Time filter

Source Type

News Article | April 29, 2016
Site: phys.org

Harvesting cauliflower is a science unto itself because the white heads are hidden beneath numerous leaves. This means that pickers have to pull back the protective leaves head for head to look at the cauliflower and decide whether it is ripe for harvesting. Pickers comb a field approximately four to five times in intervals of two to three days until the very last head of cauliflower has been harvested. This work is strenuous and backbreaking. Another challenge for farmers is their need for numerous pickers at once for a short time when harvest season is pending. Finding enough hands for this hard work is often difficult, though. Machines on the other hand would harvest the entire field at once and, since cauliflower heads ripen at different rates, thus even heads that are still too small or unripe. In the future, a machine will harvest cauliflower just as selectively as human workers would. This machine is called VitaPanther. It is being developed by researchers at the Fraunhofer Institute for Factory Operation and Automation IFF and and their colleagues at ai-solution GmbH together with five other partners: Gottfried Wilhelm Leibniz University Hannover, Steig GmbH, Beutelmann Gemüseanbau, König Sondermaschinenbau GmbH and Inokon GmbH. A prototype of VitaPanther will be finished and tested in 2017. This machine will benefit farmers in several ways: It will harvest cauliflower heads significantly faster than human pickers and could additionally work at night, too. Another plus is that farmers will be able to dispense with troublesome searches for workers. Martin Steig, farmer and CEO of Steig GmbH and and one of the future potential users of the harvester is convinced it is needed. "Farming is the last profession in which the necessary profits can only be made with a large workforce. Automation is essential for us farmers, because the minimum wage is making vegetable harvesting unfeasible. Harvesting is sustained by two components: the availability of a seasonal workforce and the pay. A shift in one of these components jeopardizes the structure. The demand for technology is thus very great." How can a machine detect the vegetable's ripeness without seeing its "whiteness", without weighing it, without knowing its size? These are the questions on which the researchers at the Fraunhofer IFF are working. They are researching and developing the necessary sensor systems along with the software that analyzes and preprocesses the data obtained so that the machine receives clear information on whether to harvest or to wait. "We are taking advantage of an effect we discovered in preliminary tests: The leaves of ripe cauliflower have a different biochemical composition than those covering unripe heads," explains Prof. Udo Seiffert, Manager of the Biosystems Engineering Expert Group at the Fraunhofer IFF. Hyperspectral cameras mounted on the harvester scan the heads of cauliflower. Whereas a conventional camera only works with visible light and produces a color picture consisting of red, green and blue tones, a hyperspectral camera scans a defined range of wavelengths beyond human vision and also encompassing infrared and ultraviolet light. Applying a mathematical model, the researchers can determine the biochemical composition of the leaves and thus the ripeness of the cauliflower based on the intensity of the light reflected in the different wavelengths scanned. The researchers are not analyzing the exact biochemical composition of the leaves, however, because the machine is only supposed to receive a yes-no command to harvest. The mathematical model that decodes the camera images into exactly this command is based on algorithms that originated with machine learning. The researchers are using examples to teach it. They "show" the camera different heads of cauliflower, which are simultaneously being inspected by a human expert. Following such a teaching phase, the system is able to decide autonomously which cauliflower should be harvested or not, even when the heads of cauliflower are unfamiliar. While the researchers from the Fraunhofer IFF are attending to the sensor systems and data analysis, their colleagues from ai-solution GmbH in Wolfsburg are working on the harvester unit that will be harvesting cauliflower heads in the future. They are building upon their asparagus harvester "Spargelpanther" for this. "We also intend to use this this asparagus harvester for other vegetables – cauliflower and head and leaf lettuce. Then, other harvester modules for other vegetables could be added in the future," says Christian Bornstein, CEO of ai-solution GmbH. "Our goal is to build a module that can be adapted to the existing unit." Farmers will only have to purchase one "vegetable harvester" in the future. Explore further: Hybrid energy harvester generates electricity from vibrations and sunlight


News Article | April 1, 2016
Site: phys.org

Logistics centers are chaotic places: forklifts, industrial trucks, motorized carts, also called ants in German, convey loads from point A to point B. Workers have been using a control panel with five to ten buttons to steer such carts. Since fully loaded carts can weigh as much as 500 kilograms, improper handling can often result in serious accidents. The tactile handles being developed by researchers at the Fraunhofer Institute for Factory Operation and Automation IFF in Magdeburg will make the operation of motorized carts more intuitive in the future. "Users will steer carts merely with hand pressure," explains Prof. Klaus Richter, expert group manager at the Fraunhofer IFF. "Whereas steering used to require effort, our handle has a kind of power steering." This means that workers can get a cart moving in the right direction with very little effort. Pressure sensors integrated in the handle make this possible. Since the handles are equipped with sensors for both hands, a cart does more than just detect whether it is being pushed or pulled. Software that compares right and left hand pressure enables a cart to recognize the particular direction specified by the user as well. The researchers are presently working to ascertain how many sensors are needed to steer a cart easily – the prototype has four sensors for the time being. "We intend to minimize the components in order to keep the price down," explains Richter. Commands given to order picking carts by workers by hand pressure are transmitted to motors like those used for electric bikes. The motor is able to execute the commands in a few milliseconds. This would be faster than an operator could handle, though. "We are developing the system to reach top speed and then we introduce artificial delays," reveals Richter. Psychological tests with subjects will pinpoint how long delays have to be to maximize user safety and ease of operation. "At the moment, we are working with our colleagues from Cloud & Heat on transmitting every worker command to a cloud where the commands will be collected and coordinated," says Richter. If a cart being taken around a blind curve is at risk of colliding with another cart, both carts will be stopped automatically – just like the vision of vehicle to vehicle communication. The researchers have already attained a latency of 10 milliseconds, i.e. the signal needs only 10 milliseconds to travel from the tactile handle to the cloud and back to the motor control unit. The researchers will be presenting their development at Hannover Messe (Halle 17, Booth C18) from April 25 to 29, 2016. Visitors will be able to test the power steering system right there, using real handles to push a virtual cart. They will also be able to vary the parameters of the order picking cart – whether these be the weight of the load or the response time with which the cart executes entered commands. The tactile handle is being developed in the FAST Realtime research project, which is primarily being advanced by the mobile communications industry and is part of the 2020 Cluster Strategy. The project with total funding of € 50 million is intended to develop new human-machine interfaces. Whereas the purpose of the "Internet of Things" is to interconnect individual objects, the "Tactile Internet" is about defining interfaces to humans, which make objects more effective and more intuitive as well as safer. Objects will be rendered smart and will communicate with each other in the "Internet of Things". This vision will be broadened in the "Tactile Internet", which will integrate humans and their behavior in real time. In other words, human-machine interfaces will be tangible and respond simultaneously, thus making them safer, more effective and more intuitive. This is why very low response times are a prerequisite for the Tactile Internet. Explore further: A tactile glove provides subtle guidance to objects in the vicinity


Heinke D.,University of Birmingham | Backhaus A.,Fraunhofer Institute for Factory Operation and Automation
Cognitive Computation | Year: 2011

In earlier work, we developed the Selective Attention for Identification Model (SAIM [16]). SAIM models the human ability to perform translation-invariant object identification in multiple object scenes. SAIM suggests that central for this ability is an interaction between parallel competitive processes in a selection stage and a object identification stage. In this paper, we applied the model to visual search experiments involving simple lines and letters. We presented successful simulation results for asymmetric and symmetric searches and for the influence of background line orientations. Search asymmetry refers to changes in search performance when the roles of target item and non-target item (distractor) are swapped. In line with other models of visual search, the results suggest that a large part of the empirical evidence can be explained by competitive processes in the brain, which are modulated by the similarity between target and distractor. The simulations also suggest that another important factor is the feature properties of distractors. Finally, the simulations indicate that search asymmetries can be the outcome of interactions between top-down (knowledge about search items) and bottom-up (feature of search items) processing. This interaction in VS-SAIM is dominated by a novel mechanism, the knowledge-based on-centre-off-surround receptive field. This receptive field is reminiscent of the classical receptive fields but the exact shape is modulated by both, top-down and bottom-up processes. The paper discusses supporting evidence for the existence of this novel concept. © 2010 The Author(s).


Naumann A.,Fraunhofer Institute for Factory Operation and Automation | Bielchev I.,Otto Von Guericke University of Magdeburg | Voropai N.,Irkutsk State University | Styczynski Z.,Otto Von Guericke University of Magdeburg
Control Engineering Practice | Year: 2014

An overview of basic IEC standards for smart grid applications is given and some examples of feasible information and communication technology for smart energy systems are shown. As ICT key standards for power grid automation, the two core standards IEC 61850 and IEC 61970 are presented in the paper. Protection automation relying on smart grid ICT technology is shown, and the hurdles to be overcome for the realization of smart grid automation are discussed. Practical examples for are demonstrated. One approach of making different standards work together is presented, which today is still not sufficiently solved and is a main hurdle on the way towards a seamless smart grid automation system. © 2013 Elsevier Ltd.


Richter K.,Fraunhofer Institute for Factory Operation and Automation | Poenicke O.,Fraunhofer Institute for Factory Operation and Automation
Communications in Computer and Information Science | Year: 2012

The paper is giving a brief overview on the Saxony-Anhalt Galileo Test Bed as an integrated environment of several real labs and test beds for research and development in the fields of logistics and transportation. As the Fraunhofer IFF is one partner of the test bed consortium, specific research topics in the application field of telematics and logistics are addressed as well. These developments are all closely connected to industrial application as the Saxony-Anhalt Galileo Test Bed offers the platform for testing the technical integration of, for instance, video- and radio-based identification and localization systems into logistics process environments. © 2012 Springer-Verlag.


Komarnicki P.,Fraunhofer Institute for Factory Operation and Automation
Archives of Electrical Engineering | Year: 2016

Current power grid and market development, characterized by large growth of distributed energy sources in recent years, especially in Europa, are according energy storage systems an increasingly larger field of implementation. Existing storage technologies, e.g. pumped-storage power plants, have to be upgraded and extended by new but not yet commercially viable technologies (e.g. batteries or adiabatic compressed air energy storage) that meet expected demands. Optimal sizing of storage systems and technically and economically optimal operating strategies are the major challenges to the integration of such systems in the future smart grid. This paper surveys firstly the literature on the latest niche applications. Then, potential new use case and operating scenarios for energy storage systems in smart grids, which have been field tested, are presented and discussed and subsequently assessed technically and economically. © Polish Academy of Sciences 2016.


Backhaus A.,Fraunhofer Institute for Factory Operation and Automation | Seiffert U.,Fraunhofer Institute for Factory Operation and Automation
Neurocomputing | Year: 2014

Against the background of classification in data mining tasks typically various aspects of accuracy, and often also of model size are considered so far. The aspect of interpretability is just beginning to gain general attention. This paper evaluates all three of these aspects within the context of several computational intelligence based paradigms for high-dimensional spectral classification of data acquired by hyperspectral imaging and Raman spectroscopy. It is focused on state-of-the-art paradigms of a number of different concepts, such as prototype based, kernel based, and support vector based approaches. Since the application point of view is emphasized, three real-world datasets are the basis of the presented study. © 2013 Elsevier B.V.


Seiffert U.,Fraunhofer Institute for Factory Operation and Automation
Neurocomputing | Year: 2014

Data and especially image compression is becoming increasingly important for efficient resource utilization. Many digital image file formats therefore include universally usable compression methods. They treat every image separately and do not profit from a larger image data set's similar image contents, which are present in numerous biomedical applications. This situation provided the impetus to develop and implement a technical system that incorporates a priori information on typical image contents in image compression on the basis of artificial neural networks and thus increases compression performance for larger image data sets with frequently recurring image contents. © 2013 Elsevier B.V.


Blumel E.,Fraunhofer Institute for Factory Operation and Automation
Procedia Computer Science | Year: 2013

Effective applied research is based on close collaboration between research and industry, which, taking the findings of basic research on customer demands as its starting point, creates new means to develop and market innovative products. What is more, growing demands for innovative and sustainable results of research and development are prompting the examination of global trends such as demographic change, growing megacities, rising energy consumption and increasing traffic and the resultant social challenges. These trends and increasing traffic in particular are giving rise to new fields of work, especially for digital technologies, as a social responsibility, e.g. on driver assistance and traffic control systems that increase safety. The social challenges are increasingly affecting markets and requiring new innovative products, efficient production processes and integrative forms of human resource development and training and qualification. The virtualization and digitization of objects and processes is becoming an enabler of the development of new strategies and concepts such as smart cities, green energy, electric vehicle networks, smart manufacturing and smart logistics. This paper examines means by which digital engineering and virtual and augmented reality technologies can support the creation of sustainable smart manufacturing and smart logistics processes as well as on-the-job training and qualification and knowledge transfer. © 2013 The Authors. Published by Elsevier B.V.


Backhaus A.,Fraunhofer Institute for Factory Operation and Automation | Seiffert U.,Fraunhofer Institute for Factory Operation and Automation
Proceedings of the 2013 IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013 | Year: 2013

Classically, machine learning methods are evaluated according to their accuracy and model size. Increasingly model parameters are used to interpret the model in order to extract information about the data it was build on. The capability of a model to deliver this kind of information, its interpretability, is so far more or less subjective. In this paper a number of quantitative measures are suggested to compare machine learning methods in their capability to offer interpretation of the underlying data. © 2013 IEEE.

Loading Fraunhofer Institute for Factory Operation and Automation collaborators
Loading Fraunhofer Institute for Factory Operation and Automation collaborators