of an automatic testing system for new energy vehicles proposed in this invention; Figure 3 This is a schematic diagram showing the components and connections of an automatic testing system for new energy vehicles proposed in this invention. Figure 4 This is a schematic diagram illustrating the working principle of an automatic testing system for new energy vehicles proposed in this invention. Figure 5 This is a comparison chart of vehicle speed tracking errors using different testing methods proposed in this invention; Figure 6 This is a comparison chart of the adaptation time for different vehicle models proposed in this invention; Figure 7 This is a comparison chart of the execution accuracy of multiple robots proposed in this invention; Figure 8 This is a comparison chart showing the efficiency improvements of different test items proposed in this invention; Figure 9 This is a comparison chart of the data acquisition integrity proposed in this invention. Detailed Implementation

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

[0023] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0024] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0025] Reference Figures 1 to 9 An automatic testing method for new energy vehicles. The test preparation steps are as follows: fix the new energy vehicle on the rotating platform, complete the CAN bus connection between the vehicle and the test system, perform visual positioning and zero-point calibration on the accelerator pedal robot, gear shifting robot and vehicle function triggering robot, build a digital twin model based on vehicle parameters, import the preset test condition curves and complete the initial parameter configuration of the rotating platform. In the working condition analysis step, the host computer decomposes the working condition curve in multiple dimensions, extracts the core parameters of vehicle speed, acceleration, slope and road resistance, and combines the digital twin model to simulate the vehicle dynamic response, generating a refined control logic table containing robot actions and hub loading strategies at each time node. The multi-device collaborative control process involves the lower-level machine receiving control commands based on a custom dual communication protocol, collaboratively controlling the accelerator pedal robot to adjust the opening degree through voltage signals, the gear shifting robot to complete gear switching, the vehicle's functions triggering the robot to execute button and touch screen operations, and the rotating platform simulating road resistance and slope in real time. The dynamic adaptive adjustment steps, based on real-time vehicle operation data and digital twin model feedback, calculate the working condition adaptation coefficient, dynamically adjust the robot's execution accuracy and hub loading parameters, and improve the phenomenon of working condition tracking lag and parameter mismatch. The full-dimensional data acquisition process involves synchronously collecting vehicle speed, motor torque, remaining battery power, voltage, current, robot execution accuracy data, and hub simulation parameters through multi-channel acquisition units, with an acquisition frequency of no less than 100Hz. The data fusion and analysis step involves using a multi-source data fusion algorithm to process the collected data, combining it with a digital twin model for simulation comparison, identifying test anomalies, and generating a comprehensive test report that includes performance indicators, operating condition tracking errors, and parameter change trends. The test closed-loop optimization steps involve adjusting the operating condition curve parameters and control strategies based on the test report, re-executing the test process until the preset test accuracy requirements are met, thus forming a test-analysis-optimization closed loop.

[0026] This invention also includes a dynamic operating condition adaptation step. Based on the vehicle's operating state and environmental parameters, an operating condition adaptation coefficient calculation model is constructed. Real-time optimization of the control strategy is achieved through multi-parameter coupled calculation, expressed as follows: in For dynamic operating condition adaptation coefficient, , , , These are weighted coefficients, and their sum is 1. For real-time vehicle speed, The standard operating speed is [not specified]. For real-time acceleration, For standard acceleration under operating conditions, For real-time road resistance, For standard road resistance, Remaining battery power The battery is fully charged. To simulate the slope of the rotating hub, multi-parameter coupled calculation can integrate vehicle dynamics, energy status and environmental simulation parameters, dynamically adapt to the control requirements of different test conditions, and improve the accuracy of condition tracking.

[0027] This invention also includes a multi-robot collaborative calibration step. The calibration process is initiated periodically before and during the test. The host computer issues calibration commands, and the slave computer controls the accelerator pedal robot, the gear shifting robot, and the vehicle's infotainment system to trigger the robots to perform preset calibration actions. The actual position data of the robots is collected through the vision positioning system, compared with the standard position data to calculate the calibration deviation, generate calibration coefficients, and send them to each robot execution unit to correct the position error and response delay of the actuators, ensuring the consistency of multi-robot action coordination. The calibration cycle can be dynamically set according to the test duration and robot operating status, and the time for a single calibration does not exceed 30 seconds.

[0028] This invention also includes a real-time feedback step using a digital twin. Based on vehicle physical parameters and test process data, the operating status of the digital twin model is dynamically updated. The dynamic response, energy consumption, and component wear of the vehicle under the current control parameters are simulated. The simulation results are compared with the actual vehicle test data in real time, and the deviation value is calculated. When the deviation value exceeds a preset threshold, the control parameter adjustment command is automatically triggered to correct the robot's execution actions and hub loading strategy, thereby realizing real-time linkage between virtual and reality and improving the stability and controllability of the test process.

[0029] This invention also includes a multi-dimensional error compensation step. A comprehensive error compensation model is constructed to address sensor measurement errors, robot execution errors, and hub simulation errors during the testing process. The control commands are corrected using a nonlinear function, expressed as follows: in To control the amount of instruction compensation, to For error compensation weighting coefficients, For sensor displacement measurement error, To perform attitude angle deviation for the robot, Due to vehicle speed measurement error, For test duration, This refers to the actual ambient temperature. For standard test temperature, To compensate for the error in the load applied to the hub, nonlinear compensation is performed by considering multiple error sources from various dimensions. This can effectively reduce the impact of system errors on test results and improve control accuracy and data reliability.

[0030] This invention also includes a test safety monitoring step, which monitors the vehicle's operating status, battery voltage and temperature, motor operating parameters, and test system communication status in real time. It sets multiple safety thresholds, and when parameters are detected to exceed the safety range or communication is interrupted, a safety protection mechanism is immediately triggered. The host computer issues an emergency stop command, the slave computer controls the robot to reset and the rotating hub to stop loading, and at the same time cuts off the vehicle's power input. Abnormal data and the time of failure are recorded, a safety warning report is generated and pushed to the test terminal, ensuring the safety of personnel and equipment during the test process.

[0031] This invention also includes a test strategy adaptive optimization step. Based on multiple rounds of test data and report results, a reinforcement learning algorithm is used to iteratively optimize the working condition analysis logic, multi-device collaborative control parameters, and dynamic adjustment strategy. A mapping relationship between the test scenario and the optimal control strategy is established. When a new test condition is introduced or the test vehicle is changed, the system can automatically call the optimal strategy and perform rapid adaptation, shortening the test preparation time, improving test efficiency and result consistency. The optimized strategy can be stored in the strategy library for subsequent use.

[0032] This invention also discloses an automatic testing system for new energy vehicles, comprising: The host computer system uses an interactive GUI interface developed based on MATLAB. It has built-in modules for working condition analysis, digital twin modeling, control strategy generation, and test management. It supports the import of various standard working condition curves and the editing of custom working conditions. The digital twin modeling module constructs a virtual model of the vehicle and the test environment. Combined with the working condition analysis module, it generates a refined control logic table. The test management module realizes test task creation, progress monitoring, and report generation. The lower-level system, with the PXIe controller as its core, is equipped with a multi-functional I / O module, a CAN / LIN interface module, a timer / counter module, and a built-in multi-device collaborative control module, a dynamic adjustment module, and a communication management module. It achieves high-speed communication with the upper-level computer and execution units through a custom JSON command protocol and a binary real-time data protocol, and coordinates the precise execution of commands by each execution device. The dynamic adjustment module optimizes control parameters based on real-time data feedback. The system comprises multiple robotic execution units, including an accelerator pedal robot, a gear shifting robot, and a vehicle infotainment system function triggering robot. The accelerator pedal robot controls the accelerator pedal opening via voltage signals with a control accuracy of ±0.1%FS. The gear shifting robot supports automatic switching between column shifter and column shifter models with a repeatability accuracy of ±0.05mm. The vehicle infotainment system function triggering robot integrates a 4-DOF robotic arm and a vision positioning system to automate the operation of the vehicle infotainment screen and steering wheel buttons. Each robot has a built-in independent calibration module. The hub simulation unit consists of a hub body, a loading system, a slope simulation system and a data acquisition submodule. It can simulate different road resistance, loading mass and slope conditions. The loading system response time is ≤10ms, the slope simulation range is 0-20%, and the data acquisition submodule collects the hub rotation speed and loading force parameters in real time and feeds them back to the lower computer. The data acquisition unit integrates multi-channel sensors, a CAN bus acquisition card, and an environmental monitoring module. It supports the acquisition of vehicle speed, motor torque, remaining battery power, voltage, current operating parameters, robot execution data, and environmental temperature and humidity data. The acquisition frequency can be adjusted within the range of 50-200Hz, and the data transmission delay is ≤10ms. The built-in data preprocessing module performs filtering and noise reduction. The data processing unit adopts an architecture that combines edge computing and cloud storage. Real-time data fusion and preliminary analysis are achieved at the edge, while the cloud performs in-depth mining of multi-source data based on big data algorithms to generate a comprehensive test report that includes performance indicators, error analysis, and optimization suggestions. The report can be exported and historical data can be traced. The safety monitoring unit consists of a status monitoring module, an emergency protection module, and an early warning push module. It monitors vehicle and system operating parameters in real time, sets multi-level safety thresholds, and quickly executes shutdown and reset operations when the safety protection mechanism is triggered. The early warning push module pushes abnormal information through audible and visual alarms and terminal notifications.

[0033] This invention also includes a remote operation and maintenance module, which supports remote status monitoring, parameter configuration, and fault diagnosis of the test system via the network. Operation and maintenance personnel can remotely view the test progress, equipment operating status, and data acquisition, and issue operating condition adjustment and system calibration commands. When a system failure occurs, the remote operation and maintenance module automatically uploads fault codes and operating logs, provides fault diagnosis suggestions, shortens fault repair time, improves system operation and maintenance efficiency, supports multi-user permission management, and ensures remote operation security.

[0034] This invention also includes an expansion interface module, which reserves access interfaces for wearable devices, wireless sensor networks, and communication interfaces for third-party testing equipment. It supports access to external devices such as dedicated battery management system testing equipment, motor efficiency testing instruments, and high-precision environmental simulation devices. It can expand the functions of battery cycle life testing, motor performance testing, and extreme environment adaptability testing. The interface supports plug-and-play and automatic adaptation. The system software supports flexible expansion of functional modules to meet the diverse adaptation needs of different testing scenarios.

[0035] The following two examples further illustrate specific embodiments of the present invention: Example 1 New energy vehicle NEDC range test This embodiment focuses on the NEDC range test of a compact pure electric sedan equipped with a ternary lithium battery, boasting a maximum nominal range of 400 kilometers. The test aims to verify the consistency between the actual range performance and the nominal value, while also examining the stability and accuracy of the testing system under long-term continuous operating conditions. The test scenarios cover typical conditions such as urban congestion, suburban constant speed driving, and highway driving, comprehensively evaluating the vehicle's battery energy consumption, motor efficiency, and overall vehicle dynamic response.

[0036] During the test preparation phase, the test vehicle was slowly driven onto the rotating platform, and the wheels were secured using a dedicated fixing device to ensure no displacement of the vehicle during the test. A CAN bus connection was established between the vehicle and the test system, enabling data interaction between the vehicle controller and the lower-level machine. Visual positioning was performed on the accelerator pedal robot, gear shifting robot, and vehicle infotainment function triggering robot. High-definition vision cameras captured the spatial positions of the vehicle's pedals, gear lever, and infotainment screen, automatically performing zero-point calibration to ensure the robot's motion accuracy met requirements. A digital twin model was constructed based on vehicle parameters including curb weight, drag coefficient, transmission efficiency, and battery capacity. The test process was rehearsed in a virtual environment to verify the adaptability to operating conditions. NEDC standard operating condition curves were imported, and initial parameters for the rotating platform were configured: road drag coefficient 0.015, load mass 1500 kg, road slope 0 degrees, data acquisition frequency set to 100 Hz, and acquisition channels covering more than 20 parameters including vehicle speed, motor torque, remaining battery power, voltage, current, motor temperature, and ambient temperature and humidity.

[0037] During the operating condition analysis phase, the host computer decomposes the NEDC operating condition curve in multiple dimensions, extracting core parameters such as target vehicle speed, acceleration, and road resistance at each time point. The urban operating condition includes four cycles with a maximum speed of 50 km / h, while the suburban operating condition has a maximum speed of 120 km / h. Combining a digital twin model to simulate the vehicle's dynamic response under different operating conditions, the required throttle opening, braking intensity, and gear status for each stage are calculated. This generates a refined control logic table, clarifying the voltage output range of the accelerator pedal robot, the timing of the gear shifting robot's actions, and the loading strategy of the hub platform. The time step of the control logic table is accurate to 10 milliseconds.

[0038] In the multi-device collaborative control phase, the lower-level machine receives the upper-level machine's control logic table via a custom JSON command protocol and achieves high-frequency data transmission based on a binary real-time data protocol. The collaboratively controlled accelerator pedal robot adjusts the accelerator pedal opening via voltage signals; during acceleration in urban conditions, the voltage gradually increases from 0.5 volts to 3.8 volts, while maintaining around 2.2 volts during constant speed. The gear-shifting robot automatically switches between N and D gears during condition changes, with a shift speed controlled within 1 second and a repeatability accuracy of ±0.05 mm. The vehicle-mounted system triggering robot operates the air conditioning, lights, and other equipment according to the logic table requirements, simulating real driving scenarios. The rotating platform tracks the operating condition curve in real time, dynamically adjusting road resistance and loading mass to ensure the simulated environment matches the actual road conditions.

[0039] During the dynamic adaptive adjustment phase, based on real-time vehicle operating data and feedback from the digital twin model, the operating condition adaptation coefficient is calculated. When the deviation between the actual vehicle speed and the target speed exceeds 0.5 km / h, the execution accuracy of the accelerator pedal robot and the hub loading parameters are automatically adjusted. For example, in suburban highway conditions, if the actual vehicle speed is lower than the target speed, the system increases the accelerator pedal opening voltage while reducing the hub loading resistance to ensure operating condition tracking accuracy. The multi-robot collaborative calibration process is initiated every 30 minutes. It collects the actual position data of the robots through the visual positioning system, compares it with the standard position to generate calibration coefficients, and corrects position errors and response delays.

[0040] During the full-dimensional data acquisition phase, multi-channel acquisition units simultaneously collect vehicle operating parameters, robot execution data, and hub simulation parameters. Key parameters such as vehicle speed, motor torque, and battery voltage and current are acquired at a frequency of 100 Hz, while parameters such as motor temperature and ambient temperature and humidity are acquired at a frequency of 10 Hz. All data is transmitted in real-time to the host computer and server via a high-speed network, with data transmission latency controlled within 10 milliseconds. A multi-dimensional error compensation mechanism is activated during the acquisition process to correct for sensor measurement errors and robot execution errors, ensuring data reliability.

[0041] In the data fusion and analysis phase, multi-source data fusion algorithms are used to process the collected data, and simulation comparisons are performed using digital twin models to identify test anomalies such as sudden changes in battery voltage and fluctuations in motor torque. A comprehensive test report is generated through data analysis, including basic test information, driving range results, average energy consumption, battery remaining charge change curves, vehicle speed tracking error curves, and motor torque change trends, clearly indicating the time, cause, and impact of any anomalies.

[0042] During the closed-loop optimization phase of the test, the test report revealed a slightly large speed tracking error during low-speed driving in urban conditions. The acceleration parameters and robot control strategy for the low-speed segment in the condition analysis logic were adjusted, and the weight of speed in the dynamic condition adaptation coefficient was increased. The test process was re-executed, and the speed tracking error in the second test was significantly reduced, meeting the preset accuracy requirements and forming a complete test-analysis-optimization closed loop. During the test, the safety monitoring unit monitored the battery temperature and voltage in real time. When the battery temperature exceeded 45 degrees Celsius, a cooling warning was automatically triggered, and no safety abnormalities occurred.

[0043] Table 1 Comparison of NEDC driving range test results Table 1 clearly demonstrates the significant advantages of the system of this invention in range testing. Traditional manual testing is affected by differences in driver operation, resulting in large speed tracking errors, long testing cycles, incomplete data collection, and poor repeatability and anomaly identification capabilities. While automated testing with a single device offers some improvement, it lacks multi-device collaboration and dynamic adjustment mechanisms, resulting in insufficient accuracy and completeness. The system of this invention, through multi-device collaborative control, dynamic adaptive adjustment, and multi-dimensional error compensation technologies, significantly reduces speed tracking errors, shortens testing cycles, improves the completeness of data collection and the repeatability of test results, and greatly enhances the accuracy of anomaly identification. It can accurately capture potential problems during the testing process, providing reliable data support for optimizing vehicle range performance and fully demonstrating the system's high precision, high efficiency, and high stability.

[0044] Example 2 Comprehensive test of the power performance of new energy vehicles This embodiment focuses on a comprehensive power performance test of a mid-size pure electric SUV. The test items cover three typical operating conditions: 0-100km / h acceleration, 100-0km / h braking, and 20% gradient climbing. The aim is to comprehensively evaluate the vehicle's acceleration capability, braking performance, and climbing power reserve, and to verify the stability and reliability of the vehicle's powertrain under extreme conditions. This model is equipped with a dual-motor four-wheel drive system with a maximum output power of 200 kW. The test needs to simulate the power response under different loads and road conditions.

[0045] During the test preparation phase, the test vehicle was fixed on the rotating platform, ensuring the vehicle's center of gravity was aligned with the center of the rotating platform. The vehicle's CAN bus was connected to the test system to achieve real-time reading of power system parameters. Visual positioning and zero-point calibration were performed on the three functional robots: the accelerator pedal robot calibrated its travel range from 0-100%, the gear shifting robot calibrated gear shifting for the SUV's column shifter structure, and the vehicle infotainment system trigger robot calibrated the operational accuracy of the steering wheel buttons and the central control screen. A digital twin model was constructed based on the vehicle's power parameters, including motor power, torque, transmission ratio, and climbing ability. Acceleration, braking, and climbing curves were imported, and rotating platform parameters were configured. The acceleration test involved a load of 1800 kg, the braking test a road resistance coefficient of 0.02, and the climbing test a 20% gradient. The data acquisition frequency was increased to 200 Hz, focusing on collecting parameters such as acceleration, braking force, motor power, and battery output current.

[0046] During the operational condition analysis phase, the host computer performs segmented analysis on the curves for the three operational conditions. For the acceleration condition, it extracts the target acceleration (0-100 km / h) and throttle opening rate of change; for the braking condition, it extracts the braking deceleration and brake pedal actuation timing; and for the climbing condition, it extracts the target vehicle speed and motor torque requirements. Combining a digital twin model, it simulates the vehicle's dynamic response under different operational conditions, generating a refined control logic table containing robot motion parameters at each time point and a hub loading strategy. In the acceleration condition, the time for the throttle pedal opening to rapidly increase from 0% to 100% is controlled within 0.5 seconds; in the braking condition, the hub loading resistance gradually increases to simulate the actual braking effect; and in the climbing condition, the hub platform maintains a 20% gradient load.

[0047] In the multi-device collaborative control phase, after receiving instructions from the host computer, the lower-level machine collaboratively controls the actions of each execution unit. During acceleration testing, the shifting robot switches to D gear, the accelerator pedal robot quickly adjusts the accelerator pedal opening according to the control logic table, and the rotating platform maintains the set load mass to simulate road resistance in real time. During braking testing, the accelerator pedal robot returns to zero, the rotating platform activates the auxiliary braking function, and adjusts the resistance according to the braking deceleration requirements. During hill climbing testing, the accelerator pedal robot maintains a stable opening to ensure sufficient torque output from the motor, and the rotating platform maintains a 20% slope load to simulate hill climbing resistance. Throughout the entire process, the lower-level machine achieves high-speed communication with each device through a custom dual protocol to ensure consistent and coordinated actions.

[0048] During the dynamic adaptive adjustment phase, control parameters are dynamically adjusted based on real-time vehicle operating data and feedback from the digital twin model. In acceleration testing, if the actual acceleration is lower than the target value, the system automatically increases the accelerator pedal opening adaptation coefficient to enhance motor output torque. In braking testing, the system adjusts the hub assist braking force according to the actual braking distance. In hill-climbing testing, if the vehicle speed decreases by more than 1 km / h, the system optimizes the accelerator pedal control strategy to maintain stable speed. Multi-robot collaborative calibration is initiated every 20 minutes to correct robot execution errors and ensure test accuracy. The digital twin model updates the vehicle status in real time, compares it with real-vehicle data to calculate deviations, and automatically triggers control parameter adjustments when deviations exceed a threshold.

[0049] During the full-dimensional data acquisition phase, multi-channel acquisition units simultaneously collect parameters such as instantaneous acceleration, motor power, and battery output current during acceleration; braking force, braking distance, and braking time during braking; and vehicle speed, motor torque, and battery voltage during hill climbing. The acquisition frequency of 200 Hz ensures the capture of instantaneous dynamic responses. The data preprocessing module filters and reduces noise in the acquired data to remove interference signals and ensure data accuracy. The safety monitoring unit monitors motor temperature, battery voltage, and current in real time, setting a maximum motor temperature threshold of 80 degrees Celsius and a maximum battery output current threshold of 300 amperes. Exceeding these thresholds immediately triggers an early warning and adjusts the test parameters.

[0050] In the data fusion and analysis phase, a multi-source data fusion algorithm is used to integrate the collected data from three operating conditions. This data is then combined with a digital twin model for simulation comparison to identify powertrain anomalies such as motor torque fluctuations during acceleration, uneven braking force during braking, and battery voltage drops during hill climbing. A comprehensive power performance test report is generated, including test results for each operating condition, dynamic parameter variation curves, anomaly analysis, and performance evaluation, providing targeted suggestions for optimizing the vehicle's powertrain.

[0051] During the closed-loop optimization phase of the test, the test report revealed unstable torque output at low speeds during the climbing condition. The climbing segment parameters and dynamic adaptation coefficients in the condition analysis logic were adjusted to optimize the motor torque control strategy. After re-executing the climbing test, the torque fluctuation issue was significantly improved, and all indicators met the preset requirements. Based on the test data, the test strategy adaptive optimization module updated the optimal control strategy for the power performance test and stored it in the strategy library for future reference in tests of similar models.

[0052] Table 2 Comparison of Comprehensive Power Performance Test Results Table 2 data fully demonstrates the technical advantages of the system of this invention in power performance testing. Traditional manual testing is affected by operating skills and reaction speed, resulting in poor repeatability of acceleration tests, large braking distance errors, and difficulty in accurately capturing power system anomalies. Although automated testing of a single working condition eliminates reliance on manual labor, it lacks the ability to coordinate multiple working conditions and dynamically adjust, and its torque stability and anomaly identification coverage are insufficient. The system of this invention, through multi-device collaborative control, real-time feedback from digital twins, and multi-dimensional error compensation technologies, significantly improves the repeatability of acceleration tests and the accuracy of braking distance control, enhances torque stability under climbing conditions, and comprehensively covers power system anomaly scenarios. At the same time, the system integrates multi-working-condition testing processes, significantly improving testing efficiency and enabling rapid completion of multi-dimensional power performance evaluation. It provides comprehensive, accurate, and efficient testing support for the research and development and improvement of vehicle power systems, fully demonstrating the system's adaptability and technological advancement under complex working conditions.

[0053] refer to Figure 3 The diagram clearly illustrates the overall structure and connections between the components of the new energy vehicle rotary drum testing system. As shown, the host computer communicates with the slave computer via test commands and feedback data; the slave computer receives commands from the host computer and simultaneously acquires data; the main controller, as the core control unit, receives control signals from the slave computer and sends control commands to the robot system and the test vehicle; the robot system, including actuators such as throttle, brake, touchscreen, and gear shifting, directly operates the test vehicle; the test vehicle is placed on the rotary drum platform, which simulates the actual road driving environment by simulating road resistance, load mass, and road gradient. The entire system forms a closed-loop control system, ensuring the accuracy and reliability of the test.

[0054] refer to Figure 4 This diagram details the complete flow of the lower-level LabVIEW program, including six main modules: upper-lower-level TCP communication, command parsing, action execution, lower-level machine-device TCP communication, data acquisition, parsing, saving, and uploading, and data processing result uploading. The upper-lower-level TCP communication module connects to the upper-level machine, receives commands, performs CRC checks, and generates corresponding feedback based on the check results. The command parsing module parses the received commands, distinguishes between different types of commands such as TE, TC, and EM, and generates corresponding feedback. The action execution module maps actions and configures parameters based on the command content. The lower-level machine-device TCP communication module is responsible for issuing action commands and acquiring status. The data acquisition module is responsible for acquiring parameter configurations, parsing data, and generating and uploading data. The data processing module is responsible for data processing script mapping, data processing, and generating and uploading data. The entire program flow forms a complete closed-loop control system, ensuring accurate execution of the experiment and reliable data acquisition.

[0055] refer to Figure 5This invention visually demonstrates its core advantage in vehicle speed tracking accuracy, overcoming the accuracy bottlenecks of traditional testing and single automated testing. Traditional manual testing, affected by operational consistency and fatigue, results in errors exceeding ±2.8 km / h across various operating conditions, with errors reaching over ±3.5 km / h at high speeds and during uphill climbs, failing to meet high-precision testing requirements. Single automated testing lacks dynamic adaptive adjustment and multi-dimensional error compensation, maintaining errors within ±1.3-2.5 km / h, making it difficult to handle complex operating conditions. This invention, through dynamic operating condition adaptation, real-time feedback from digital twins, and multi-dimensional error compensation technology, controls errors across various operating conditions within ±0.5-0.9 km / h, significantly reducing vehicle speed fluctuations. It maintains high-precision tracking even under high load and dynamic transition conditions, fully verifying the effectiveness of multi-device collaborative control and adaptive adjustment strategies, providing a core guarantee for the accuracy of test data.

[0056] refer to Figure 6 This invention clearly demonstrates its groundbreaking improvement in vehicle model adaptation efficiency, addressing the industry pain point of long adaptation cycles in traditional testing methods. Traditional testing methods require manual recalibration of parameters and adjustment of robot movements, resulting in adaptation times of 8.5-14.5 hours, especially for large vehicles and hybrid vehicles, severely impacting testing efficiency. This invention integrates multi-robot collaborative calibration and controller parameter self-learning strategies, using visual positioning for automatic calibration and digital twin model rapid adaptation to shorten adaptation time to 35-55 minutes, significantly reducing manual intervention. Whether it's a compact sedan or a large pickup truck, it can quickly learn vehicle characteristics, improving adaptation efficiency by more than 10 times. This significantly enhances the system's rapid response capability to different vehicle models, meeting the batch testing needs of automakers for multiple models and reducing R&D cycles and testing costs.

[0057] refer to Figure 7 This invention highlights the advantages of its multi-robot integrated control and collaborative calibration technology, solving the problems of low robot execution accuracy and poor coordination in traditional automated testing. In traditional automated testing, the accelerator pedal robot has an error of ±0.8%FS, and the repeatability error of the gear shifting and vehicle-mounted robot exceeds ±0.3mm, resulting in insufficient overall execution accuracy and affecting the reliability of test data. This invention, through visual positioning zero-point calibration, dynamic adaptive adjustment, and multi-device collaborative control, improves the accelerator pedal robot accuracy to ±0.1%FS, and controls the repeatability of the gear shifting and vehicle-mounted robot to within ±0.05mm, significantly optimizing overall execution accuracy and repeatability. The coordinated and consistent movements of the three robots effectively simulate real driving operations, ensuring the accuracy and stability of the testing process, providing reliable operational simulation support for vehicle performance evaluation, and demonstrating the system's technological breakthrough in refined control.

[0058] refer to Figure 8This invention fully demonstrates the efficiency advantages of its fully automated process and multi-device collaborative design, solving the problems of long cycles and high labor intensity in traditional manual testing. Traditional manual testing is hampered by cumbersome procedures, time-consuming data recording, and slow switching between operating conditions. A single test item can take 2.5-8.5 hours, and comprehensive performance testing can take as long as 8.5 hours, resulting in low efficiency. This invention, through automatic analysis of operating conditions, collaborative execution by multiple devices, and automatic collection and fusion analysis of all-dimensional data, reduces the time required for each test item to 1.1-3.8 hours, improving efficiency by 40%-58%. Especially in braking performance testing and acceleration testing, efficiency is improved by over 55% thanks to the robot's rapid response and precise control. The entire process requires no manual intervention, achieving closed-loop automation of "test-analysis-optimization," significantly shortening the testing cycle, meeting the large-scale, high-efficiency testing needs of automakers, and reducing testing labor and time costs.

[0059] refer to Figure 9 This invention directly reflects the advantages of its comprehensive data acquisition and multi-source fusion technology, solving the problems of incomplete and one-sided data acquisition in traditional testing. Traditional manual testing relies on manual recording, covering only about 60% of the comprehensive parameters, with environmental and energy consumption parameters accounting for as low as 62%, failing to meet the needs of comprehensive performance analysis. While single automated testing offers improvements, robot execution data and hub simulation parameters cover less than 75%, lacking multi-source data collaborative support. This invention, through multi-channel acquisition units and a comprehensive acquisition design, integrates multiple parameters such as vehicle operation, robot execution, hub simulation, and environmental energy consumption, increasing the acquisition coverage to 96%-99% and achieving 97% comprehensive parameter coverage across all dimensions. The rich and complete data provides comprehensive support for digital twin simulation comparison, test anomaly identification, and performance index analysis, ensuring the scientific and comprehensive nature of test reports and providing accurate data for vehicle R&D improvement, demonstrating the system's advantages in the integrity and reliability of data acquisition.

[0060] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An automatic testing method for new energy vehicles, characterized in that, include: The test preparation steps are as follows: fix the new energy vehicle on the rotating platform, complete the CAN bus connection between the vehicle and the test system, perform visual positioning and zero-point calibration on the accelerator pedal robot, gear shifting robot and vehicle function triggering robot, build a digital twin model based on vehicle parameters, import the preset test condition curves and complete the initial parameter configuration of the rotating platform. In the working condition analysis step, the host computer decomposes the working condition curve in multiple dimensions, extracts the core parameters of vehicle speed, acceleration, slope and road resistance, and combines the digital twin model to simulate the vehicle dynamic response, generating a refined control logic table containing robot actions and hub loading strategies at each time node. The multi-device collaborative control process involves the lower-level machine receiving control commands based on a custom dual communication protocol, collaboratively controlling the accelerator pedal robot to adjust the opening degree through voltage signals, the gear shifting robot to complete gear switching, the vehicle's functions triggering the robot to execute button and touch screen operations, and the rotating platform simulating road resistance and slope in real time. The dynamic adaptive adjustment steps are based on real-time vehicle operation data and feedback from the digital twin model to calculate the working condition adaptation coefficient and dynamically adjust the robot's execution accuracy and hub loading parameters. The full-dimensional data acquisition process involves simultaneously collecting vehicle speed, motor torque, remaining battery power, voltage, current operating parameters, robot execution accuracy data, and hub simulation parameters through multi-channel acquisition units. The data fusion and analysis step involves using a multi-source data fusion algorithm to process the collected data, combining it with a digital twin model for simulation comparison, identifying test anomalies, and generating a comprehensive test report that includes performance indicators, operating condition tracking errors, and parameter change trends. The test closed-loop optimization steps involve adjusting the operating condition curve parameters and control strategies based on the test report, and re-executing the test process until the preset test accuracy requirements are met.

2. The automatic testing method for new energy vehicles according to claim 1, characterized in that, It also includes a dynamic operating condition adaptation step, which constructs an operating condition adaptation coefficient calculation model based on vehicle operating status and environmental parameters. Real-time optimization of the control strategy is achieved through multi-parameter coupled calculation, expressed as follows: in For dynamic operating condition adaptation coefficient, , , , These are weighted coefficients, and their sum is 1. For real-time vehicle speed, The standard operating speed is [not specified]. For real-time acceleration, For standard acceleration under operating conditions, For real-time road resistance, For standard road resistance, Remaining battery power The battery is fully charged. To simulate the slope of the rotating hub, multi-parameter coupled calculation can integrate vehicle dynamics, energy status and environmental simulation parameters to dynamically adapt to the control requirements of different test conditions.

3. The automatic testing method for new energy vehicles according to claim 1, characterized in that, It also includes a multi-robot collaborative calibration step, initiating the calibration process periodically before and during the test. The host computer issues calibration commands, and the slave computer controls the accelerator pedal robot, the gear shifting robot, and the vehicle's infotainment system to trigger the robots to perform preset calibration actions. The actual position data of the robots is collected through the vision positioning system, compared with the standard position data to calculate the calibration deviation, generate calibration coefficients, and send them to each robot execution unit to correct the position error and response delay of the actuator. The calibration cycle can be dynamically set according to the test duration and the robot's operating status.

4. The automatic testing method for new energy vehicles according to claim 1, characterized in that, It also includes a real-time feedback step for digital twins, which dynamically updates the operating status of the digital twin model based on vehicle physical parameters and test process data, simulates the dynamic response, energy consumption and component wear of the vehicle under the current control parameters, compares the simulation results with the actual vehicle test data in real time, calculates the deviation value, and automatically triggers control parameter adjustment instructions when the deviation value exceeds the preset threshold, correcting the robot's execution actions and hub loading strategy.

5. The automatic testing method for new energy vehicles according to claim 1, characterized in that, It also includes a multi-dimensional error compensation step, which constructs a comprehensive error compensation model to address sensor measurement errors, robot execution errors, and hub simulation errors during the testing process. The control commands are corrected using a nonlinear function, expressed as follows: in To control the amount of instruction compensation, to For error compensation weighting coefficients, For sensor displacement measurement error, To perform attitude angle deviation for the robot, To account for speed measurement error, For test duration, This refers to the actual ambient temperature. For standard test temperature, To compensate for the error in the load applied to the hub, nonlinear compensation is performed by considering multiple sources of error from various dimensions.

6. The automatic testing method for new energy vehicles according to claim 1, characterized in that, It also includes test safety monitoring steps, real-time monitoring of vehicle operating status, battery voltage and temperature, motor operating parameters, and test system communication status, setting multiple safety thresholds, and immediately triggering a safety protection mechanism when parameters are detected to exceed the safety range or communication is interrupted. The host computer issues an emergency stop command, the slave computer controls the robot to reset and the hub to stop loading, and at the same time cuts off the vehicle's power input, records abnormal data and the time of failure, generates a safety warning report and pushes it to the test terminal.

7. The automatic testing method for new energy vehicles according to claim 1, characterized in that, It also includes a test strategy adaptive optimization step. Based on multiple rounds of test data and report results, reinforcement learning algorithms are used to iteratively optimize the working condition analysis logic, multi-device collaborative control parameters, and dynamic adjustment strategies. A mapping relationship between test scenarios and optimal control strategies is established. When new test conditions are introduced or test vehicles are changed, the system can automatically call the optimal strategy and perform rapid adaptation, shorten test preparation time, improve test efficiency and result consistency, and the optimized strategy can be stored in the strategy library for subsequent use.

8. A system for an automatic testing method for new energy vehicles according to any one of claims 1-7, characterized in that, include: The host computer system uses an interactive GUI interface developed based on MATLAB. It has built-in modules for working condition analysis, digital twin modeling, control strategy generation, and test management. It supports the import of various standard working condition curves and the editing of custom working conditions. The digital twin modeling module constructs a virtual model of the vehicle and the test environment. Combined with the working condition analysis module, it generates a refined control logic table. The test management module realizes test task creation, progress monitoring, and report generation. The lower-level system, with the PXIe controller as its core, is equipped with a multi-functional I / O module, a CAN / LIN interface module, a timer / counter module, and a built-in multi-device collaborative control module, a dynamic adjustment module, and a communication management module. It achieves high-speed communication with the upper-level computer and execution units through a custom JSON command protocol and a binary real-time data protocol, and coordinates the precise execution of commands by each execution device. The dynamic adjustment module optimizes control parameters based on real-time data feedback. The multi-robot execution unit includes an accelerator pedal robot, a gear shifting robot, and a vehicle function triggering robot. The accelerator pedal robot controls the opening of the accelerator pedal through voltage signals. The gear shifting robot supports automatic switching between column shifter models. The vehicle function triggering robot integrates a 4-DOF robotic arm and a vision positioning system to achieve automated operation of the vehicle screen and steering wheel buttons. Each robot has a built-in independent calibration module. The hub simulation unit consists of a hub body, a loading system, a slope simulation system, and a data acquisition submodule. It simulates different road resistance, loading mass, and slope conditions. The data acquisition submodule collects hub rotation speed and loading force parameters in real time and feeds them back to the lower-level computer. The data acquisition unit integrates multi-channel sensors, a CAN bus acquisition card, and an environmental monitoring module. It supports the acquisition of vehicle speed, motor torque, remaining battery power, voltage, current operating parameters, robot execution data, and environmental temperature and humidity data. The built-in data preprocessing module performs filtering and noise reduction. The data processing unit adopts an architecture that combines edge computing and cloud storage. Real-time data fusion and preliminary analysis are achieved at the edge, while the cloud performs in-depth mining of multi-source data based on big data algorithms to generate a comprehensive test report that includes performance indicators, error analysis, and optimization suggestions. The report can be exported and historical data can be traced. The safety monitoring unit consists of a status monitoring module, an emergency protection module, and an early warning push module. It monitors vehicle and system operating parameters in real time, sets multi-level safety thresholds, and quickly executes shutdown and reset operations when the safety protection mechanism is triggered. The early warning push module pushes abnormal information through audible and visual alarms and terminal notifications.

9. The automatic testing system for new energy vehicles according to claim 8, characterized in that, It also includes a remote operation and maintenance module, which supports remote status monitoring, parameter configuration and fault diagnosis of the test system via the network. Operation and maintenance personnel can remotely view the test progress, equipment operation status and data acquisition, and issue operating condition adjustment and system calibration instructions. When the system fails, the remote operation and maintenance module automatically uploads fault codes and operation logs, provides fault diagnosis suggestions, and supports multi-user permission management.

10. The automatic testing system for new energy vehicles according to claim 8, characterized in that, It also includes an expansion interface module, which reserves access interfaces for wearable devices, wireless sensor networks, and communication interfaces for third-party testing equipment. It supports access to external devices such as dedicated battery management system testing equipment, motor efficiency testing instruments, and high-precision environmental simulation devices, and expands the functions of battery cycle life testing, motor performance testing, and extreme environment adaptability testing. The interface supports plug-and-play and automatic adaptation, and the system software supports flexible expansion of functional modules to meet the diverse adaptation needs of different testing scenarios.

Images