Heinrich Wörtche

Prof. Dr. Heinrich Wörtche

Institute of Engineering, Hanze University of Applied Sciences, 9747 AS Groningen, The Netherlands
orcid

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

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

Representative Publications

View All Publications
21 March 2022 in Automation
Verified
Journal Article
Open access
PDF file
NGNatanael Magno Gomes
Natanael Magno Gomes
Felipe Nascimento MartinsFelipe Nascimento Martins
Felipe Nascimento Martins
Jose LimaJose Lima
Jose Lima
Heinrich WörtcheHeinrich Wörtche
Heinrich Wörtche
The number of applications in which industrial robots share their working environment with people is increasing. Robots appropriate for such applications are equipped with safety systems according to ISO/TS 15066:2016 and are often referred to as collaborative robots (cobots). Due to the nature of human-robot collaboration, the working environment of cobots is subjected to unforeseeable modifications caused by people. Vision systems are often used to increase the adaptability of cobots, but they usually require knowledge of the objects to be manipulated. The application of machine learning techniques can increase the flexibility by enabling the control system of a cobot to continuously learn and adapt to unexpected changes in the working environment. In this paper we address this issue by investigating the use of Reinforcement Learning (RL) to control a cobot to perform pick-and-place tasks. We present the implementation of a control system that can adapt to changes in position and enables a cobot to grasp objects which were not part of the training. Our proposed system uses deep Q-learning to process color and depth images and generates an ϵ-greedy policy to define robot actions. The Q-values are estimated using Convolution Neural Networks (CNNs) based on pre-trained models for feature extraction. To reduce training time, we implement a simulation environment to first train the RL agent, then we apply the resulting system on a real cobot. System performance is compared when using the pre-trained CNN models ResNext, DenseNet, MobileNet, and MNASNet. Simulation and experimental results validate the proposed approach and show that our system reaches a grasping success rate of 89.9% when manipulating a never-seen object operating with the pre-trained CNN model MobileNet.

21 January 2023 in Sensors
Verified
Journal Article
Open access
PDF file
MGMatheus Araújo Gava
Matheus Araújo Gava
HRHelder Roberto de Oliveira Rocha
Helder Roberto de Oliveira Rocha
MFMenno Jan Faber
Menno Jan Faber
MSMarcelo Segatto
Marcelo Segatto
Heinrich WörtcheHeinrich Wörtche
Heinrich Wörtche
JSJair Adriano Lima Silva
Jair Adriano Lima Silva
A resource optimization methodology is proposed for application in long range wide area networks (LoRaWANs). Using variable neighborhood search (VNS) and a minimum-cost spanning tree algorithm, it reduces the implementation and the maintenance costs of such low power networks. Performance evaluations were conducted in LoRaWANs with LoRa repeaters to increase coverage, in scenario where the number and the location of the repeaters are determined by the VNS metaheuristic. Parameters such as spread factor (SF), bandwidth and transmission power were adjusted to minimize the network’s total energy per useful bit (Ebit) and the total data collection time. The importance of the SF in the trade-off between (Ebit) and time on-air is evaluated, considering a device scaling factor. Simulation results, obtained after model adjustments with experimental data, show that, in networks with few associated devices, there is a preference for small values of SF aiming at reduction of Ebit. The usage of large SF’s becomes relevant when reach extensions are required. The results also demonstrate that, for networks with high number of nodes, the scaling of devices over time become relevant in the fitness function, forcing an equal distribution of time slots per SF to avoid discrepancies in the time data collection.

1 October 2013
Verified
Conference Paper
Closed access
HBH. H. W. J. Bosman
H. H. W. J. Bosman
ALAntonio Liotta
Antonio Liotta
Giovanni IaccaGiovanni Iacca
Giovanni Iacca
Heinrich WörtcheHeinrich Wörtche
Heinrich Wörtche
The maturing field of Wireless Sensor Networks (WSN) results in long-lived deployments that produce large amounts of sensor data. Lightweight online on-mote processing may improve the usage of their limited resources, such as energy, by transmitting only unexpected sensor data (anomalies). We detect anomalies by analyzing sensor reading predictions from a linear model. We use Recursive Least Squares (RLS) to estimate the model parameters, because for large datasets the standard Linear Least Squares Estimation (LLSE) is not resource friendly. We evaluate the use of fixed-point RLS with adaptive thresholding, and its application to anomaly detection in embedded systems. We present an extensive experimental campaign on generated and real-world datasets, with floating-point RLS, LLSE, and a rule-based method as benchmarks. The methods are evaluated on prediction accuracy of the models, and on detection of anomalies, which are injected in the generated dataset. The experimental results show that the proposed algorithm is comparable, in terms of prediction accuracy and detection performance, to the other LS methods. However, fixed-point RLS is efficiently implement able in embedded devices. The presented method enables online on-mote anomaly detection with results comparable to offline LS methods.

15 March 2023 in Sensors
Verified
Journal Article
Open access
PDF file
Luan Carlos KleinLuan Carlos Klein
Luan Carlos Klein
João Afonso Braun NetoJoão Afonso Braun Neto
João Afonso Braun Neto
JMJoão Mendes
João Mendes
Vítor H. PintoVítor H. Pinto
Vítor H. Pinto
Felipe Nascimento MartinsFelipe Nascimento Martins
Felipe Nascimento Martins
Jose LimaJose Lima
Jose Lima
Localization is a crucial skill in mobile robotics because the robot needs to make reasonable navigation decisions to complete its mission. Many approaches exist to implement localization, but artificial intelligence can be an interesting alternative to traditional localization techniques based on model calculations. This work proposes a machine learning approach to solve the localization problem in the RobotAtFactory 4.0 competition. The idea is to obtain the relative pose of an onboard camera with respect to fiducial markers (ArUcos) and then estimate the robot pose with machine learning. The approaches were validated in a simulation. Several algorithms were tested, and the best results were obtained by using Random Forest Regressor, with an error on the millimeter scale. The proposed solution presents results as high as the analytical approach for solving the localization problem in the RobotAtFactory 4.0 scenario, with the advantage of not requiring explicit knowledge of the exact positions of the fiducial markers, as in the analytical approach.

1 January 2021
Verified
Book Chapter
Closed access
PDF file
NGNatanael Magno Gomes
Natanael Magno Gomes
Felipe Nascimento MartinsFelipe Nascimento Martins
Felipe Nascimento Martins
Jose LimaJose Lima
Jose Lima
Heinrich WörtcheHeinrich Wörtche
Heinrich Wörtche

8 June 2023 in IEEE Sensors Journal
Verified
Journal Article
Closed access
HRHelder Roberto de Oliveira Rocha
Helder Roberto de Oliveira Rocha
RNRicardo Natale
Ricardo Natale
AEArthur C. Estevão
Arthur C. Estevão
DLDanilo M. Louzada
Danilo M. Louzada
Arnaldo Leal-JuniorArnaldo Leal-Junior
Arnaldo Leal-Junior
JSJair Adriano Lima Silva
Jair Adriano Lima Silva
In this work, a feasible and low-cost approach is proposed for level measurement in multiphase systems inside tanks used for petroleum-derived oil production. The developed level sensor system consisted of light-emitting diodes (LED), light-dependent resistor (LDR) and a low-cost microprocessor. Two different types of oil were tested: AW460 and AW68. Linear regression (LR) was applied for 11 scenarios and showed a direct correlation between the level of oil and the sensor’s output. The measurement with AW460 oil presented a perfect linear behaviour, while for AW68, a higher standard deviation was obtained justifying the occurrence of the non linearity in several scenarios. In order to overcome the nonlinear effect, two machine learning techniques were tested: K-nearest neighbors regression (KNNR) and multi-layer perceptron (MLP) neural network regression. The highest correlation coefficient (R 2 ) and the lowest root mean squared error (RMSE) were obtained for AW68 with MLP. Therefore, MLP was used for regression (level prediction for water, oil and emulsion) as well as classification (identify the type of oil in the reservoir) simultaneously. The suggested network exhibited a high accuracy for oil identification (99.801%) and improved linear performance in regression (R 2 = 0.9989 and RMSE = 0.065).