Diyar Altinses, M.Sc.

September 2019 - April 2021
Double Degree: Master of Science (M.Sc.)
South Westphalia University of Applied Sciences, Soest (Germany)
University of Bolton, Bolton (UK)
Course: Systems Engineering and Engineering Management

since November 2021
Degree: Doctor of natural sciences (Dr. rer. nat.)
Graduate School for Applied Research, North Rhine-Westphalia (Germany)
South Westphalia University of Applied Sciences, Soest (Germany)
Title: Advancing Multimodal Fault-tolerance: Leveraging Intermodal and Intramodal Correlations using ML Algorithms

Optimization of industrial processes has become increasingly important since unplanned downtime in large industries is associated with increasingly higher costs. These costs come from production lines with higher capacity and higher spending on energy, labor, and material wastage. Predictive maintenance plays a critical role in reducing costs and increasing productivity by predicting wear and aging processes. Nevertheless, predictive maintenance can only detect predictable wear and aging processes. However, failures can also occur randomly and usually lead to unpredictable downtime. A promising and innovative approach to address failures and ensure robust operation is to apply fault-tolerant machine learning techniques to reconstruct corrupt sensor data. This Fault-tolerant systems comprise three integral components: fault injection, fault detection, and fault correction. In this research, we mainly focus on multimodal real-world failure injection methods and self-supervised, as well as generative corretion approaches. We do not specifically address fault detection, as the correction mechanism is consistently applied, rendering detection an inherent and integrated aspect of the correction process.

SeGuForm: The goal of the project is to develop a self-optimizing system for scrap reduction and customer-oriented packaging in stamping and forming technology. This will be achieved through flexible post-processing and sensor-assisted handling systems. The main objective is to reduce the scrapping of defective parts in the production of formed metal components.

SIDDA: The SIDDA project researches automated, intermodal drone networks for climate-neutral goods distribution in suburban areas. By integrating road and air traffic, airspace will be utilized for logistics solutions. Both static and mobile micro-hubs will be configured into a drone airline and integrated into tour and route planning. A U-space service provider will be prototypically developed to simulate the drone airline’s behavior within U-space.

Drones4Parcel5G: Development of a system for planning and movement control of delivery flights using parcel drones based on 5G wireless technology.

PSAR: Development of a learning, AI-based, markerless AR assistance system for various industrial operating tasks involving complex systems, featuring automated AR content creation (time savings > 90%).

2025

Altinses, D., Torres, D. O. S., Lier, S., & Schwung, A. (2025, February). Neural Data Fusion Enhanced PD Control for Precision Drone Landing in Synthetic Environments. In 2025 IEEE International Conference on Mechatronics (ICM) (pp. 1-7). IEEE.

Torres, D. O. S., Altinses, D., & Schwung, A. (2025, March). Data Imputation Techniques Using the Bag of Functions: Addressing Variable Input Lengths and Missing Data in Time Series Decomposition. In 2025 IEEE International Conference on Industrial Technology (ICIT) (pp. 1-7). IEEE.

Altinses, D., & Schwung, A. (2025). Performance benchmarking of multimodal data-driven approaches in industrial settings. Machine Learning with Applications, 100691.

2024

Altinses, D., Torres, D. O. S., Gobachew, A. M., Lier, S., & Schwung, A. (2024). Synthetic Dataset Generation for Optimizing Multimodal Drone Delivery Systems. Drones (2504-446X), 8(12).

Altinses, D., Salazar Torres, D. O., Schwung, M., Lier, S., & Schwung, A. (2024). Optimizing Drone Logistics: A Scoring Algorithm for Enhanced Decision Making across Diverse Domains in Drone Airlines. Drones, 8(7), 307.

2023

Altinses, D., & Schwung, A. (2023, October). Deep Multimodal Fusion with Corrupted Spatio-Temporal Data Using Fuzzy Regularization. In IECON 2023-49th Annual Conference of the IEEE Industrial Electronics Society (pp. 1-7). IEEE.

Altinses, D., & Schwung, A. (2023, October). Multimodal Synthetic Dataset Balancing: a Framework for Realistic and Balanced Training Data Generation in Industrial Settings. In IECON 2023-49th Annual Conference of the IEEE Industrial Electronics Society (pp. 1-7). IEEE.