Matthias Kahl

Matthias Kahl

Auto-generated profile, based on publicly available information

Matthias Kahl

Chair for Application and Middleware Systems Technische UniversitÃďt MÃijnchen Germany

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

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

Representative Publications

View All Publications
16 October 2018 in TM - Technisches Messen
Verified
Journal Article
Closed access
MKMatthias Kahl
Matthias Kahl
VKVeronika Krause
Veronika Krause
RHRudolph Hackenberg
Rudolph Hackenberg
AHAnwar Ul Haq
Anwar Ul Haq
AHAnton Horn
Anton Horn
AUAnton Udalzow
Anton Udalzow
To support a rational and efficient use of electrical energy in residential and industrial environments, Non-Intrusive Load Monitoring (NILM) provides several techniques to identify state and power consumption profiles of connected appliances. Design requirements for such systems include a low hardware and installations costs for residential, reliability and high-availability for industrial purposes, while keeping invasive interventions into the electrical infrastructure to a minimum. This work introduces a reference hardware setup that allows an in depth analysis of electrical energy consumption in industrial environments. To identify appliances and their consumption profile, appropriate identification algorithms are developed by the NILM community. To enable an evaluation of these algorithms on industrial appliances, we introduce the Laboratory-measured IndustriaL Appliance Characteristics (LILAC) dataset: 1302 measurements from one, two, and three concurrently running appliances of 15 appliance types, measured with the introduced testbed. To allow in-depth appliance consumption analysis, measurements were carried out with a sampling rate of 50 kHz and 16-bit amplitude resolution for voltage and current signals. We show in experiments that signal signatures, contained in the measurement data, allows one to distinguish the single measured electrical appliances with a baseline machine learning approach of nearly 100 % accuracy.

16 May 2017
Verified
Conference Paper
Open access
TKThomas Kriechbaumer
Thomas Kriechbaumer
AHAnwar Ul Haq
Anwar Ul Haq
MKMatthias Kahl
Matthias Kahl
HJHans-Arno Jacobsen
Hans-Arno Jacobsen
Traditional energy measurement fails to provide support to consumers to make intelligent decisions to save energy. Non-intrusive load monitoring is one solution that provides disaggregated power consumption profiles. Machine learning approaches rely on public datasets to train parameters for their algorithms, most of which only provide low-frequency appliance-level measurements, thus limiting the available feature space for recognition. In this paper, we propose a low-cost measurement system for high-frequency energy data. Our work utilizes an off-the-shelf power strip with a voltage-sensing circuit, current sensors, and a single-board PC as data aggregator. We develop a new architecture and evaluate the system in real-world environments. The self-contained unit for six monitored outlets can achieve up to 50 kHz for all signals simultaneously. A simple design and off-the-shelf components allow us to keep costs low. Equipping a building with our measurement systems is more feasible compared to expensive existing solutions. We used the outlined system architecture to manufacture 20 measurement systems to collect energy data over several months of more than 50 appliances at different locations, with an aggregated size of 15 TB.

1 January 2015
Verified
Book Chapter
Closed access
PDF file
MKMatthias Kahl
Matthias Kahl
CGChristoph Goebel
Christoph Goebel
AHAnwar Ul Haq
Anwar Ul Haq
TKThomas Kriechbaumer
Thomas Kriechbaumer
HJHans-Arno Jacobsen
Hans-Arno Jacobsen

24 April 2019 in arXiv
Verified
Preprint
Open access
PDF file
MKMatthias Kahl
Matthias Kahl
TKThomas Kriechbaumer
Thomas Kriechbaumer
DJDaniel Jorde
Daniel Jorde
AHAnwar Ul Haq
Anwar Ul Haq
HJHans-Arno Jacobsen
Hans-Arno Jacobsen
Non-intrusive load monitoring (NILM) is a modern and still expanding technique, helping to understand fundamental energy consumption patterns and appliance characteristics. Appliance event detection is an elementary step in the NILM pipeline. Unfortunately, several types of appliances (e.g., switching mode power supply (SMPS) or multi-state) are known to challenge state-of-the-art event detection systems due to their noisy consumption profiles. Classical rule-based event detection system become infeasible and complex for these appliances. By stepping away from distinct event definitions, we can learn from a consumer-configured event model to differentiate between relevant and irrelevant event transients. We introduce a boosting oriented adaptive training, that uses false positives from the initial training area to reduce the number of false positives on the test area substantially. The results show a false positive decrease by more than a factor of eight on a dataset that has a strong focus on SMPS-driven appliances. To obtain a stable event detection system, we applied several experiments on different parameters to measure its performance. These experiments include the evaluation of six event features from the spectral and time domain, different types of feature space normalization to eliminate undesired feature weighting, the conventional and adaptive training, and two common classifiers with its optimal parameter settings. The evaluations are performed on two publicly available energy datasets with high sampling rates: BLUED and BLOND-50.

14 March 2026
Verified
Conference Paper
Open access
MKMatthias Kahl
Matthias Kahl
MSMartin Schodlock
Martin Schodlock
The retrieval of drill cores is a costly component of mineral exploration. Improving the spatial overview of mineral abundances within a deposit can substantially reduce the need for drilling. We present an unsupervised, automated annotation strategy for pixel-wise mineral labeling in hyperspectral imagery of simple deposit styles. In this context, a simple deposit style refers to deposits with very low or no mineral transitions and predominantly homogeneous, dominant mineral occurrences.The automated annotation is based on handcrafted, mineral- and deposit-specific normalized difference indices (NDI). The objective is to extract a large number of representative mineral spectra for each occurring mineral. These spectra are subsequently used as training data for a targeted hyperspectral neural network with positional encoding, which is expected to generalize better to more complex deposit styles.As a first step, the normalized mineral indices were successfully learned by the network, achieving an F-score of 0.98. This result represents a promising step toward physics-informed, neural-network-based mineral classification in hyperspectral imagery.