Data processing unit for predictive diagnostics in software-defined vehicle embedding systems

Abstract

Data processing device for predictive diagnostics in software-defined vehicle embedded systems, comprising a data acquisition unit for recording runtime and status data from several embedded subsystems arranged in a vehicle, a correlation evaluation unit for determining temporal and functional relationships between software-related and hardware-related diagnostic indicators, a dependency mapping unit for generating a machine-readable structure of functional dependencies between multiple vehicle embedding systems, a prognostic unit for determining at least one probable diagnostic state for at least one subsystem or system function based on the defined relationships and the machine-readable structure, and a response unit for triggering, selecting or providing at least one technical measure depending on the specific diagnostic condition.

Description

Technical field

[0001] The invention relates to a data processing device for diagnosing technical states in vehicle embedded systems. In particular, the invention relates to a device for the predictive acquisition, processing, and evaluation of state and runtime data from several embedded subsystems distributed within a vehicle. Furthermore, the invention relates to a technical diagnostic architecture that jointly evaluates software-related and hardware-related diagnostic indicators, maps functional dependencies within the vehicle system, and, based on this, detects impending functional deviations or degradation states. State of the art

[0002] Modern vehicles feature a multitude of embedded electronic and software-controlled subsystems. These include, in particular, control units, sensor interfaces, communication modules, actuator units, and software-based function modules, all interconnected via internal vehicle networks. As vehicle functions increasingly shift to software-defined architectures, the complexity of fault detection and system monitoring rises considerably.

[0003] Common diagnostic systems often operate reactively. An error is only detected or stored once a predefined limit is exceeded or a malfunction has already occurred. Such systems are often limited to monitoring individual error codes, sensor signals, or communication failures in isolation. Early detection of gradual malfunctions, which initially manifest only in temporal patterns, load conditions, signal correlations, or gradual deviations, is therefore only possible to a limited extent.

[0004] Another disadvantage of known solutions is that software-related runtime phenomena and hardware-related state changes are often considered separately. This frequently leaves system-wide relationships between computational load, message delay, memory utilization, signal drift, communication errors, and function calls undetected. Likewise, known systems regularly lack a structured representation of functional dependencies between multiple embedded vehicle systems.

[0005] Furthermore, conventional diagnostic systems often only provide a warning or error message after detecting a fault, without deriving technical countermeasures from the identified pattern. Therefore, there is a need for a data processing system that continuously evaluates distributed vehicle states, predicts future degradation or malfunctions, and prepares or triggers technical countermeasures based on this information. Object of the invention

[0006] The invention is based on the objective of providing a data processing device which enables improved predictive diagnostics in software-defined vehicle embedding systems.

[0007] In particular, a system should be created that combines distributed runtime and status data from several embedded subsystems, correlates software-related and hardware-related diagnostic indicators, maps functional dependencies between the subsystems, and determines a likely state of degradation or failure before an actual functional failure occurs.

[0008] Another task is to provide a device that derives or triggers technical response measures from the predicted diagnostic state in order to increase the operational stability of the vehicle system. Summary of the invention

[0009] The task is solved by a data processing unit for predictive diagnostics in software-defined vehicle embedding systems with a data acquisition unit, a correlation evaluation unit, a dependency mapping unit, a forecasting unit and a reaction unit.

[0010] The acquisition unit is designed to record runtime and status data from several embedded subsystems located within the vehicle. This runtime and status data includes, in particular, processing load data, memory state data, communication latencies, signal waveforms, sensor signal deviations, control unit states, function call data, and actuator responses.

[0011] The correlation evaluation unit is designed to determine temporal and functional relationships between software-related and hardware-related diagnostic indicators. This reveals patterns that indicate an impending malfunction, instability, or degradation.

[0012] The dependency mapping unit is designed to generate a machine-readable structure between multiple vehicle embedding systems, which maps functional dependencies, communication relationships, and signal paths.

[0013] The prognostic unit is designed to determine at least one likely diagnostic state for at least one subsystem or system function, based on the correlated diagnostic indicators and the dependency structure.

[0014] The response unit is configured to trigger, select, or provide at least one technical measure, depending on the specific diagnostic state. This measure may include, in particular, prioritized data acquisition, switching to an alternative functional path, activating a redundant channel, restricting non-critical functions, or providing maintenance or control information.

[0015] The combination of these units creates a technical diagnostic architecture that not only detects existing faults, but also predicts future failure or degradation states and links them to technical responses. Detailed description of the invention

[0016] The invention relates to a data processing device for predictive diagnostics in software-defined vehicle embedded systems. The device comprises a data acquisition unit for recording distributed runtime and status data, a correlation evaluation unit for the joint evaluation of software-related and hardware-related diagnostic indicators, a dependency mapping unit for generating a machine-readable structure of functional relationships, a forecasting unit for determining a likely diagnostic state, and a reaction unit for triggering technical measures. This enables the early detection of gradual degradation and impending functional deviations in distributed vehicle embedded systems and the implementation of corresponding technical countermeasures.

[0017] In one embodiment, the data processing device comprises a data acquisition unit configured to continuously or event-driven acquire diagnostic and status data from multiple subsystems distributed throughout the vehicle. The data can originate from control units, sensor interfaces, communication modules, memory areas, software runtime environments, and actuator paths. The data acquisition unit can standardize the received data, assign it to a specific time period, and make it available for further diagnostic processing.

[0018] In a preferred embodiment, the acquisition unit comprises software-related diagnostic indicators and hardware-related diagnostic indicators. Software-related diagnostic indicators include, in particular, runtime deviations, errors in function calls, memory access patterns, task switching behavior, delays in process sequences, or communication stack events. Hardware-related diagnostic indicators include, in particular, sensor signal drift, voltage deviations, temperature states, communication errors, actuator response times, or device-related state changes.

[0019] The correlation analysis unit processes the acquired data and determines temporal, functional, or causal relationships between multiple diagnostic indicators. It preferentially identifies patterns that emerge from the combined development of several parameters. For example, the simultaneous occurrence of increased computational load, extended message latency, and a sensor signal deviation may indicate impending functional instability in a specific subsystem. The combined evaluation of multiple indicators enables a more accurate and earlier diagnosis than an isolated analysis of each individual indicator.

[0020] In another embodiment, the data processing unit includes a dependency mapping unit. This unit is configured to generate a machine-readable structure that describes functional relationships between several embedded subsystems of the vehicle. The structure can include control units, sensor signals, communication buses, software functions, actuator chains, and safety-related function paths. In a preferred embodiment, the structure is designed as a graph model with nodes and edges. This allows the system to identify which other vehicle functions are directly or indirectly affected by a detected deviation.

[0021] Based on the correlated diagnostic indicators and the dependency structure, the forecasting unit determines a likely future diagnostic state. This diagnostic state can include, in particular, an impending failure, a degradation stage, a risk level, a remaining operational probability, or a time window until an expected loss of function. In a preferred embodiment, both current operating data and historical operating data are incorporated to improve forecast accuracy.

[0022] In another embodiment, the forecasting unit takes into account the criticality of the affected vehicle function. This allows predicted deviations to be classified not only according to their probability of occurrence, but also according to their potential impact on vehicle-related operational or safety functions. This enables a prioritized technical response.

[0023] The response unit is designed to trigger, select, or provide at least one technical measure when a predefined diagnostic or prognostic threshold is reached or exceeded. Such a measure could, for example, consist of acquiring additional diagnostic data with increased priority, activating an alternative software path, switching on a redundant function channel, restricting a non-critical function, or transmitting technical maintenance or control information to another vehicle system or an operator output.

[0024] In a preferred embodiment, the acquisition unit, the correlation evaluation unit, the dependency mapping unit, the forecasting unit, and the reaction unit are coupled within a common data processing architecture. This enables continuous feedback between data acquisition, pattern evaluation, system modeling, forecasting, and reaction triggering.

[0025] In another embodiment, the data processing unit can include a learning unit that evaluates historical diagnostic histories, actual fault events, and the effects of previously triggered measures in order to adjust weights or thresholds within the correlation evaluation or forecasting unit. This allows for further improvement of diagnostic performance over the operating lifetime.

[0026] The invention is not limited to the described embodiments. Rather, modifications are possible in which individual diagnostic indicators, dependency models, or reaction types are adapted according to the requirements of the respective vehicle embedding system.

Claims

[1] Data processing device for predictive diagnostics in software-defined vehicle embedded systems, comprising a data acquisition unit for recording runtime and state data from several embedded subsystems arranged in a vehicle, a correlation evaluation unit for determining temporal and functional relationships between software-related and hardware-related diagnostic indicators, a dependency mapping unit for generating a machine-readable structure of functional dependencies between multiple vehicle embedding systems, a prognostic unit for determining at least one probable diagnostic state for at least one subsystem or system function based on the defined relationships and the machine-readable structure, and a response unit for triggering, selecting or providing at least one technical measure depending on the specific diagnostic condition. [2] Data processing equipment according to claim 1, characterized by that the runtime and state data include at least computational load data, memory state data, communication latencies, sensor signal profiles, control unit states, function call data or actuator response data. [3] Data processing equipment according to claim 1, characterized by , that the machine-readable structure is designed as a graph structure to represent control units, sensor signals, communication paths, software functions and actuator chains. [4] Data processing equipment according to claim 1, characterized bythat the expected diagnostic state includes at least one degradation level, one failure risk level, one residual operating probability, or one time window until an expected loss of function. [5] Data processing equipment according to claim 1, characterized by that the response unit, as a technical measure, triggers or provides prioritized data acquisition, switching to an alternative functional path, activating a redundant channel, restricting non-critical functions, or technical maintenance or control information. [6] Data processing equipment according to claim 1, characterized by , that a learning unit is provided which processes historical diagnostic histories, error events or effects of measures for the adjustment of evaluation parameters, weightings or forecast thresholds.