Sensor Fusion and Virtual Sensor Design for Enhanced Multi-Sensor Data Accuracy in Autonomous Systems

Sensor fusion and virtual sensors play a pivotal role in improving data quality for multi-sensor systems used in real-time applications. This paper explores an advanced sensor fusion approach integrating multiple sensor inputs and virtual sensor designs to optimize accuracy and reliability in autonomous systems. By leveraging statistical and machine-learning-based fusion techniques, the proposed method synthesizes redundant and complementary data sources, forming a robust virtual sensor model. This paper investigates the mathematical underpinnings of sensor fusion, proposes a comprehensive simulation framework, and benchmarks two distinct fusion models for their effectiveness. Simulation results validate the capability of the proposed models in enhancing predictive accuracy and resilience against sensor faults, underscoring the method's potential for autonomous applications.