Browsing by Author "Uzun, Mevlut"

Now showing 1 - 6 of 6
  • Thumbnail Image
    ItemOpen Access
    A certifiable AI-based braking control framework for landing using scientific machine learning
    (IEEE, 2024-09-29) Uzun, Mevlut; Celik, Ugurcan; Guner, Guney; Ozdemir, Orhan; Inalhan, Gokhan
    This paper proposes an AI-based braking control system for aircraft during landing. Utilizing scientific machine learning, we train an agent to apply the most effective braking strategy under various landing conditions. This approach ensures physically consistent outputs by grounding the algorithm in the principles of landing physics. Our results demonstrate that the aircraft can successfully decelerate without skidding across all runway conditions and landing speeds. Additionally, the algorithm maintains performance and safety even when brake performance degradation and initial yaw angles are introduced. This robustness is crucial for the certification of AI in safety-critical systems, as the proposed framework provides a reliable and effective solution.
  • Thumbnail Image
    ItemOpen Access
    AMU-LED Cranfield flight trials for demonstrating the advanced air mobility concept
    (MDPI, 2023-08-31) Altun, Arinc Tutku; Hasanzade, Mehmet; Saldiran, Emre; Guner, Guney; Uzun, Mevlut; Fremond, Rodolphe; Tang, Yiwen; Bhundoo, Prithiviraj; Su, Yu; Xu, Yan; Inalhan, Gokhan; Hardt, Michael W.; Fransoy, Alejandro; Modha, Ajay; Tena, Jose Antonio; Nieto, Cesar; Vilaplana, Miguel; Tojal, Marta; Gordo, Victor; Menendez, Pablo; Gonzalez, Ana
    Advanced Air Mobility (AAM) is a concept that is expected to transform the current air transportation system and provide more flexibility, agility, and accessibility by extending the operations to urban environments. This study focuses on flight test, integration, and analysis considerations for the feasibility of the future AAM concept and showcases the outputs of the Air Mobility Urban-Large Experimental Demonstration (AMU-LED) project demonstrations at Cranfield University. The purpose of the Cranfield demonstrations is to explore the integrated decentralized architecture of the AAM concept with layered airspace structure through various use cases within a co-simulation environment consisting of real and simulated standard-performing vehicle (SPV) and high-performing vehicle (HPV) flights, manned, and general aviation flights. Throughout the real and simulated flights, advanced U-space services are demonstrated and contingency management activities, including emergency operations and landing, are tested within the developed co-simulation environment. Moreover, flight tests are verified and validated through key performance indicator analysis, along with a social acceptance study. Future recommendations on relevant industrial and regulative activities are provided.
  • Thumbnail Image
    ItemOpen Access
    Data-driven synthetic air data estimation system development for a fighter aircraft
    (AIAA, 2023-06-08) Karali, Hasan; Uzun, Mevlut; Yuksek, Burak; Inalhan, Gokhan
    In this paper, we propose an AI-based methodology for estimating angle-of-attack and angle-of-sideslip without the need for traditional vanes and pitot-static systems. Our approach involves developing a custom neural-network model to represent the input-output relationship between air data and measurements from various sensors such as inertial measurement units. To generate the training data required for the neural network, we use a 6-degrees-of-freedom F-16 simulator, which is further modified to simulate more realistic flight data. The training data covers the full flight envelope, allowing the neural network to generate accurate predictions in all feasible flight conditions. Our methodology achieves high-accuracy estimations of angle-of-attack and angle-of-sideslip, with mean absolute errors of 0.534 deg and 0.247 deg, respectively, during the test phase. The results demonstrate the potential of the proposed methodology to accurately estimate important flight parameters without the need for complex and costly instrumentation systems. The proposed methodology could have significant practical applications in the aviation industry, particularly in next-generation aircraft instrumentation and control. Future research could focus on further refining the neural-network model and exploring its application in other aircraft systems to improve safety and reduce costs.
  • Thumbnail Image
    ItemOpen Access
    The development of an advanced air mobility flight testing and simulation infrastructure
    (MDPI, 2023-08-17) Altun, Arinc Tutku; Hasanzade, Mehmet; Saldiran, Emre; Guner, Guney; Uzun, Mevlut; Fremond, Rodolphe; Tang, Yiwen; Bhundoo, Prithiviraj; Su, Yu; Xu, Yan; Inalhan, Gokhan; Hardt, Michael W.; Fransoy, Alejandro; Modha, Ajay; Tena, Jose Antonio; Nieto, Cesar; Vilaplana, Miguel; Tojal, Marta; Gordo, Victor; Mendendez, Pablo; Gonzalez, Ana
    The emerging field of Advanced Air Mobility (AAM) holds great promise for revolutionizing transportation by enabling the efficient, safe, and sustainable movement of people and goods in urban and regional environments. AAM encompasses a wide range of electric vertical take-off and landing (eVTOL) aircraft and infrastructure that support their operations. In this work, we first present a new airspace structure by considering different layers for standard-performing vehicles (SPVs) and high-performing vehicles (HPVs), new AAM services for accommodating such a structure, and a holistic contingency management concept for a safe and efficient traffic environment. We then identify the requirements and development process of a testing and simulation infrastructure for AAM demonstrations, which specifically aim to explore the decentralized architecture of the proposed concept and its use cases. To demonstrate the full capability of AAM, we develop an infrastructure that includes advanced U-space services, real and simulated platforms that are suitable for future AAM use cases such as air cargo delivery and air taxi operations, and a co-simulation environment that allows all of the AAM elements to interact with each other in harmony. The considered infrastructure is envisioned to be used in AAM integration-related efforts, especially those focusing on U-space service deployment over a complex traffic environment and those analyzing the interaction between the operator, the U-space service provider (USSP), and the air traffic controller (ATC).
  • Thumbnail Image
    ItemOpen Access
    Physics guided deep learning for data-driven aircraft fuel consumption modeling
    (MDPI, 2021-02-08) Uzun, Mevlut; Demirezen, Mustafa Umut; Inalhan, Gokhan
    This paper presents a physics-guided deep neural network framework to estimate fuel consumption of an aircraft. The framework aims to improve data-driven models’ consistency in flight regimes that are not covered by data. In particular, we guide the neural network with the equations that represent fuel flow dynamics. In addition to the empirical error, we embed this physical knowledge as several extra loss terms. Results show that our proposed model accomplishes correct predictions on the labeled test set, as well as assuring physical consistency in unseen flight regimes. The results indicate that our model, while being applicable to the aircraft’s complete flight envelope, yields lower fuel consumption error measures compared to the model-based approaches and other supervised learning techniques utilizing the same training data sets. In addition, our deep learning model produces fuel consumption trends similar to the BADA4 aircraft performance model, which is widely utilized in real-world operations, in unseen and untrained flight regimes. In contrast, the other supervised learning techniques fail to produce meaningful results. Overall, the proposed methodology enhances the explainability of data-driven models without deteriorating accuracy.
  • Thumbnail Image
    ItemOpen Access
    Scientific machine learning based pursuit-evasion strategy in unmanned surface vessel defense tactics
    (IEEE, 2024-09-29) Celik, Ugurcan; Uzun, Mevlut; Inalhan, Gokhan; Woods, Mike
    In this work we develop an AI-aided tactics generator for uncrewed surface vessels (USVs) for protection of critical national infrastructure and maritime assets in face of surface vehicle attacks. Our scientific machine learning (SciML) based methodology incorporates physical principles into the learning process, enhancing the model's ability to generalize and perform accurately in scenarios not encountered during training. This innovation addresses a critical gap in existing AI applications for maritime defense: the ability to operate effectively in novel or changing conditions without the need for retraining.