Artificial Intelligence

The human being has the ability to adapt to physical changes in the production and logistic environment. The aim of this research project is to equip series transport vehicle with additional intelligent technique so that the vehicle becomes an automated guided vehicle (AGV) and to reduce the commissioning effort for AGVs. The user can control the AGV by voice and gesture so the AGV can do the work automatically. To realize such an interactive system it is important to develop decentralized controllers for the AGV. Thus, it is possible that the AGV is able to flexibly adapt his own behaviour.

The major components of a machine vision system are image acquisition and image processing. The complexity of image acquisition is based on the huge number of degrees of freedom. This article introduced an extension of a sensor simulation tool which enables a user to simulate the current sensor positions and field of views. The results are registered point clouds from different sensor positions merged to a final object point cloud. Now it is possible to evaluate image processing techniques based on the simulated 3D data, without explicit results from the real sensor behaviour. The benefits of the tool were demonstrated in a real micro cold forming scenario. Future work focuses on additional sensor noise models. Adding these models results in a more realistic sensor simulation framework.

Today sensors and actuators gain increased importance in industrial production processes. Traditionally centralized processing architectures are used for sensor data processing and actuator control. Increasing sensor and actuator densities require new decentralized and distributed processing methods and architectures. Artificial Intelligence, part of computer science, can contribute to robustness and autonomy aspects concerning data processing and distribution.

Surface inspection and defect detection during manufacturing is a fundamental task for industrial image processing. Based on reliable and accurate defect detection systems changing process behaviour, like blunt tools, can be detected. In many areas like Si-wafer production, optical quality control is still state of the art. Nevertheless, in many cases a completely new and specialised image processing technique will be developed for every new problem. This contribution demonstrates a method for a general image processing technique for industrial surface inspection which is not limited to a specific hardware or defect type. The system is part of the Collaborative Research Center (CRC) 747 “Micro Cold Forming” and will be evaluated in a real cold forming process.

The rise of the complexity of products and the product development process in addition to shorter development cycles and the customers’ high demands on quality, increase the pressure on producing as accurately and economically as possible. Hence, it is necessary to avoid errors in the early phases of development instead of fixing them at a later stage, which can be costly. This article uses the product configuration of milling robot units to show an innovative solution for the early development phase. This approach uses established technologies from the area of semantic technologies and problem solving methods. It also shows how to improve the reliability of product configuration by making use of the expertise from later phases in the development.

A process integrated approach for the quality management of metal working fluids in metal working processes The state of the art of the quality management for metal working fluids (MWF) in metal working processes is an offline measurement of the stability or particle size distribution of the emulsion. A German-Brazilian research project within the BRAGECRIM initiative (Brazilian - German Collaborative Research Initiative on Manufacturing Technology) heads for the in-line measurement of the quality and stability of MWF by measurement and evaluation of turbidity spectroscopy. The resulting MWF-Emulsion-Monitor should increase the economical and ecological balance of metal working processes.

There are numerous and relevant application domains like inspection, maintenance, surveillance, and handling of non-uniform goods where the use of robots would be beneficial. This article addresses the reasons that make it difficult for a present-day robot to master these domains. The dual arm system AILA will be presented, which takes further steps in the direction of such a use case.

Product piracy has risen to a serious problem for almost every manufacturer in the capital goods industry. Especially if the customer unwittingly mistake counterparts for original products, the manufacturer will lose parts of reputation. So manufacturers have to expand their quality management with new tasks to ensure that their original products are delivered to their customers. A holistic anti-counterfeiting-strategy provides adequate counter measures for this challenge.

Dynamics and complexity of today`s production systems bring established approaches for production planning and control to their limits. Accordingly, developing new concepts and methods is a key point for research in this area. The combination of a decentralized control structure and innovative methods from the field of artificial intelligence seems promising here. Open source tools have proven their applicability to implement those methods. They are disposable and can be flexibly adapted to many problems. This contribution introduces an approach for the decentralized control of a shop floor. Here, artificial neuronal networks are used as adaptive control instruments. The simulation of these networks is performed with the open source tool Stuttgart Neural Network Simulator (SNNS) and its successor Java Neural Network Simulator (JNNS).

Due to the increasing globalization of commodity flow a growth of mass good transportation is perceptible. Therefore automatic unloading of goods and their automatic transfer to logistic systems is one of the technical challenges. Opportunities arising from automatic goods unloading are explored in [1]. A system for unloading cubic goods automatically was developed and its introduction to market started last year. Changing environments and various operational areas are the motivation for research on a cognitive system for unloading goods. Hence, a concept for a cognitive system for unloading containers is introduced in this article. The components for such a system are image processing, robot control handling system and kinematics. For each of these components cognitive methods and technologies are proposed to be integrated in an overall system.