A software and hardware decoupling simulation test system

Abstract

The application discloses a kind of hardware and software decoupling simulation analog test system, belong to vehicle simulation analog test technical field, solve the technical problem that traditional test system is excessively considered from theoretical angle, actual situation in running process is often ignored.System includes: environment simulation module, as through analog simulation technology realizes the integrated tool and platform of automobile virtual research and development, vehicle driving environment model, vehicle-mounted environment sensing model and traffic model are integrated in one, provide three-dimensional digital virtual test scene modeling and editing function;Vehicle dynamics simulation module provides vehicle dynamics simulation, scene construction, sensor construction, data interface construction;Evaluation index and evaluation software module are used to receive the result output by environment simulation module and vehicle dynamics simulation module, and output report.The application focuses on simulating real vehicle working condition, road condition and scene to test designed algorithm model, software and hardware, so as to find problems as soon as possible and solve problems.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle simulation test, more particularly, it relates to a kind of soft and hardware decoupling simulation test system. BACKGROUND

[0002] The existing whole vehicle and spare parts laboratory verification system usually designs the related closed-loop test system from the three levels of MIL (Model in the loop), SIL (software in the loop) and HiL (Hardware-in-the-Loop). Since the functional requirements of MIL / SIL / HIL are considered separately from the initial design, the MIL / SIL / HIL test systems are built independently, which often meets the needs of engineers at different levels (algorithm engineers, software engineers and system engineers) and separates the whole engineering system, making it difficult to combine the test requirements at different levels with the overall engineering requirements and build a comprehensive and systematic test system.

[0003] In the process of building the test system, the traditional method focuses on the test design based on kinematic model and dynamic model from the theoretical point of view, and verifies the performance of the model under different working conditions and road conditions. Excessive consideration from the theoretical point of view often ignores the actual situation during operation. SUMMARY

[0004] The technical problem to be solved by the present application is to solve the above-mentioned deficiencies of the prior art. The purpose of the present application is to provide a soft and hardware decoupling simulation test system, which can highly fit the actual scene, start from the actual scene, verify the performance in different fitting scenes, and optimize the design.

[0005] The technical solution of the present application is: a soft and hardware decoupling simulation test system, comprising:

[0006] An environment simulation module, as an integrated tool and platform for virtual R&D of automobiles through simulation technology, integrates automobile driving environment model, vehicle-mounted environment sensing model and traffic model, can be seamlessly linked with Matlab / Simulink and supports offline and real-time simulation functions, and provides three-dimensional digital virtual test scene modeling and editing functions;

[0007] A vehicle dynamics simulation module for providing vehicle dynamics simulation, scene construction, sensor construction and data interface construction, building control algorithms through the Simulink platform, and realizing dynamics simulation by connecting hardware, software and images, wherein the images are automatically generated by the environment simulation module;

[0008] The evaluation index and evaluation software module are used for receiving the perception result output by the environment simulation module, receiving the planning target trajectory, vehicle speed, control state quantity and control execution quantity output by the vehicle dynamics simulation module, defining the evaluation index and objective function, outputting the model algorithm capability report, and outputting the algorithm module and strategy algorithm parameter to be optimized for the report output.

[0009] As a further improvement, the environment simulation module designs a plurality of scene libraries according to different working conditions and ground conditions, and the operation coefficient K of the scene type can be calculated according to the following linear formula:

[0010] K=α* K1+β* K2+γ*K3+δ*K4;

[0011] Wherein, α is the ground condition consideration ratio, K1 is the ground comprehensive coefficient, β is the weather condition consideration ratio, K2 is the weather comprehensive coefficient, γ is the crop growth condition consideration ratio, K3 is the crop growth condition comprehensive coefficient, δ is the operation scene consideration ratio, and K4 is the operation scene comprehensive coefficient.

[0012] Further, the vehicle dynamics simulation module includes a power unit module, a chassis unit module and an intelligent driving unit module.

[0013] The power unit module is used to simulate the driving process of the power system, and the power output is calculated using the following formula:

[0014] P=2π×T×n / 60;

[0015] Wherein, P is the power, T is the torque, and n is the speed.

[0016] The chassis unit module is used to calibrate the energy loss k generated in the engine power transmission process and calculate the final power output value.

[0017] P1=P*(1-k);

[0018] Wherein, P1 is the output power.

[0019] The intelligent driving unit module uses the PID formula to calculate and generate the control quantity according to the difference error between the target speed and the measured speed, and the PID formula is as follows:

[0020] ;

[0021] Wherein, is the target speed, is the proportional gain, is the integral gain, is the differential gain.

[0022] Further, the torque and speed have the following relationship:

[0023] T = F * L;

[0024] F = M * a;

[0025] v(t) = ∫a(t)dt;

[0026] Where F is force, L is the distance of force action, M is the mass of the vehicle, a is acceleration, and v(t) is velocity.

[0027] Further, a data sharing interface is provided between the power unit module, chassis unit module, and intelligent driving unit module.

[0028] Further, the power system includes a diesel engine system, a range-extending hybrid system, and a pure electric system.

[0029] Further, an adaptive PID adjustment strategy is used to adaptively adjust the KP, KI, and KD of the PID controller to adapt to each road segment. The adaptive PID adjustment strategy includes:

[0030] Data-driven adaptive adjustment: using historical data and machine learning techniques, analyze the performance of different vehicle models in different road conditions, train models to predict the optimal PID parameters;

[0031] Modular PID control: configure independent PID controller modules for each vehicle model and road condition, each module has its adaptive adjustment logic;

[0032] Real-time road condition feedback: collect real-time road condition information through sensors, including road friction coefficient, slope, and curvature, dynamically adjust PID parameters based on road condition information;

[0033] Vehicle dynamics model: establish detailed dynamics models for each vehicle model to simulate their behavior in different road conditions, adjust PID parameters based on model predictions;

[0034] Adaptive control strategy: implement adaptive control algorithms to dynamically adjust PID parameters based on current and predicted errors;

[0035] Fuzzy logic system: use fuzzy logic to handle multiple input variables and nonlinear system characteristics, output adaptive PID parameters;

[0036] Genetic algorithm optimization: apply genetic algorithms to optimize PID parameters, find the best parameters that adapt to the current road segment and vehicle model through iterative processes;

[0037] Hierarchical control structure: design a hierarchical control structure, the lower layer is responsible for real-time PID parameter adjustment, and the upper layer is responsible for selecting control strategies according to vehicle models and road conditions;

[0038] Driver Behavior Simulation: Simulate the driving behavior of different drivers in the simulation system, adjust the PID parameters according to the simulation results to adapt to different driving styles;

[0039] Multi-objective Optimization: Considering the conflicting control objectives of different vehicle models and road conditions, use multi-objective optimization method to balance these objectives and find the most suitable PID parameters;

[0040] Reinforcement Learning: Use reinforcement learning algorithm to let the control system learn from the interaction with the environment, continuously optimize the PID parameters to achieve better control effect;

[0041] System Integration and Coordination: Ensure that all modules can work together and share information to better understand the vehicle state and environmental conditions, and make more effective PID parameter adjustments;

[0042] Real-time Monitoring and Adjustment: Implement real-time monitoring system to continuously evaluate control performance, and adjust PID parameters immediately once performance degradation or specific trigger conditions are detected;

[0043] User Customization: Allow users to customize PID parameter adjustment strategies based on experience or specific needs, and the system will adjust accordingly based on user input.

[0044] Advantages

[0045] Compared with the prior art, the present application has the following advantages:

[0046] The present application focuses on simulating real vehicle working conditions, road conditions and scenarios to ensure data flow transmission between modules. According to the needs of various test environments in the development process, relevant models are designed in the loop, software in the loop and hardware in the loop to ensure that the test environment is highly similar to the real vehicle situation, so as to test the designed algorithm model, software and hardware, thereby discovering and solving problems as early as possible. The present application starts from the working scene and has the following characteristics: coupling with actual scene data, high data sharing between subsystems, decoupling of software and hardware, modularization of subsystems, and flexibility in building. BRIEF DESCRIPTION OF DRAWINGS

[0047] Fig. 1 is the framework diagram of the present application;

[0048] Fig. 2 is the PID controller framework diagram. DETAILED DESCRIPTION

[0049] The present application will be further described below in conjunction with specific embodiments in the drawings.

[0050] Reference Figs. 1-2The application discloses a hardware and software decoupling simulation test system, which comprises an environment simulation module, a vehicle dynamics simulation module, an evaluation index and evaluation software module.

[0051] The environment simulation module is an integrated tool and platform for virtual R&D of the automobile by means of simulation technology, integrates an automobile driving environment model, a vehicle-mounted environment sensing model and a traffic model, can be seamlessly linked with Matlab / Simulink, supports offline and real-time simulation functions, and provides three-dimensional digital virtual test scene modeling and editing functions.

[0052] The environment simulation module supports modeling and editing of the automobile driving environment of roads and road textures, lane lines, traffic signs and facilities, weather and night scenes. The model establishment and setting of various types of sensors can be supported, including fisheye cameras, monocular cameras, binocular cameras, millimeter wave radars, laser radars, ultrasonic radars, V2X communication sensors, lane line sensors and target object identification sensors. The road model can be set and edited, the road database can be set with slope, curvature, roll and viaduct, the parameterized road surface model is provided, the automatic test is supported, the non-paved road simulation is supported, the third-party map OSM / OpenDRIVE high-precision map import is supported. The environment model can be built, including the road surface and roadside facility database, the traffic sign database, the building and green belt database, and the user-defined 3D model import is supported. The creation and editing of road users are supported, including cars and motorcycles, commercial vehicles, pedestrians and bicycles and test balloon vehicles. The weather conditions such as weather, light and the like are supported, including the creation of day, night, rain, snow and fog weather, the setting of car lights and street lamps. The light model has the sensor response characteristics. The uneven road (such as the deceleration belt and the pothole) can be simulated, and the real road environment can be directly imported from the HERE map. The driver model can define the driving style (aggressive type and stable type), can support complex driving operations, has the learning function and can adapt to different vehicles and road characteristics. The traffic flow model can define almost infinite number of traffic participants, each participant can add a dynamics model, and the participants can be finely controlled according to events.

[0053] The vehicle dynamics simulation module supports modeling and simulation analysis of large, medium and small cars with various typical driving modes and suspension forms, is used for providing vehicle dynamics simulation, scene construction, sensor construction and data interface construction, constructing a control algorithm through a Simulink platform, and realizing dynamics simulation by connecting hardware, software and images. The image is automatically generated through the environment simulation module. The hardware connection I / O involves the communication between the single-chip microcomputer of the embedded system hardware and the PC system, CAN signal reading, and real-time feedback signals to the roll, pitch and yaw control systems on the steering wheel and the base.

[0054] The evaluation index and evaluation software module are used to receive the perception result output by the environment simulation module, and receive the planning target trajectory, vehicle speed, control state quantity and control execution quantity output by the vehicle dynamics simulation module, define the evaluation index and objective function, output the model algorithm capability report, and output the algorithm module and strategy algorithm parameter to be optimized. The output capability report includes perception capability, planning capability, control capability, communication capability and intelligent driving comprehensive capability evaluation. According to the model selection and evaluation index and automatic test software and test case, the optimization variable selection, optimization variable and constraint condition definition are carried out, and finally the algorithm optimization of the measured object is carried out.

[0055] The environment simulation module designs a plurality of scene libraries according to different working conditions and ground conditions, mainly considering the ground conditions (the hardness, flatness and slope of the ground, the ground friction and adhesion), the weather conditions (rain and snow conditions), the crop growth conditions and the operation scene (harvesting scene, seeding scene, field management scene and plowing scene). The operation coefficient K of the scene type can be calculated according to the following linear formula:

[0056] K=α* K1+β* K2+γ*K3+δ*K4;

[0057] Wherein, α is the proportion of ground condition, K1 is the ground comprehensive coefficient, β is the proportion of weather condition, K2 is the weather comprehensive coefficient, γ is the proportion of crop growth condition, K3 is the crop growth condition comprehensive coefficient, δ is the proportion of operation scene, K4 is the operation scene comprehensive coefficient. According to the above conditions, the scene coefficient corresponding table is made, and the table format is shown in Table 1 as follows:

[0058]

[0059] Table 1

[0060] The vehicle dynamics simulation module includes a power unit module, a chassis unit module and an intelligent driving unit module, and the software and hardware are decoupled as follows:

[0061] The power unit module is used to simulate the driving process of the power system, and the power system includes a diesel engine system, a range extender hybrid system and a pure electric system. The power output is calculated by the following formula:

[0062] P=2π×T×n / 60;

[0063] Wherein, P is the power, T is the torque, and n is the speed.

[0064] The chassis unit module is used to calibrate the energy loss k generated in the process of engine power transmission, and calculate the final power output value. According to the above power output, the relationship between power output and vehicle longitudinal acceleration and longitudinal speed is calibrated by imitating the actual working condition, and the difference between diesel engine system, hybrid and pure electric system is distinguished by coefficient.

[0065] P1=P*(1-k);

[0066] Wherein, P1 is the output power.

[0067] The intelligent driving unit module uses the PID formula to calculate and generate the control amount according to the difference error between the target speed and the measured speed, and the PID formula is as follows:

[0068] ;

[0069] Wherein, is the target speed, is the proportional gain, is the integral gain, is the differential gain.

[0070] The relationship between torque and speed is as follows:

[0071] T=F*L;

[0072] F=M*a;

[0073] ;

[0074] Wherein, F is force, L is force action distance, M is vehicle mass, a is acceleration, and v(t) is speed. Therefore, the above three modules of power unit module, chassis unit module and intelligent driving unit module have certain correlation through torque.

[0075] Data sharing interfaces are provided between the power unit module, the chassis unit module and the intelligent driving unit module to realize data sharing between the power unit module, the chassis unit module and the intelligent driving unit module, and ensure that the required data is transmitted between different modules or systems. For example, the acceleration output calculated according to the engine output power will be used in the actual parameter simulation of the intelligent driving system, especially in the PID control, to calibrate the difference between the measured speed and the target speed.

[0076] In the embodiment, the adaptive PID adjustment strategy is adopted to adaptively adjust KP, KI and KD of the PID controller to adapt to each road section. The adaptive PID adjustment strategy includes:

[0077] 1. Data-driven adaptive tuning: Utilize historical data and machine learning techniques to analyze the performance of different vehicle models on various road conditions, training models to predict optimal PID parameters.

[0078] 2. Modular PID control: Configure independent PID controller modules for each vehicle model and road condition, each with its own adaptive tuning logic.

[0079] 3. Real-time road condition feedback: Collect real-time road condition information through sensors, including road friction coefficient, slope, and curvature, and dynamically adjust PID parameters based on this information.

[0080] 4. Vehicle dynamics modeling: Establish detailed dynamics models for each vehicle model to simulate their behavior on different road conditions, and adjust PID parameters based on model predictions.

[0081] 5. Adaptive control strategy: Implement adaptive control algorithms (such as Model Predictive Control (MPC)) to dynamically adjust PID parameters based on current and predicted errors.

[0082] 6. Fuzzy logic system: Use fuzzy logic to handle multiple input variables (such as vehicle speed, steering angle, acceleration, etc.) and non-linear system characteristics, outputting adaptive PID parameters.

[0083] 7. Genetic algorithm optimization: Apply genetic algorithms to optimize PID parameters, finding the best parameters for the current road segment and vehicle model through an iterative process.

[0084] 8. Hierarchical control structure: Design a hierarchical control structure, with the lower layer responsible for real-time PID parameter adjustment and the upper layer responsible for selecting control strategies based on vehicle model and road conditions.

[0085] 9. Driver behavior simulation: Simulate the driving behavior of different drivers in a simulation system, and adjust PID parameters to accommodate different driving styles based on simulation results.

[0086] 10. Multi-objective optimization: Consider conflicting control objectives for different vehicle models and road conditions, and use multi-objective optimization methods to balance these objectives, finding the most suitable PID parameters.

[0087] 11. Reinforcement learning: Use reinforcement learning algorithms to allow the control system to learn from interactions with the environment, continuously optimizing PID parameters for better control performance.

[0088] 12. System integration and coordination: Ensure that all modules (such as perception, decision-making, and control) work together, sharing information to better understand vehicle state and environmental conditions, resulting in more effective PID parameter adjustments.

[0089] 13. Real-time monitoring and adjustment: Implement a real-time monitoring system to continuously evaluate control performance and adjust PID parameters immediately upon detecting performance degradation or specific trigger conditions.

[0090] 14. User-defined settings: Allow users to customize PID parameter adjustment strategies based on experience or specific needs, with the system making corresponding adjustments based on user input.

[0091] The system highly fits the actual scene, starting from the actual scene, and optimizing the design by verifying the performance in different fitting scenarios. In terms of data application, some test modes focus more on the independent application of data. Since the system focuses on the virtual integration and data integration between different subsystems (power unit module, chassis unit module, and intelligent driving unit module), data is coupled in different subsystems. The software and hardware of each system are decoupled, data is shared, and subsystems can be flexibly arranged to serve different testing purposes.

[0092] The above is only the preferred embodiment of the present application. It should be noted that for those skilled in the art, without departing from the structure of the present application, a number of modifications and improvements can be made, which will not affect the effect of the implementation of the present application and the practicality of the patent.

Claims

1. A hardware- software decoupled simulation test system, comprising: The application relates to a vehicle simulation system, which comprises the following modules: An environment simulation module, which is an integrated tool and platform for virtual research and development of a vehicle through simulation technology, integrates a vehicle driving environment model, a vehicle-mounted environment sensing model and a traffic model, can be seamlessly linked with Matlab / Simulink, supports offline and real-time simulation functions, and provides three-dimensional digital virtual test scene modeling and editing functions; A vehicle dynamics simulation module, which is used for providing vehicle dynamics simulation, scene construction, sensor construction, data interface construction, constructing a control algorithm through a Simulink platform, and realizing dynamics simulation by connecting hardware, software and images, wherein the images are automatically generated by the environment simulation module; An evaluation index and evaluation software module, which is used for receiving perception results output by the environment simulation module, receiving planning target trajectories, vehicle speeds, control state quantities and control execution quantities output by the vehicle dynamics simulation module, defining evaluation indexes and target functions, outputting a model algorithm capability report, and outputting an algorithm module and strategy algorithm parameters which need to be optimized according to the report; The environment simulation module designs a plurality of scene libraries according to different working conditions and ground conditions, and the operation coefficient K of the scene type is calculated according to the following linear formula: K = alpha * K1 + beta * K2 + gamma * K3 + delta * K4; Wherein, alpha is the proportion of ground condition consideration, K1 is the ground comprehensive coefficient, beta is the proportion of weather condition consideration, K2 is the weather comprehensive coefficient, gamma is the proportion of crop growth condition consideration, K3 is the crop growth condition comprehensive coefficient, delta is the proportion of operation scene consideration, and K4 is the operation scene comprehensive coefficient; The vehicle dynamics simulation module comprises a power unit module, a chassis unit module and an intelligent driving unit module; The power unit module is used for simulating a driving process of a power system, and power output is calculated according to the following formula: P = 2pi * T * n / 60; Wherein, P is power, T is torque, and n is rotating speed; The chassis unit module is used for calibrating energy loss k generated in the power transmission process of an engine and calculating a final power output value; P1 = P * (1-k); Wherein, P1 is output power; The intelligent driving unit module calculates and generates a control quantity according to the difference error between a target speed and a measured speed by using a PID formula, and the PID formula is as follows: ； wherein, is a target speed, is a proportional gain, is an integral gain, is a derivative gain; A data sharing interface is arranged between the power unit module, the chassis unit module and the intelligent driving unit module.

2. The hardware- software decoupled simulation test system of claim 1, wherein, There is the following relationship between torque and speed: T = F * L; F = M * a; v(t) = integral(a(t)dt); Wherein, F is force, L is force action distance, M is vehicle mass, a is acceleration, and v(t) is speed.

3. The hardware- software decoupled simulation test system of claim 1, wherein, The power system comprises a diesel engine system, an extended-range hybrid system and a pure electric system.

4. The hardware- software decoupled simulation test system of any one of claims 1-3, wherein, An adaptive PID adjustment strategy is used to adaptively adjust KP, KI and KD of a PID controller to adapt to various road sections, and the adaptive PID adjustment strategy comprises the following steps: Data-driven adaptive adjustment: historical data and machine learning technology are used to analyze the performance of different vehicle models under different road conditions, and a model is trained to predict optimal PID parameters. Modular PID Control: Configure independent PID controller modules for each vehicle model and road condition, each module has its adaptive adjustment logic. Real-time Road Condition Feedback: Collect road condition information in real-time through sensors, including road friction coefficient, slope, and curvature, dynamically adjust PID parameters based on road condition information. Vehicle Dynamics Model: Establish detailed dynamics models for each vehicle model, simulate the behavior of each vehicle in different road conditions, adjust PID parameters based on model prediction. Adaptive Control Strategy: Implement adaptive control algorithms to dynamically adjust PID parameters based on current and predicted errors. Fuzzy Logic System: Use fuzzy logic to handle multiple input variables and nonlinear system characteristics, output adaptive PID parameters. Genetic Algorithm Optimization: Apply genetic algorithms to optimize PID parameters, find the best parameters that adapt to the current road segment and vehicle model through an iterative process. Hierarchical Control Structure: Design a hierarchical control structure, the bottom layer is responsible for real-time PID parameter adjustment, the upper layer is responsible for selecting control strategies according to vehicle models and road conditions. Driver Behavior Simulation: Simulate the driving behavior of different drivers in the simulation system, adjust PID parameters to adapt to different driving styles according to simulation results. Multi-objective Optimization: Consider the conflicting control objectives of different vehicle models and road conditions, use multi-objective optimization methods to balance these objectives, find the most suitable PID parameters. Reinforcement Learning: Use reinforcement learning algorithms to let the control system learn from interaction with the environment, continuously optimize PID parameters to achieve better control effect. System Integration and Coordination: Ensure that all modules can work together and share information to better understand the vehicle state and environmental conditions, and make more effective PID parameter adjustments. Real-time Monitoring and Adjustment: Implement a real-time monitoring system to continuously evaluate control performance, adjust PID parameters immediately once performance degradation or specific trigger conditions are detected. User Customization Settings: Allow users to customize PID parameter adjustment strategies based on experience or specific needs, the system adjusts accordingly based on user input.

Images