Autonomous System Digital Twin to test Machine Vision

Autonomous systems are playing crucial role in numerous industrial processes. However, while autonomous systems could offer numerous benefits in productivity and efficiency, their interaction with human operators introduces unique challenges. Autonomous vehicles equipped with combinations of sensors have immense potential in industrial environments, yet the integration and optimization of sensors like LIDAR and video cameras with machine vision pose complex challenges. This article highlights the role of simulation in development and testing of the sensor combination for autonomous vehicles to navigate safely in industrial settings, characterized by high levels of dust and noise as well as by presence of human operators. Simulation emerges as a pivotal tool to replicate realistic environments, enabling comprehensive testing of the sensor combination's performance under diverse and challenging scenarios. Sensor fusion, a critical aspect of obstacle detection, receives validation and fine-tuning through repeatable simulations, enhancing the overall system efficiency. By harnessing simulation's capabilities, developers could iteratively optimize sensor combinations, supporting the advancement of autonomous vehicles in industrial environments. The article describes a holistic approach that combines testing in synthetic environment with real-world validation, proposing the way for safer, more efficient, and reliable autonomous systems.