Diagnostic de Fonctions Réparties
Online diagnostics to improve the safety and maintenance of trucks: detect, locate and automatically identify sensor failures in trucks during operation. The project objective is to propose solutions for future electronic design by including at the same time the means of diagnosis. The increasing complexity of electronics in trucks makes more difficult the identification of the faulty sensors. Fault diagnosis of electronics in garage (offline) exists today. The automatic diagnosis of faults during operation of the truck (online) is still a challenge.

From modelling of breakdowns to integration of diagnostics in embedded electronics: the working method in the project followed the V’s cycle.

A first step involved modelling breakdowns in the electronic sensors of the truck, described in a rigorous language, non verbose (e.g. Simulink, Formal language). In this representation, the electronic functions of the truck are described and the physical medium is abstracted (ECU, bus, sensors).

The second step involved the development and integration of diagnostic features in all. The latter raises an indicator output when the failure is detected on an input function under diagnosis. Another output provides a code indicating the type of failure (e.g. failure of the speed sensor of the third rear wheel).

The third step was the programming of these new features (software) in the electronic controllers of the truck with the constraint to make this integration as transparent as possible in the existing functions.

Number of scientific articles published: 4

Number of patents filed: 0

Number of product’s innovation: 3

[1] a prototype tool to analyze the diagnosability of an electronic system. The demonstration of this prototype is done on the case study (SDK Smart Distance Keeping) simulated. It can conclude on the diagnosis of failure ie, there exists at least one instance of the operating system for which the failure can not be discriminated;

[2] a co-real/simulated online diagnostic demonstrator applied to the case of the SDK. The diagnosers identify the fault using analytical redundancy. Validation tests show that the diagnosers identify systematically failures injected. The diagnosis latency due to CAN communications remains quasi constant;

[3] a prototype tool for analyzing the observability of failures, once the electronic system integrated into the truck. The demonstration of this prototype is done on the SDK implemented in the previous demonstrator.

Number of product’s innovation service: 0

Number of projected jobs created: 0

Number of jobs maintained: 0

Number of related companies creation: 0