Volumetric Reconstruction of Combustion Dynamics with Multimodal Sensor Fusion

Vadum proposes a state-of-the-art solution for volumetric reconstruction of turbulent combustion dynamics through the integration of multimodal sensor fusion and advanced neural network-based tomographic techniques. By combining high-speed frame-based cameras for high spatial and color resolution with event-based sensors for ultra-high temporal resolution, this system captures the full complexity of combustion processes, including rapid ignition and turbulent flame propagation. The data are processed in real-time using physics-informed convolutional neural networks (CNNs) and deployed on the low-SWaP Nvidia Jetson AGX Orin platform, enabling efficient 3D volumetric reconstruction and seamless user interaction. The innovative aspects of Vadum’s approach include the fusion of asynchronous event-based and synchronous frame-based imaging data streams, homographic alignment for multimodal data fusion, and the use of machine learning algorithms to accelerate and enhance tomographic accuracy. The solution also incorporates a robust software suite with an intuitive GUI for real-time visualization, actionable reporting, and adaptable scalability to meet diverse mission requirements. The proposed system will have transformative impacts on the Army’s ability to optimize propellant design, improve artillery range and lethality, and enhance the safety and efficiency of combustion testing. Beyond military applications, the technology will benefit scientific research, autonomous vehicles, medical imaging, and energy systems. Vadum’s experienced interdisciplinary team, composed of applied scientists, engineers, and military SMEs with proven expertise in imaging systems, neural networks, and real-time data fusion, will ensure the successful development and deployment of this innovative solution.