Continuous transformation of computer systems from large-scale isolated units to the application-specific distributed units are rising the challenge of time synchronisation due to their associated and time-critical actions. These kinds of transformation are noticeable in many sectors like, Automation, Distributed Monitoring, Power Distribution Systems (SmartGrids) and Media networking (AVB*). As the large-scale distributed systems are becoming more complex and modular, where each distributed module/node is communicating via a standardised communication medium, the demand for precision and accuracy of time synchronisation is increasing.
Precision Time Protocol [1, 2] claims to overcome time synchronisation challenges; it provides a self-organising time synchronisation protocol while utilising the existing ethernet based network capabilities for timekeeping applications. The protocol operates on master/slave hierarchy on local area networked devices and provides the opportunity to achieve the sub-microsecond level accuracy on existing (multicast supported) LAN network. Time the message is sent or received at master/slave device is recorded (Timestamped), the precision of this time measurements is essential, it sets a fundamental level of accuracy for time synchronisation. The Timestamp for a PTP* message is generated either on Software (Application Layer) or Hardware level (MAC or PHY).
A research objective of the study is to find out the maximum attainable accuracy in both hardware and software timestamping-based implementation of the protocol and to analyse the precision uncertainties between Hardware and Software-based solutions. As the PTP protocol does not require additional physical network infrastructure to establish a primary PTP network So, It is probable that the present resource exhausting applications running on the same system node or in the overall network may affect the performance of the PTP clock synchronisation process. In order to find out the possible effect on the performance of established PTP network, the simulated load is applied in different dimensions (CPU, I/O and Network) on PTP nodes in the established PTP network.
The protocol hierarchy is established using multiple IEEE 1588 clock based SoC* (hardware-assisted) platform in different test case scenarios (Hardware/Software Assisted and with Simulated load), an open-source solution [3] is used to establish the network in Linux environment. The log data generated in a specific scenario is prepared and imported to MATLAB workspace, where comparison of results in different scenarios are compiled in graphical form. The paper presents the summary of the PTP protocol, PTP infrastructure in Linux, and it presents the results collected during the different test case scenario, which are designed to achieve the research objectives, the results are then further analysed and evaluated to draw a conclusion.
The result shows the significant differences between the hardware and software-based synchronisation, in both accuracy and precision. On the other hand, in simulated load scenarios, network tests showed some considerable effect. The results point out some factors and their magnitude of disruption in synchronisation process, which will be helpful in designing synchronisation network.
* PTP: Precision Time Protocol || AVB: Audio Video Bridging || SoC: System-on-Chip
[1] IEEE Std 1588 -2002. IEEE Standard for a Precision Clock Synchronization Protocol for Networked Measurement and Control Systems.
url:
https://standards.ieee.org/findstds/standard/1588-2002.html.
[2] IEEE Std 1588-2008 (Revision of IEEE Std 1588-2002). IEEE Standard for a Precision Clock Synchronization Protocol for Networked Measurement and
Control Systems. July 24, 2008.
url: https://standards.ieee.org/findstds/standard/1588-2008.html.
[3] Cristian Marinescu Richard Cochran and Christian Riesch. “Synchronizing the Linux System Time to a PTP Hardware Clock.” Precision Clock Synchronization
for Measurement Control and Communication (ISPCS), International IEEE Symposium on 12-16 Sept. 2011.
In this paper authors consider the organizational, methodological and technical aspects of the training professionals call in the field of embedded automation. Also proposed the approaches to solution of described needs. Given the example of the organization of the Interdepartmental Center of Embedded Systems of Automation and Computing, using the developed methodologies and microcontroller-based equipment.
Besides, the authors consider the development of a fuzzy solver which is a function-oriented controller. the technique of performance comparison of different fuzzy algorithms implementation options is worked out; operation on algorithm coding in the VHDL language is carried out.
In any complex audio video networks such as professional audio recording studios, automotive or in-flight infotainment systems, concert avenues or even home entertainment systems, the connection between the various audio/video sources and sinks are mostly analog, point to point and serve a single purpose. This leads to tonnes of confusing cables, each cable serving a specific data exchange. Even the digital solutions such as I2S, S/PDIF, AES3 for short distance connections, most for automotive applications and Firewire (IEEE 1394), HDMI or audio over USB for high bandwidth applications, still require purpose built cables and proprietary software to work correctly and yet they still lack interoperability. Ethernet is now ubiquitous and offers high bandwidth for low cost over reasonable distances, so it makes sense to use Ethernet for audio/video streaming.
The only limitation of Ethernet networks is that, since it is a packet switched network offering reliable, real time delivery of media is a challenge. To overcome this limitation and use the Ethernet networks for reliable, real time, flexible audio/video streaming, a set of protocols have been developed. These set of protocols required for a synchronised real time streaming are collectively called the Audio Video Bridging(AVB). This is a relatively new standard and currently there are only a handful of implementations that make use of this standard.
This paper discusses the implementation of AVB as a part of the open source Linux kernel which can be used and improved by anyone irrespective of applications. The study aims at building AVB stack for an open source hardware Beagle Board and study the various metrics such as latency, synchronisation, throughput etc... between devices like Beagle Bone Black and BeagleBoard X15 along with documenting the interoperability with off the shelf AVB devices such as MAC OS X laptops.
At Kiel University of Applied Sciences the project “Application-oriented Robotics in elderly care” (ARiA) is being carried out in cooperation with the University of Siegen. The project investigates future robotic working environments in elderly care facilities based on participatory approaches anchored in practice in order to establish innovative solution models for the challenges in the field of care robotics
This paper presents the design of an anthropomorphic transradial prosthesis that incorporates a 13 degrees of freedom robotic hand with force touch sensing capabilities. The articulated structure of the fingers and the thumb is designed as a compliant mechanism based on rigid phalanges interconnected with helical extension springs. An underactuated design has been utilized that makes use of tendon drive mechanisms, which simplifies the fabrication and the assembly of the prosthetic hand. The hand is able to actuate each finger individually, as well as rotate the thumb to create opposition or a lateral pinch grip. An artificial forearm is designed for the prosthesis to base the palm of the hand and also to be connected with the remaining limb of the patient. Finally we present early experimental results of the robotic hand in performing common gestures and grasping tasks.
Figure 1: The developed transradial prosthesis
There are three parties involving in a typical product examination through sampling procedure. The authority, the auditor and the examination Laboratory. The main idea of the invention is to isolate these three involving parts and assure anonymity for a forensically approved sampling procedure. Our solution is an intelligent parcel. The parcel creation is commanded through automated call by the authority to an auditor. He collects the investigation sample, he puts it into the intelligent parcel and locks it. Then he places it to a courier service with unknown delivery address which is revealed by the courier at the first transshipment hub. Finally, the parcel is unpacked and examined in the lab. For the initial stage the parcel records every transport detail to a giant undeniable auditing path with full anonymity. Application are sampling in restaurant food and beverages, supermarket products, organic agriculture, ecology assurance, secure filing etc.
Preventative maintenance of wind energy systems using condition monitoring technologies have become very important in today as the demand for renewable energy sources has dramatically increased. We have applied data driven methods using big data architectures in order to meet all the requirements. We show some results of our recent research project.
Magnetic Induction Tomography (MIT) can be used to measure the conductivity of objects contactlessly. This has many advantages in applications from healtcare to the food industry. Especially the contactless measurement of biological tissues in the food industry is interesting, allowing a way to measure the state of the food without distruction and through packaging. This paper presents a system using a purpose-build RF power amplifier to improve the measurements of an existing MIT/MIS measurement system for the use with biological tissues (banana and apples).
When measuring the conductivity of biological tissues over a range of frequencies, one can obtain information about the cellular structure of the tissue. The electrical conductivity of biological tissues is frequency-dependent and ranges from DC to 1 GHz. Of the three main dielectric dispersions, the -dispersion lies in the radio-frequency range. The Maxwell–Wagner relaxation, an interfacial relaxation process occurring in materials containing boundaries between two different dielectrics, produces the -dispersion, which is very distinct in biological tissues. The information in the -dispersion contains the information about the cellular structure.
Biological tissues have the disadvantage of a relativ low conducticity, thus requiring a much higher measurement frequency (up to 20 MHz) compared to godd conductors. At the same time it is important to measure at low frequencies, in order to obtain the -dispersion. Combining both, low and high frequency measurements using one transmitter coil, is challenging for the power amplifier driving the transmitter coil. A typical MIT/MIS systems consists of different components: A signal generator, a RF power amplifier, a transmitter and a receiver coil, a receiver pre-amplifier and a measurement system. The Keil MIT/MIS system is based on the Red Pitaya processor platform (Xilinx Zynq FPGA based) and uses a single receiver coil with a wideband pre-amplifier. The transmitting section contains the Red Pitaya processing board as signal generator and a standard RF power amplifier (Pötschke 5600) with a single transmitter coil. Driving the single transmitter coil at 200 kHz as well as at 20 MHz is demanding. While the system provides a phase precision of 100 at 10 MHz, the phase precision at 200 kHz is much lower. One factor is the ability of the RF amplifier to provide sufficient current at lower frequencies.
The aim of this project was to develop a new wideband RF power amplifier for the existing Kiel MIT/MIS system. The bandwith of the power amplifier was set to 200 kHz to 20 MHz with the ability to deliver up to 3 A at low frequencies. In order to achieve this, four ADA4870 operational amplifiers, each capable of 1 A output and a bandwith of 52 MHz, were used in parallel. The new power amplifier was developed, tested and compared to the commercially available Pöttscke 5600. The results showed an improved performance, in terms of phase precision, at lower frequencies, and the same performance at higher frequencies. Furthermore an improved sensitivity was observed.
The new power amplifier was evaluated and compared further using biological tissue. Two different types of experiments were set up. The first one was to measure the ripening of bananas over seven days. The second experiment was set to try to determine the type of apple. Both experiments were conducted with the Pötschke 5600 and the new, purpose-build power amplifier. The results showed in the first experiment that the gradient between 500 kHz and 2 MHz of the -dispersion of the bananas provided a good indicator for the ripening using the new power amplifier. The second experiment showed a different conductivity for each type of apple at 1 MHz, thus it was possible to differentiate between them. However, factors like different size of the objects and longterm drift was excluded in these experiments.
Robotics is playing an increasingly important role in industry, especially with regard to human-machine collaboration. Companies such as KUKA or Universal Robots offer lightweight robots for sensitive and filigree assembly applications. These robots impress with unique kinematics with high-quality drive systems combined with a high security level.
The aim of this project was to develop a realistic model for teaching purposes, which can simulate the essential kinematic properties of a real iiwa made by KUKA. Mechatronics students achieved the solution by miniaturizing the original with CAD and 3D printing. Concerning drive technology we focused on compact stepper motors with integrated electronics and sensors, resulting in a desktop model that is cost-effective to manufacture, easy to transport and interface with the control software of the original robot.
The work was created as part of a six-month project, which is completed in the sixth semester of the bachelor's degree in mechatronics at Kiel University of Applied Sciences. As a central computer a Raspberry PI of the third generation was used. Python 3 together with high level libraries for the drive control and the graphical user interface was applied.
In the course of the digitization of production facilities, the tracking of assets in the supply chain becomes increasingly relevant for the manufacturing industry. Asset tracking refers to the method of tracking physical production orders, either by scanning bar code labels on production bins or by using tags with UWB, GPS, BLE or RFID technology attached to the bins that transmit their location to suitable system. Current research and development projects use the Bluetooth standard 4.2.
Although Bluetooth 4.2 is very energy-efficient and good results are achieved, the RSSI value loses its significance after five meters over the distance of the object to the receiver. Taking into account the test results of current BT 4 AT systems, this paper describes the structure and prototypical implementation of a Low Cost BT5 Asset Tracking System.
This open system is characterized by the reduction of unnecessary data transmission, greater configurability (e.g. transmission interval) and a generally higher intelligence of the individual receiving stations.
Therefore, we expect a higher accuracy and reliability due to the collaboration of the receiving stations.
The objective is to create a system that can be used economically for industrial companies while still providing adequate results in the area recognition of assets. The proposed solution is expected to reduce the costs associated with tracking and managing assets and improve asset utilization and operational efficiency.
The next steps are the evaluation of the system in the industrial environment as well as a quantitative comparison of BT4 and BT5 in the use of AT systems.
For various reasons pseudo random sequences play a significant role in several technical areas. Such sequences can easily be generated by Linear Feedback Shift Registers (LFSR). The design of such LFSRs is based on the mathematical theory of Finite Fields, so called Galois.
The goal of this presentation is to build a bridge from the mathematics of Galois Fields, in particular Extension Fields towards the related circuit theory in terms of Linear Feedback Shift Registers and their usage in several technical applications. Therefore the presentation is structured by the following chapters:
The International Assisted Communications for Education Project, I-ACE is a 2-year EU funded project started in October 2016 and it involves 7 partners across Europe. The number of people who are deaf or have hearing issues rises with age and, in all age bands, is rising over time. The project examines the possibilities and prospects deaf students have in education, and also outlines the obstacles they face due to lack of effective communication. I-ACE aims to provide an effective tool to the deaf and hard hearing, to overcome barriers that exist, so that they achieve their full potential. Research has shown that there are approximately 200 different sign languages across the world (47 of them in Europe), with different language origins and limited similarity. I-ACE project delivers an automatic bi-directional translation between sign language and written speech across 6 different languages (Portuguese, Slovenian, Greek, German, Cyprus and UK sign languages).
The project is a communication bridge between conventional classroom and the deaf student and between deaf and non-deaf people. Consequently, this will result to better integration and better education. This paper describes the rationale, aims and objectives of the project and presents the outcomes and the acceptance of the automatic bi-directional translation tool from the deaf individuals and their hearing classmates.
The automatic detection and accurate localization of key points is a crucial task in medical imaging. It is necessary for tasks like diagnosis, surgical planning, and post-operative assessment. A common approach to localize multiple key points is to combine multiple independent localizers for individual key points with a spatial regularizer, e.g., a conditional random field (CRF). Its configuration, e.g., the CRF topology and potential functions, often has to be manually specified w.r.t. the target application.
We present a general framework to automatically learn the optimal configuration of a CRF for localizing multiple key points. Furthermore, we introduce a novel “missing” label for each key point (node in the CRF). Key idea is to define a pool of potentials and optimize their CRF weights and the potential values for missing key points in a learning framework. Potentials with a low weight are removed, thus optimizing the graph topology. This allows to easily transfer our framework to new applications, and to integrate different localizers. Further advantages of our algorithm are its low test runtime, low amount of training data, and interpretability. We illustrate its feasibility in a detailed evaluation on three medical datasets featuring high degrees of pathologies and outliers.
Despite the high demand of skilled technical staff by the industry, it has been challenging for Universities to attract new students in the Science, Technology, Engineering and Mathematics (STEM or MINT in German) subjects. Furthermore, STEM disciplines suffer a significant dropout rate. Reasons for this are multi-faceted including hard subjects as well as a strong gender imbalance. Schools, Universities and Industry are spending significant efforts to improve the situation by introducing orientation weeks, workshops and tailoring lighter weight study programs. For example, robotic competitions starting even in primary school (e.g. Lego First League) have resulted in stronger affection of kids towards STEM subjects and a higher likelihood of them to actually consider a career in that domain. Similar effects result from computer gaming and the interest in how gaming technology actually works.
We follow this thought by capitalizing on the interest in electronic music of prospective students. Music is of many peoples interest, whereas in particular electronic music has been becoming ever more popular of the last 20 years. Many young adults identify themselves strongly with specific genres and are deeply involved in respective communities. Frequently, propelled by the omnipresence of affordable home studio technology, they produce their own music and release it to the market.
About five years ago, we have founded the Creative Technology AG, which is a working group with focus on music technology. Since then the group has attracted new students to the University with particular interest in that field. In several instances students who are engaged in the group successfully embarked voluntarily in hard subjects such digital signal processing, which were not part of their compulsory curriculum. This makes us confident that with focusing on music technology we are able to spark further interest in STEM subjects as well as increase the intrinsic motivation to successfully complete a technical degree.
A concrete example of our work includes the teaching of sound synthesis by means of modular synthesizers. Such a synthesizer consists of several electronic modules such as an oscillator, filter, envelope generator, voltage control, amplifiers etc. which can be interconnected by cables. We use real hardware as well as virtual software instruments. This allows students to have a direct access and grasp the technology. We offer student project to build an own module, which includes aspects such as electronics, software engineering, design all the way to creating own products and going to market. Within our new study program Media Engineer (Medien Ingenieur) students of both faculties Media as well as Computer Science and Electrical Engineering will have the opportunity to have the right mix of deepened technical knowledge and media application (here in particular musical) focus.
Information security becomes ever entangled in many aspects of our societies because of ubiquitous adoption of information technology. Nowadays it’s hard to imagine that societies would survive without information technology. Although this technology enables us to be much more productive and makes access and distribution of information really easy, it also carries with itself a bunch of security issues. If the information on the systems used by our employers or banks becomes exposed to an attacker, the consequences can be drastic. A few common software development vulnerabilities can lead to security issues in our applications. The main categories of software development vulnerabilities include input validation attacks, authentication attacks, authorization attacks, cryptographic attacks, buffer overflows and race conditions
In this presentation we go through each of these topics and see how we can defend ourselves against these threats while developing applications.
This paper presents the design and control of a low cost cart-pendulum platform, developed for educational purposes. The cart-pendulum is one of the most popular laboratory experiments for teaching and demonstrating underactuated mechanical systems Fieldsand non-linear control methods, but the cost of commercially available units is typically quite high.
Our implementation was built around a linear slide mechanism, salvaged from an old dot-matrix printer, which employs cable transmission to actuate the cart via a DC-motor, driven by a current-control servo amplifier. The platform is outfitted with high-precision optical encoders to measure the pendulum’s angle and the cart’s position. Following the development of the system’s equations of motion, we present a variety of stabilizing controllers in both the upright (inverted pendulum) and the downward (gantry) configurations. A Rapid Control Prototyping approach, based on the WinCon real-time software extension of Matlab/Simulink, is adopted for implementation of these controllers on the developed platform. The experimental results are in very good agreement with those obtained in simulation from a virtual model of the system, which is supplied to students as an accompanying educational tool.
Figure 1. The developed cart-pendulum educational platform.
In this paper a case study of Bluetooth 4.0 sensor network is presented, which features an improved energy performance due to including the sensor data into the advertising packets of the peripheral device so that the subsequent connection becomes unnecessary.
By using the presented technique, a reduction in energy consumption of approximately 40%, as compared with the usual communication mode, was experimentally verified with a built proof-of-concept BLE network comprising different sensors.
Professional life is nowadays heavily depending on mobility and demands professionals to excel in communication skills at an international, cross-cultural environment. However, soft skills, as well as international exposure, are rarely addressed by undergraduate courses. Each year an Academic International Mobility program takes place in Universities, blending students from different EU countries who participate in designing and developing a business idea. The aim of the project is to remodel international mobility and empower students’ employability through blended mobility. There were 2 different projects at Blended Aim 2018. In the first project, students had to develop an application for people that suffer from epilepsy, in the second project students had to develop a commercial Chatbot.
Application for people with epilepsy
People with epilepsy are threatened by social isolation as seizures come unexpected. We envision a concept that allows people with epilepsy to go out with their friends and family safely, without stigmatization. In this project we are going to examine the case of a kid suffering from epilepsy and our solution is helping both kids and parents to ensure that their kids can lead to a normal life. The solution we came up with was Epigo, a hybrid mobile application for the parents. With Epigo, parents are now aware when their child is having a seizure. Using an EEG headset, the child’s application can detect absence seizures. A notification and live-video stream of the child is sent to the parent app once a seizure event is detected. Records of the seizures and the medication are stored on a cloud database. There is also website for parents and doctors were the child’s seizure records are visualized to the associated doctor, in order to help with the diagnosis and the medication.
The students were divided into three scrum teams, and the development team was divided into three main sections, the 2 hybrid applications and the backend service. In each of these sections, tasks were assigned and the developer had to handle them. Ioannis, student of TEI of Crete, worked on the hybrid applications and mostly on the parent application. His main task was the implementation of the design of the application, including all the functionalities such as send and receive notifications and live-video stream from the child’s app, display a daily calendar view with all the seizure records and insert records manually, medication reminders, registration to the application and the login-logout function. In the notification section, a notification is sent from the child’s app once a seizure is detected. The parent app receives this notification and live-video of the child is streamed. During a seizure, the child’s location is sent and displayed on the parent app. The communication between the applications was accomplished through the use of the REST API and Google Cloud Functions. All the previously mentioned capabilities were implemented by using technologies like Android, iOS, Node.js, npm, Apache Cordova, Ionic Framework, AngularJS, Spring Boot RESTful Web Service, Java, Google Maps, Google Cloud Platform and Firebase.
Chatbot assistant for retail stores
Nowadays the amount of information on the internet is enormous and most of the people want to find something easily and without much time spent. Furthermore, every retail store spends valuable time on training new employees and it would be a waste of time and effort to make them memorize products information on top of all the other skills they learn during it. That’s the reason that we set as our main target to create a way to get information fast and without any difficulties. Our approach to the solution of these problems was to develop a chat bot for retail shops that can help a company and its workers to have quick and easy access to information, uses and all the important properties of a searched product. It can also understand natural language and return answers to specific or general questions that a user asked. In case an answered question already exists in the database(in our case the company’s REST API), it can return the answer immediately without the need to analyze the natural language.
Our project team was divided in three group: designers, business development/management and developers. The application was also divided into three categories: backend(bot), frontend(webpage that hosted it), LUIS(natural language understanding) and etc. Each category had its own tasks that was being assigned to the developers. Giorgos, student of TEI of Crete, worked on the initiation of LUIS and its implementation and afterwards, he worked on the bot programming(backend). His main task on the bot programming part was the question handling, which is the part that the bot searches for an existing question on the KnowledgeFlow API and finds the correct answer(or multiple correct answers) for the question. In case it does find the question with the answer, it returns it. Otherwise if it finds the question but it’s unanswered it informs the user about it. Also, there is an option that if you don’t find any similar question, it registers the question back in the KnowledgeFlow API where the developers can answer it for future search. The communication between our chatbot and the company’s API( RESTful API) was accomplished through the use axios which handles HTTP request on node.js. All the previously mentioned capabilities were implemented by using technologies like Bot Builder SDK, Node.js, npm, LUIS and React.
The Open Source Applications for Industrial Automation (OpenIn) project is a three-year duration Project funded by the European Union Erasmus+ Strategic Partnerships Programme, started in October 2016 and it involves 4 partners across Europe. OpenIn project focuses to help higher VET and HEIs institutions, Technical Universities as well as enterprises to integrate innovative automation systems by providing them with Training Courses for both students and teachers on free open source hardware and software for designing automated systems. Arduino was selected to be the teaching and training platform.
During the last year a training meeting and a summer school were held in Portο and in Heraklion respectively. In this paper, the aims and objectives of the project are presented along with the outcomes of the training events in Porto and Heraklion. Moreover, the structure and the modules of the training course are outlined along with the results of the summer school. The answers given by the summer school participants to relevant questionnaires showed that the project fulfilled its aims and objectives
Nanopore DNA sequencing enables the sequence determination of single DNA molecules up to 10,000 times longer than that currently permitted by second-generation sequencing platforms. Nanopore sequencing gives real-time access to sequencing data and enables the detection of epigenetic modifications. This unique feature set is poised to foster the development of novel biomedical applications previously deemed unfeasible.
Nanopore sequencing is based on picoampere scale measurement of current modulated by DNA or RNA polymers traveling through a nanometer opening between two compartments. Each of the five canonical nucleobases (A, T, G, C, U) has a characteristic electrical resistance, which ultimately enables the determination of the precise base sequence. However, a substantial computational effort is required to resolve the underlying sequence from a time-warped and noisy stream of digitized current measurements.
Recently, a wide range of digital signal analysis and machine learning methods have been developed for Nanopore sequencing applications. Clinically relevant questions, such as the quantification of short repetitive DNA sequences remain an unresolved challenge to current generic, state-of-the-art nanopore data analysis methods. We believe realistic simulation of the signal stream can be instrumental in the development of tailored algorithms for such novel biomedical applications.
Based on our work with the Oxford Nanopore Technologies nanopore sequencing platform, we have developed a software package for the in silico generation of realistic, raw-signal-level data. It starts from a reference genome in conjunction with a configuration and model file derived from real-world nanopore sequencing experiments as input. To validate our algorithm, we have sequenced custom synthetic DNA, and in so doing have generated a “ground-truth” data set potentially useful for downstream algorithm development. Additionally, we demonstrate the utility for method development in typical clinical use cases.
Supplementary examples, raw data obtained from our synthetic DNA sequencing experiments and the software are available under the open Mozilla Public license.
The development of affordable electric vehicles (EV) for mass production is a main target of car manufactures nowadays. One key component of an EV is the inverter that converts the dc current of the battery to a three-phase ac current to drive the electric machine. In order to obtain sufficient design freedom for the electrical drive system and to reduce the total cost, it is necessary to minimize the volume of the inverter and thus the components inside the inverter. This is reached by the integration of several subcomponents, i.e. gate-drivers or temperature and current sensors.
This paper proposes the development of an integrated current sensing method for a new ultra compact traction inverter with a power density of 100kW/l. The current measurement is part of a cylindrical 80kW inverter assembly for the direct integration into the frame of an electric machine. The paper presents, the innovative strategic approach for cost-efficient miniaturization of the current measurement. The concept allows a direct integration of the current measurement in the internal ac-bus of the inverter, which enables a reduction of the size of more than 90% compared to a standard current sensor.
The projects objective was to visualise digital content within the windows of a building. As a first application the game "TicTacToe" for two players was chosen. Nine windows were used to display the state of the game play.
The system was implemented using LED stripes controlled by an ESP8266-Wifi-chip (clients) for each of the nine windows. These nine processors communicated using MDNS and sockets over Wifi with a central unit which was also implemented on an ESP8266-Wifi-chip. Each of the nine clients at the windows has a hand-assigned ID. The ESP8266-Wifi-chip were programmed in C/CPP via the Arduino IDE. The whole system runs on 5V power supplies. Furthermore, a mobile phone app was developed using the Android IDE in order to offer a dynamic generated playground matrix depending on the game, in this case with the known parts of "TicTacToe".
To connect to a game, the Wifi of the smartphones has to be the same as the one of the central unit. Once connected the central unit worked using the "first come first served" concept to pick the two players of a round.
To make it possible for players to stand outside the building a project goal is to connect to the central unit via webadress. During the tests, it appeared that the performance of ESP8266 in the central unit does not suffice to have a comfortable game flow. Because of this, one additional project goal for the future is a platform change of the central unit from the ESP8266-chip to the Raspberry PI.
Electrochemical processes in battery systems represent a complex multiphysics problem, which is far too difficult to model in a reasonable time. For example the estimation of the state of charge (SoC) of a battery cell is one of the key diagnostic states for a safe and reliable operation of complete a battery system. It turns out that this quantity is heavily dependent on the actual operation point. In many BMS the current, the cell voltage and temperature of the cell are often used to calculate the state of charge by means of look-up tables and simple current integration over time.
Problems occur, when it comes to cells with a very flat open circuit voltage over nearly the complete SoC-range like in LiFePo4 or Lithium-Titanate-cells. In this case difficulties persist to estimate the SoC with a sufficient accuracy on controller units in comibation with acceptable runtimes.
In this publication, a solution is presented with which the SoC of each cell of the battery system can be estimated based on model previously derived by a supervised machine learning-training. The model represents an artificial feed forward network, which is applicable on each battery cell of a complete battery system. For this purpose, cell-individual impedance spectra are measured by means of a special developed measurement unit as an add-on component of the battery management system. Additionally the temperatures of each cell and the distribution of the former current flow of the cell are further important features, which have to be taken into account for the training process.
Beside the derived model for the estimation of the state of charge, this paper gives an overview of the developed hardware components. It is demonstrated that the presented concept can be implemented in smaller microcontroller units with an acceptable execution time and accuracy, making this solution suitable for industrial applications.
As we have witnessed this era is for electrical vehicles. Every day we used to see new applications for EV (Electrical Vehicles) systems such bikes, cars, skate boards, surf boards and so on. Doubtless secure and reliable communication is vital for systems as well. Thus, to give more flexibility for communications between ev systems we have designed MQTT(MQ Telemetry Transport) based communication and called it E-Pit Stop.
With this application we aim to connect ev systems easily and to get datas such cell diagnostic, voltages, current consumption, charge, discharge status and so on. E-pit stop allows to get all provided data from the ev system easily by any MQTT client application in pc, laptop, mobile phone etc. Since e-pit stop is MQTT based it provides robust and secure communication as well. Thus, it would be easy, secure and reliable to connect EV systems. To sum up, we believe that e-pit stop concept would be beneficial and give an inspiration for EV applications.