Sebastian Kümmritz

Dr. Sebastian Kümmritz


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 March 2024 in Drones
Verified
Journal Article
Open access
PDF file
SKSebastian Kümmritz
Sebastian Kümmritz
In response to the growing challenges in drone security and airspace management, this study introduces an advanced drone classifier, capable of detecting and categorizing Unmanned Aerial Vehicles (UAVs) based on acoustic signatures. Utilizing a comprehensive database of drone sounds across EU-defined classes (C0 to C3), this research leverages machine learning (ML) techniques for effective UAV identification. The study primarily focuses on the impact of data augmentation methods—pitch shifting, time delays, harmonic distortion, and ambient noise integration—on classifier performance. These techniques aim to mimic real-world acoustic variations, thus enhancing the classifier’s robustness and practical applicability. Results indicate that moderate levels of augmentation significantly improve classification accuracy. However, excessive application of these methods can negatively affect performance. The study concludes that sophisticated acoustic data augmentation can substantially enhance ML-driven drone detection, providing a versatile and efficient tool for managing drone-related security risks. This research contributes to UAV detection technology, presenting a model that not only identifies but also categorizes drones, underscoring its potential for diverse operational environments.

30 January 2024 in Preprints.org
Verified
Preprint
Open access
PDF file
SKSebastian Kümmritz
Sebastian Kümmritz
In response to the growing challenges in drone security and airspace management, this study introduces an advanced drone classifier, capable of detecting and categorizing Unmanned Aerial Vehicles (UAVs) based on acoustic signatures. Utilizing a comprehensive database of drone sounds across EU-defined classes (C0 to C3), this research leverages machine learning (ML) techniques for effective UAV identification. The study primarily focuses on the impact of data augmentation methods—pitch shifting, time delays, harmonic distortion, and ambient noise integration—on classifier performance. These techniques aim to mimic real-world acoustic variations, thus enhancing the classifier's robustness and practical applicability. Results indicate that moderate levels of augmentation significantly improve classification accuracy. However, excessive application of these methods can negatively affect performance. The study concludes that sophisticated acoustic data augmentation can substantially enhance ML-driven drone detection, providing a versatile and efficient tool for managing drone-related security risks. This research contributes to UAV detection technology, presenting a model that not only identifies but also categorizes drones, underscoring its potential for diverse operational environments.

3 July 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
SKSebastian Kümmritz
Sebastian Kümmritz
Classroom acoustic design is often strongly guided by reverberation-time criteria, particularly in regulatory and practical planning contexts. Beyond reverberation time, speech intelligibility is commonly assessed using established descriptors such as C50, D50, and STI, which capture early-to-late energy balance and modulation transfer. This field study examines whether edge-based acoustic treatments influence speech intelligibility thresholds under steady background noise in ways that are not fully reflected by these conventional scalar room-acoustic descriptors. Four geometrically comparable primary-school classrooms were investigated: one acoustically optimized ceiling reference, one untreated control, and two rooms equipped with edge-based systems (primarily absorptive vs. primarily reflective/scattering). Room impulse responses were measured at multiple source–receiver configurations, and standard parameters (STI, C50, D50, T30) were derived. Psychoacoustic sentence-repetition tests with children (7–9 years) under controlled, fixed noise-playback conditions were used to determine speech-level and SNR thresholds at a predefined sentence-repetition intelligibility criterion. Both edge-based treatments showed higher STI values than the untreated control room, while C50 and D50 improved relative to the corresponding untreated room states. Notably, both edge-based systems reached the same median STI (0.68), intermediate between the untreated control (0.60) and the ceiling-treated reference (0.80). In the psychoacoustic evaluation, however, SNR thresholds in the treated rooms were already comparable to or better than in the ceiling-treated reference, despite their clearly lower STI. Although not all SNR comparisons reached statistical significance after correction for multiple testing, effect sizes indicated consistent directional improvements. This study was designed as an exploratory pilot investigation to assess the plausibility and directionality of potential effects rather than to provide definitive statistical confirmation. This discrepancy between intermediate STI values and near-reference-level psychoacoustic performance findings suggest that intelligibility under noise is not fully reflected by the conventional scalar room-acoustic descriptors considered here and motivates further investigation of structural room-acoustic characteristics in small classrooms.