Refine
Document Type
- Conference Proceeding (6)
- Article (reviewed) (2)
- Doctoral Thesis (1)
- Master's Thesis (1)
Conference Type
- Konferenzartikel (6)
Language
- English (10)
Keywords
- machine learning (10) (remove)
Institute
- Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) (6)
- ivESK - Institut für verlässliche Embedded Systems und Kommunikationselektronik (3)
- Fakultät Elektrotechnik und Informationstechnik (E+I) (bis 03/2019) (2)
- Fakultät Medien (M) (ab 22.04.2021) (1)
- INES - Institut für nachhaltige Energiesysteme (1)
Open Access
- Open Access (7)
- Diamond (3)
- Closed Access (2)
- Gold (2)
- Bronze (1)
- Closed (1)
The last decades have seen the evolution of industrial production into more sophisticated processes. The development of specialized, high-end machines has increased the importance of predictive maintenance of mechanical systems to produce high-quality goods and avoid machine breakdowns. Predictive maintenance has two main objectives: to classify the current status of a machine component and to predict the maintenance interval by estimating its remaining useful life (RUL). Nowadays, both objectives are covered by machine learning and deep learning approaches and require large training datasets that are often not available. One possible solution may be transfer learning, where the knowledge of a larger dataset is transferred to a smaller one. This thesis is primarily concerned with transfer learning for predictive maintenance for fault classification and RUL estimation. The first part presents the state-of-the-art machine learning techniques with a focus on techniques applicable to predictive maintenance tasks (Chapter 2). This is followed by a presentation of the machine tool background and current research that applies the previously explained machine learning techniques to predictive maintenance tasks (Chapter 3). One novelty of this thesis is that it introduces a new intermediate domain that represents data by focusing on the relevant information to allow the data to be used on different domains without losing relevant information (Chapter 4). The proposed solution is optimized for rotating elements. Therefore, the presented intermediate domain creates different layers by focusing on the fault frequencies of the rotating elements. Another novelty of this thesis is its semi and unsupervised transfer learning-based fault classification approach for different component types under different process conditions (Chapter 5). It is based on the intermediate domain utilized by a convolutional neural network (CNN). In addition, a novel unsupervised transfer learning loss function is presented based on the maximum mean discrepancy (MMD), one of the state-of-the-art algorithms. It extends the MMD by considering the intermediate domain layers; therefore, it is called layered maximum mean discrepancy (LMMD). Another novelty is an RUL estimation transfer learning approach for different component types based on the data of accelerometers with low sampling rates (Chapter 6). It applies the feature extraction concepts of the classification approach: the presented intermediate domain and the convolutional layers. The features are then used as input for a long short-term memory (LSTM) network. The transfer learning is based on fixed feature extraction, where the trained convolutional layers are taken over. Only the LSTM network has to be trained again. The intermediate domain supports this transfer learning type, as it should be similar for different component types. In addition, it enables the practical usage of accelerometers with low sampling rates during transfer learning, which is an absolute novelty. All presented novelties are validated in detailed case studies using the example of bearings (Chapter 7). In doing so, their superiority over state-of-the-art approaches is demonstrated.
Artificial intelligence (AI), and in particular machine learning algorithms, are of increasing importance in many application areas but interpretability and understandability as well as responsibility, accountability, and fairness of the algorithms' results, all crucial for increasing the humans' trust into the systems, are still largely missing. Big industrial players, including Google, Microsoft, and Apple, have become aware of this gap and recently published their own guidelines for the use of AI in order to promote fairness, trust, interpretability, and other goals. Interactive visualization is one of the technologies that may help to increase trust in AI systems. During the seminar, we discussed the requirements for trustworthy AI systems as well as the technological possibilities provided by interactive visualizations to increase human trust in AI.
The use of artificial intelligence continues to impact a broad variety of domains, application areas, and people. However, interpretability, understandability, responsibility, accountability, and fairness of the algorithms' results - all crucial for increasing humans' trust into the systems - are still largely missing. The purpose of this seminar is to understand how these components factor into the holistic view of trust. Further, this seminar seeks to identify design guidelines and best practices for how to build interactive visualization systems to calibrate trust.
Much of the research in the field of audio-based machine learning has focused on recreating human speech via feature extraction and imitation, known as deepfakes. The current state of affairs has prompted a look into other areas, such as the recognition of recording devices, and potentially speakers, by only analysing sound files. Segregation and feature extraction are at the core of this approach.
This research focuses on determining whether a recorded sound can reveal the recording device with which it was captured. Each specific microphone manufacturer and model, among other characteristics and imperfections, can have subtle but compounding effects on the results, whether it be differences in noise, or the recording tempo and sensitivity of the microphone while recording. By studying these slight perturbations, it was found to be possible to distinguish between microphones based on the sounds they recorded.
After the recording, pre-processing, and feature extraction phases we completed, the prepared data was fed into several different machine learning algorithms, with results ranging from 70% to 100% accuracy, showing Multi-Layer Perceptron and Logistic Regression to be the most effective for this type of task.
This was further extended to be able to tell the difference between two microphones of the same make and model. Achieving the identification of identical models of a microphone suggests that the small deviations in their manufacturing process are enough of a factor to uniquely distinguish them and potentially target individuals using them. This however does not take into account any form of compression applied to the sound files, as that may alter or degrade some or most of the distinguishing features that are necessary for this experiment.
Building on top of prior research in the area, such as by Das et al. in in which different acoustic features were explored and assessed on their ability to be used to uniquely fingerprint smartphones, more concrete results along with the methodology by which they were achieved are published in this project’s publicly accessible code repository.
Learning to Walk With Toes
(2020)
This paper explains how a model-free (with respect to the robot model and the behavior to learn) approach can facilitate learning to walk from scratch. It is applied to a simulated Nao robot with toes. Results show an improvement of 30% in speed compared to a model without toes and also compared to our model-based approach, but with less stability.
The conversion of space heating for private households to climate-neutral energy sources is an essential component of the energy transition, as this sector as of 2018 was responsible for 9.4 % of Germany’s carbon dioxide emissions. In addition to reducing demand through better insulation, the use of heat pumps fed with electricity from renewable energy sources, such as on-site photovoltaics (PV) systems, is an important solution approach.
Advanced energy management and control can help to make optimal use of such heating systems. Optimal here can e.g. refer to maximizing self-consumption of self-generated PV power, extended component lifetime or a grid-friendly behavior that avoids load peaks. A powerful method for this is model predictive control (MPC), which calculates optimal schedules for the controllable influence variables based on models of the system dynamics, current measurements of system states and predictions of future external influence parameters.
In this paper, we will discuss three different use cases that show how artificial intelligence can contribute to the realization of such an MPC-based energy management and control system. This will be done using the example of a real inhabited single family home that has provided the necessary data for this purpose and where the methods are implemented and tested. The heating system consists of an air-water heat pump with direct condensation, a thermal stratified storage tank, a pellet burner and a heating rod and provides both heating and hot water. The house generates a significant portion of its electricity needs through a rooftop PV system.
Positioning mobile systems with high accuracy is a prerequisite for intelligent autonomous behavior, both in industrial environments and in field robotics. This paper describes the setup of a robotic platform and its use for the evaluation of simultaneous localization and mapping (SLAM) algorithms. A configuration using a mobile robot Husky A200, and a LiDAR (light detection and ranging) sensor was used to implement the setup. For verification of the proposed setup, different scan matching methods for odometry determination in indoor and outdoor environments are tested. An assessment of the accuracy of the baseline 3D-SLAM system and the selected evaluation system is presented by comparing different scenarios and test situations. It was shown that the hdl_graph_slam in combination with the LiDAR OS1 and the scan matching algorithms FAST_GICP and FAST_VGICP achieves good mapping results with accuracies up to 2 cm.
An Overview of Technologies for Improving Storage Efficiency in Blockchain-Based IIoT Applications
(2022)
Since the inception of blockchain-based cryptocurrencies, researchers have been fascinated with the idea of integrating blockchain technology into other fields, such as health and manufacturing. Despite the benefits of blockchain, which include immutability, transparency, and traceability, certain issues that limit its integration with IIoT still linger. One of these prominent problems is the storage inefficiency of the blockchain. Due to the append-only nature of the blockchain, the growth of the blockchain ledger inevitably leads to high storage requirements for blockchain peers. This poses a challenge for its integration with the IIoT, where high volumes of data are generated at a relatively faster rate than in applications such as financial systems. Therefore, there is a need for blockchain architectures that deal effectively with the rapid growth of the blockchain ledger. This paper discusses the problem of storage inefficiency in existing blockchain systems, how this affects their scalability, and the challenges that this poses to their integration with IIoT. This paper explores existing solutions for improving the storage efficiency of blockchain–IIoT systems, classifying these proposed solutions according to their approaches and providing insight into their effectiveness through a detailed comparative analysis and examination of their long-term sustainability. Potential directions for future research on the enhancement of storage efficiency in blockchain–IIoT systems are also discussed.
Subspace clustering aims to find all clusters in all subspaces of a high-dimensional data space. We present a massively data-parallel approach that can be run on graphics processing units. It extends a previous density-based method that scales well with the number of dimensions. Its main computational bottleneck consists of (sequentially) generating a large number of minimal cluster candidates in each dimension and using hash collisions in order to find matches of such candidates across multiple dimensions. Our approach parallelizes this process by removing previous interdependencies between consecutive steps in the sequential generation process and by applying a very efficient parallel hashing scheme optimized for GPUs. This massive parallelization gives up to 70x speedup for
the bottleneck computation when it is replaced by our approach and run on current GPU hardware. We note that depending on data size and choice of parameters, the parallelized part of the algorithm can take different percentages of the overall runtime of the clustering process, and thus, the overall clustering speedup may vary significantly between different cases. However, even
in our ”worst-case” test, a small dataset where the computation makes up only a small fraction of the overall clustering time, our parallel approach still yields a speedup of more than 3x for the complete run of the clustering process. Our method could also be combined with parallelization of other parts of the clustering algorithm, with an even higher potential gain in processing speed.
The paper describes the methodology and experimental results for revealing similarities in thermal dependencies of biases of accelerometers and gyroscopes from 250 inertial MEMS chips (MPU-9250). Temperature profiles were measured on an experimental setup with a Peltier element for temperature control. Classification of temperature curves was carried out with machine learning approach.
A perfect sensor should not have thermal dependency at all. Thus, only sensors inside the clusters with smaller dependency (smaller total temperature slopes) might be pre-selected for production of high accuracy inertial navigation modules. It was found that no unified thermal profile (“family” curve) exists for all sensors in a production batch. However, obviously, sensors might be grouped according to their parameters. Therefore, the temperature compensation profiles might be regressed for each group. 12 slope coefficients on 5 degrees temperature intervals from 0°C to +60°C were used as the features for the k-means++ clustering algorithm.
The minimum number of clusters for all sensors to be well separated from each other by bias thermal profiles in our case is 6. It was found by applying the elbow method. For each cluster a regression curve can be obtained.