CN-121979174-A - Automobile intelligent chassis fault prediction self-healing method, device, equipment and storage medium

CN121979174ACN 121979174 ACN121979174 ACN 121979174ACN-121979174-A

Abstract

The invention discloses a self-healing method, device, equipment and storage medium for automobile intelligent chassis fault prediction, which are characterized in that bus network data, controller internal variables, physical sensor data and environment perception data of a current new energy automobile are synchronously collected in real time, a space-time feature matrix is constructed through space-time alignment and feature level fusion, fault prediction and root cause positioning are carried out based on the space-time feature matrix, fault types, fault occurrence probability and fault root cause positions are determined, corresponding self-healing decision instructions are generated, the chassis multi-subsystem is coordinated and controlled through a multi-objective optimization algorithm according to the self-healing decision instructions, the self-healing strategy is executed, comprehensive perception and accurate characterization of chassis system states can be achieved, early fault feature identification capability is remarkably improved, safety, reliability and user experience of the new energy automobile intelligent chassis system can be greatly improved, and automobile intelligent chassis fault prediction self-healing speed and efficiency are improved.

Inventors

PAN ZIAN
WANG ZHAODONG
Liu lv

Assignees

东风汽车集团股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260108

Claims (10)

1. The automobile intelligent chassis fault prediction self-healing method is characterized by comprising the following steps of: Synchronously acquiring bus network data, controller internal variables, physical sensor data and environment sensing data of a current new energy vehicle in real time, and constructing a space-time feature matrix through space-time alignment and feature level fusion; performing fault prediction and root cause positioning based on the space-time feature matrix, determining fault types, fault occurrence probability and fault root cause positions, and generating corresponding self-healing decision instructions; And according to the self-healing decision instruction, the chassis multi-subsystem is coordinated and controlled through a multi-objective optimization algorithm to execute a self-healing strategy.
2. The self-healing method for predicting failure of an intelligent chassis of an automobile according to claim 1, wherein the real-time synchronous acquisition of bus network data, controller internal variables, physical sensor data and environment sensing data of a current new energy vehicle and construction of a space-time feature matrix through space-time alignment and feature level fusion comprises: The method comprises the steps of collecting bus network data of a current new energy vehicle in real time through a vehicle-mounted communication bus protocol, synchronously collecting internal variables of a controller through a diagnosis interface or an internal memory of an electronic control unit ECU, synchronously collecting physical sensor data through a measuring device, and synchronously collecting environment sensing data through external sensing equipment; And carrying out space-time alignment and feature level fusion on the bus network data, the controller internal variables, the physical sensor data and the environment sensing data, and constructing and generating a space-time feature matrix.
3. The method for predicting and self-healing an automobile intelligent chassis fault according to claim 2, wherein the performing space-time alignment and feature level fusion on the bus network data, the controller internal variables, the physical sensor data and the environment sensing data to construct and generate a space-time feature matrix comprises: Performing time synchronization on time deviation of the bus network data, the controller internal variables, the physical sensor data and the environment sensing data by adopting an IEEE 1588 precise network time protocol PTP; Carrying out space alignment on each data source based on a vehicle kinematic model and a coordinate transformation algorithm, and uniformly mapping sensor data distributed at different positions of a vehicle into a global coordinate system taking a vehicle mass center as an origin; Extracting signal periodic characteristics and abnormal fluctuation indexes from the bus network data, carrying out state space reconstruction and outlier detection on the internal variables of the controller, applying wavelet packet decomposition to extract energy entropy characteristics from the physical sensor data, calculating time-frequency domain statistics, and carrying out target recognition and scene semantic segmentation on environment perception data to extract road curvature and obstacle position characteristics; and organizing the processed bus network data, the controller internal variables, the physical sensor data and the environment perception data according to a unified time sequence to construct a space-time feature matrix with N multiplied by M dimensions, wherein N represents the time sequence length and M represents the feature dimension after fusion.
4. The method for predicting and self-healing an automobile intelligent chassis fault according to claim 1, wherein the performing fault prediction and root cause positioning based on the space-time feature matrix, determining a fault type, a fault occurrence probability and a fault root cause position, and generating a corresponding self-healing decision instruction comprises: When fault prediction and root cause positioning are carried out based on the space-time feature matrix, an LSTM-transducer mixed neural network architecture is adopted to process and carry out multi-level feature extraction time sequence features on the input space-time feature matrix, fault probability distribution and fault type prediction of each key component in a preset time period in the future are output, wherein an LSTM layer in the LSTM-transducer mixed neural network architecture captures short-term dynamic features, and a transducer layer captures long-term dependence through a self-attention mechanism; Meanwhile, a system dependency relationship model based on a graph neural network is constructed, and the fault root cause position is determined according to the system dependency relationship model; and generating a self-healing decision instruction containing an optimal executor combination scheme, an execution priority and a safety constraint condition according to the fault type, the fault occurrence probability and the fault root cause position, a preset fault strategy mapping rule and a preset real-time optimization algorithm.
5. The method for predicting and self-healing an automobile intelligent chassis fault according to claim 4, wherein the simultaneously constructing a system dependency model based on a graph neural network, determining a fault root cause position according to the system dependency model comprises: Meanwhile, a system dependency relation model based on a graph neural network is constructed, all subsystems of a chassis of the current new energy vehicle are abstracted into graph nodes, signal interaction among all subsystems is abstracted into edges, the propagation path of an abnormal mode in a feature matrix in a system graph is analyzed through an abnormal propagation algorithm, and the position of a fault root cause is determined by combining Bayesian reasoning.
6. The method for predicting and self-healing an intelligent chassis fault of an automobile according to claim 1, wherein the step of coordinating and controlling the chassis multiple subsystems to execute the self-healing strategy through a multi-objective optimization algorithm according to the self-healing decision instruction comprises the following steps: Analyzing the current fault type information contained in the self-healing decision instruction, and determining a chassis subsystem combination to be activated and a corresponding safety boundary according to the current fault type information; Establishing a multi-actuator cooperative control model comprising a steering system, a braking system, a suspension system and a driving system of the current new energy vehicle according to the chassis subsystem combination and the safety boundary, and determining a multi-objective optimization function according to the following formula: Wherein, the For the purpose of optimizing the function for multiple objectives, 、、 And To dynamically adjust the weighting coefficients according to the severity of the fault and the operating conditions of the vehicle, In order to track the error in the tracking, In order to be able to consume energy, Is comfortable; By applying a safety constraint condition in the process of solving the multi-objective optimization function, the current new energy vehicle is ensured to be always in a stable domain; distributing the optimal control quantity obtained by the optimization solution to a corresponding executor to generate a control instruction sequence of each subsystem; And controlling the control chassis multi-subsystem to execute a self-healing strategy according to the control instruction sequence.
7. The method for predicting and self-healing an intelligent chassis fault of an automobile according to claim 6, wherein the controlling the control chassis multi-subsystem to execute the self-healing strategy according to the control instruction sequence comprises: When the failure of the steering torque sensor of the current new energy vehicle is detected, automatically switching to a redundant sensor channel according to a steering self-healing instruction in the control instruction sequence, and generating auxiliary steering torque through differential braking within a preset steering response time; When the defect of insufficient brake fluid pressure of the current new energy vehicle is detected, a motor is driven according to a brake fluid pressure self-healing instruction in the control instruction sequence to provide regenerative braking force compensation within a preset motor response time, and meanwhile, the damping force of a shock absorber is regulated in real time through a suspension continuous damping control system CDC, so that the pitch angle change rate of a vehicle body is controlled to meet a preset change rate threshold; Under the condition that the congestion of the communication bus of the current new energy vehicle is detected, the transmission priority of the non-key signal is dynamically degraded according to the communication self-healing instruction in the control instruction sequence, and the transmission delay of the steering or braking key instruction is ensured.
8. The utility model provides an automobile intelligent chassis trouble prediction self-healing device which characterized in that, automobile intelligent chassis trouble prediction self-healing device includes: the space-time feature matrix construction module is used for synchronously acquiring bus network data, controller internal variables, physical sensor data and environment perception data of the current new energy vehicle in real time and constructing a space-time feature matrix through space-time alignment and feature level fusion; the self-healing decision instruction generation module is used for carrying out fault prediction and root cause positioning based on the space-time feature matrix, determining fault types, fault occurrence probability and fault root cause positions and generating corresponding self-healing decision instructions; And the strategy execution module is used for coordinating and controlling the chassis multi-subsystem to execute the self-healing strategy through a multi-objective optimization algorithm according to the self-healing decision instruction.
9. An automobile intelligent chassis fault prediction self-healing device, characterized in that the automobile intelligent chassis fault prediction self-healing device comprises a memory, a processor and an automobile intelligent chassis fault prediction self-healing program which is stored on the memory and can run on the processor, wherein the automobile intelligent chassis fault prediction self-healing program is configured to realize the steps of the automobile intelligent chassis fault prediction self-healing method according to any one of claims 1 to 7.
10. A storage medium, wherein a vehicle intelligent chassis fault-prediction self-healing program is stored on the storage medium, and the vehicle intelligent chassis fault-prediction self-healing program, when executed by a processor, implements the steps of the vehicle intelligent chassis fault-prediction self-healing method according to any one of claims 1 to 7.

Description

Automobile intelligent chassis fault prediction self-healing method, device, equipment and storage medium Technical Field The invention relates to the technical field of intelligent chassis control of new energy automobiles, in particular to an intelligent chassis fault prediction self-healing method, device and equipment for an automobile and a storage medium. Background With the rapid development of new energy automobile technology, the safety and reliability of an intelligent chassis system are increasingly concerned, and at present, the main vehicle fault diagnosis and treatment technology in the market mainly comprises the following modes: The threshold alarm system is the most widely applied technical scheme at present, the working principle is that a fault code is triggered based On a preset fixed threshold (such as temperature overrun, current abnormality and the like) and is reported by a vehicle-mounted diagnosis system (On-Board Diagnostics, OBD) system, however, the system can only give an alarm afterwards after the fault occurs, lacks fault prediction capability, cannot realize active repair, and causes the potential safety hazard of a vehicle to be eliminated in advance. Part of high-end vehicle models adopt a single system fault-tolerant control technology, for example, a braking system adopts a double-electronic control unit (Electronic Control Unit, ECU) redundancy design, and the technology is limited to local hardware backup, and when a cross-subsystem coupling fault (such as electronic stability program (Electronic Stability Program, ESP) intervention delay caused by steering system failure) occurs, driving safety cannot be effectively ensured due to lack of a cooperative mechanism among systems. In addition, with the development of the internet of vehicles technology, cloud fault diagnosis technology is gradually rising, vehicle data is uploaded to the cloud for analysis and processing through a 4G/5G network, but due to inherent delay (usually more than 500 ms) of network transmission and cloud computing, the technology is difficult to meet the requirement of millisecond-level response of a control instruction under the high-speed driving working condition of a vehicle, and real-time self-healing control cannot be supported. The prior art has the following defects: firstly, the effective identification capability of early fault characteristics is lacking, and prediction cannot be performed before the fault occurs; secondly, bus data, sensor data and internal variables of the controller are often analyzed independently, and dynamic coupling relations among subsystems of the chassis are ignored; Furthermore, the fault handling strategy is limited to single-point degradation, and the chassis global actuator resources cannot be fully utilized for collaborative compensation; finally, the response delay of the prior art is larger, and the real-time requirement of the modern intelligent electric automobile on safety control is difficult to meet. Disclosure of Invention The invention mainly aims to provide a self-healing method, device, equipment and storage medium for predicting faults of an intelligent chassis of an automobile, and aims to solve the technical problems that in the prior art, fault prediction is inaccurate, cross-subsystem faults lack of collaborative self-healing capability and millisecond-level real-time control is difficult to realize due to multi-source data fracture analysis. In a first aspect, the present invention provides an automobile intelligent chassis fault prediction self-healing method, which includes the following steps: Synchronously acquiring bus network data, controller internal variables, physical sensor data and environment sensing data of a current new energy vehicle in real time, and constructing a space-time feature matrix through space-time alignment and feature level fusion; performing fault prediction and root cause positioning based on the space-time feature matrix, determining fault types, fault occurrence probability and fault root cause positions, and generating corresponding self-healing decision instructions; And according to the self-healing decision instruction, the chassis multi-subsystem is coordinated and controlled through a multi-objective optimization algorithm to execute a self-healing strategy. Optionally, the real-time synchronous acquisition of bus network data, internal variables of a controller, physical sensor data and environmental perception data of the current new energy vehicle, and the construction of a space-time feature matrix through space-time alignment and feature level fusion comprises the following steps: The method comprises the steps of collecting bus network data of a current new energy vehicle in real time through a vehicle-mounted communication bus protocol, synchronously collecting internal variables of a controller through a diagnosis interface or an internal memory of an electronic control unit ECU, synchronous