Development of Sensor Models with Embedded Failure Simulation for Co-Simulation

Introduction:

As complexity and level of automation increases for engineering systems, simulation-based validation emerges as a crucial tool for ensuring system safety and compliance. Particularly in the realm of autonomous vehicles and ships, comprehensive testing spanning from individual software functions to end-to-end system integration is essential. End-to-end testing mandates sophisticated system simulations that accurately represent interactions among components, environmental factors, control systems, software, and operators. In this context, co-simulation, involving the joint simulation of loosely coupled stand-alone sub-simulators, offers significant advantages over traditional modeling approaches. Co-simulation not only facilitates modular model reuse but also enables the establishment of standardized interfaces for seamless integration into system simulators.

Objective:

This specialization/master project aims to develop reusable sensor models with embedded failure simulation capabilities, focusing primarily on GPS and radar models. These sensor models are pivotal for conducting realistic end-to-end testing of autonomous vehicles, providing accurate measurement signals reflective of the simulated environment while also accommodating the simulation of relevant sensor failure modes.

Proposed Approach:

  1. Literature Review: Conduct an extensive review of literature on co-simulation methodologies, sensor modeling techniques, and failure simulation strategies in the context of autonomous systems validation.
  2. Model Development: Develop robust sensor models for GPS and radar systems, ensuring fidelity to real-world behavior and incorporating mechanisms for simulating various failure modes realistically.
  3. Implementation: Implement the developed sensor models as FMUs (functional mock-up units) using the fmiCpp framework.
  4. Validation and Testing: Validate the sensor models through comprehensive testing, assessing their accuracy in generating realistic measurement signals under varying environmental conditions and failure scenarios.
  5. Documentation: Prepare a comprehensive thesis report documenting the models, and summarizing the findings, methodologies, and contributions of the project.

This specialization/master project will be conducted in collaboration with SINTEF Ålesund.