CSChristos Skliros

Christos Skliros

*This is an algorithmically generated author record based on publicly available information.

Christos Skliros has a total of 16 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • NEFELI: A deep-learning detection and tracking pipeline for enhancing autonomy in advanced air mobility

    by Anastasios Arsenos, Evangelos Petrongonas, Orfeas Filippopoulos, Christos Skliros, Dimitrios Kollias, Stefanos Kollias


  • Common Corruptions for Evaluating and Enhancing Robustness in Air-to-Air Visual Object Detection

    by Anastasios Arsenos, Vasileios Karampinis, Evangelos Petrongonas, Christos Skliros, Dimitrios Kollias, Stefanos Kollias, Athanasios Voulodimos

    The main barrier to achieving fully autonomous flights lies in autonomous aircraft navigation. Managing non-cooperative traffic presents the most important challenge in this problem. The most efficient strategy for handling non-cooperative traffic is based on monocular video processing through deep learning models. This letter contributes to the vision-based deep learning aircraft detection and tracking literature by investigating the impact of data corruption arising from environmental and hardware conditions on the effectiveness of these methods. More specifically, we designed 7 types of common corruptions for camera inputs taking into account real-world flight conditions. By applying these corruptions to the Airborne Object Tracking (AOT) dataset we constructed the first robustness benchmark dataset named AOT-C for air-to-air aerial object detection. The corruptions included in this dataset cover a wide range of challenging conditions such as adverse weather and sensor noise. The second main contribution of this letter is to present an extensive experimental evaluation involving 8 diverse object detectors to explore the degradation in the performance under escalating levels of corruptions (domain shifts). Based on the evaluation results, the key observations that emerge are the following: 1) One-stage detectors of the YOLO family demonstrate better robustness, 2) Transformer-based and multi-stage detectors like Faster R-CNN are extremely vulnerable to corruptions, 3) Robustness against corruptions is related to the generalization ability of models. The third main contribution is to present that finetuning on our augmented synthetic data results in improvements in the generalisation ability of the object detector in real-world flight experiments.


  • Aircraft system-level diagnosis with emphasis on maintenance decisions

    by Christos Skliros, Fakhre Ali, Steve King, Ian Jennions

    This paper proposes a diagnostic technique that can predict component degradation for a number of complex systems. It improves and clarifies the capabilities of a previously proposed diagnostic approach, by identifying the degradation severity of the examined components, and uses a 3D Principal Component Analysis approach to provide an explanation for the observed diagnostic accuracy. The diagnostic results are then used, in a systematic way, to influence maintenance decisions. Having been developed for the Auxiliary Power Unit (APU), the flexibility and power of the diagnostic methodology is shown by applying it to a completely new system, the Environmental Control System (ECS). A major conclusion of this work is that the proposed diagnostic approach is able to correctly predict the health state of two aircraft systems, and potentially many more, even in cases where different fault combinations result in similar fault patterns. Based on the engineering simulation approach verified here, a diagnostic methodology suitable from aircraft conception to retirement is proposed.


  • Experimental Investigation and Simulation of a Boeing 747 Auxiliary Power Unit

    by Christos Skliros, Fakhre Ali, Ian K. Jennions

    Auxiliary power units (APUs) are a major driver of maintenance on civil aircraft. However, experimental data and performance simulations are rarely seen in public domain literature. While there is recourse to aircraft engine experience, this does not address the loading and the failure modes of an APU. This work aims to add to the literature, by experimentally investigating a Boeing 747 APU, collecting data under various power settings and ambient conditions, and using these data to calibrate a simple simulation model. This simulation model will subsequently be used to explore failure modes in the APU and hence what sensors may be needed for health monitoring purposes in future work. In this paper, a Boeing 747 APU rig development process and the testing strategy are presented. The rig is validated through a process that includes uncertainty analysis, repeatability tests, consistency tests, and comparison of the collected data with the calibrated simulation model. The results from the rig's validation indicate that the data collected from the APU is independent of its running time or the order of loading cycles imposed on it, i.e., the results are path independent. Changes in pneumatic and electrical power result in small changes in the rotational speed despite the fact that the rotational speed should remain constant. The rotational speed shows a slightly increasing trend when the extracted power rises, and this affects the APU thermodynamic characteristics. This work has resulted in a calibrated simulation model that will be further used in examining fault mode scenarios, as injecting these directly into the rig is seen as high risk.


  • Open Access

    A CASE STUDY OF VIBRATION FAULT DIAGNOSIS APPLIED AT ROLLS-ROYCE T-56 TURBOPROP ENGINE

    by Christos Skliros

    Gas turbine engines include a plethora of rotating modules, and each module consists of numerous components. A component’s mechanical fault can result in excessive engine vibrations. Identification of the root cause of a vibration fault is a significant challenge for both engine manufacturers and operators. This paper presents a case study of vibration fault detection and isolation applied at a Rolls-Royce T-56 turboprop engine. In this paper, the end-to-end fault diagnosis process from starting system faults to the isolation of the engine’s shaft that caused excessive vibrations is described. This work contributes to enhancing the understanding of turboprop engine behaviour under vibration conditions and highlights the merit of combing information from technical logs, maintenance manuals and engineering judgment in successful fault diagnosis.


  • Open Access

    A review of model based and data driven methods targeting hardware systems diagnostics

    by Christos Skliros, Manuel Esperon Miguez, Ali Fakhre, Ian Jennions

    System health diagnosis serves as an underpinning enabler for enhanced safety and optimized maintenance tasks in complex assets. In the past four decades, a wide-range of diagnostic methods have been proposed, focusing either on system or component level. Currently, one of the most quickly...