As advancements in artificial intelligence (AI) and neural networks continue to reshape industries, the aviation sector is increasingly leveraging these technologies to enhance safety and operational efficiency. Global spending on AI in the aviation market is projected to exceed $6 billion by 2025, underscoring that the future of flight will be defined by technological innovation. Central to this transformation are AI-driven systems that improve equipment reliability, enhance pilot visibility, and mitigate risks in complex flight environments.
Yevgeni Yermolin, a seasoned computer engineer, has dedicated much of his career to advancing AI applications in aviation technology. With extensive experience at Elbit Systems, Yermolin contributed to developing image processing solutions that significantly enhanced pilot situational awareness in low-visibility conditions. Elbit Systems Ltd. is recognized worldwide for its distinguished contributions to defense and aviation technology. The company ranks 28th in the SIPRI Arms Industry Database and 30st in the Defense News Top 100 list, highlighting its commitment to high-quality innovation and strong partnerships across military and civilian sectors. “Working with Elbit’s ClearVision Enhanced Flight Vision System (EFVS) allowed us to combine visible and invisible spectrum imaging, giving pilots a clearer view than the human eye could achieve,” Yermolin explains. This multispectral visual solution, augmented with real-time synthetic data, enables safer navigation through fog, rain, and smoke — conditions that would typically impair visibility.
Yermolin recalls a pivotal moment during development when the team encountered a real-world test scenario: “A fire broke out near our office, and we seized the opportunity to test our algorithms in real smoke conditions. The system filtered out the smoke, which wasn’t part of the original scope but demonstrated the robustness of our technology.” His work at Elbit established a strong foundation in AI applications for aviation, particularly in systems that require high reliability for pilot safety.
Building on his experience with Elbit, Yermolin later co-authored research on energy-efficient neural networks for aviation, which was presented at IJCNN 2020 (International Joint Conference on Neural Networks). This prestigious conference, a gathering point for researchers and industry professionals, promotes collaboration and innovation within the global neural network community. At IJCNN 2020, the team presented research on optimizing Convolutional Neural Networks (CNNs) for resource-efficient applications in safety-critical fields, including aviation. “IJCNN 2020 was pivotal, offering insights into both theoretical and practical applications of neural networks in high-stakes environments like aviation,” Yermolin reflects. He notes that their research achieved models with approximately 75% accuracy, with the primary focus being energy efficiency rather than accuracy. However, he has also worked with models achieving over 95% accuracy, a standard he believes is essential for civil aviation. Achieving such accuracy requires rigorous validation, as it is crucial to prove the correctness of the models and not rely solely on statistical probability. The accuracy must be near perfection to ensure the reliability that safety agencies demand,” he adds.
This balance between innovative technologies and the stringent safety demands of aviation is central to Yermolin’s work. His contributions to image processing algorithms and hardware integration ensure that pilots have access to advanced visual aids in the cockpit. “Every second counts in the cockpit, and the hardware must support real-time decision-making without compromising safety. The collaboration between hardware and software is critical for these complex systems to function seamlessly,” he explains. “While it will be challenging to transition such products entirely to AI, this is the future of aviation. Development will accelerate, but we must find a way to demonstrate the smooth operation of these algorithms.”
Looking ahead, Yermolin anticipates even greater integration of AI in aviation, particularly in developing autonomous systems that could assist pilots during critical flight phases. “We are working toward systems that can anticipate issues before they occur, thanks to AI-driven predictive maintenance and decision support,” he adds. These advancements have the potential to revolutionize flight safety by reducing pilot workload and enhancing the precision of in-flight decision-making.
As AI technologies evolve, Yermolin’s ongoing work in aviation safety systems is shaping the future of flight. His contributions to hardware design and AI integration are paving the way for a new generation of aircraft that are not only safer but also more efficient and reliable than ever before.
