Robotic welding system

The developers of the Trans Steel Robotics system by Fronius have focused entirely on robotic steel welding. The welding uses the digital Trans Steel inverter. They can stop and start the arc via an automated interface and transmit analog or digital signals via CANopen or Devicenet. At the touch of a button, the Easy Job function conveniently saves up to five parameter data records. Fronius is an Austrian company, which has also production facilities in Czech Republic and Ukraine. The company is active in the fields of battery charging systems, welding technology, and solar electronics. www.fronius.com

The fishs fin: Bionic concepts in Automation

t the Hanover Trade Fair 2010, Festo presented some highlights from the Bionic Learning Network. The shown bionic concepts were inspired by nature. With the BionicTripod 2.0 and Modular Lightweight Handling, the company introduced solutions for a wide variety of handling tasks. It could be used with the CAN-connectable CMXR robot controller.

Due to its lightweight design, BionicTripod is efficient in terms of energy consumption in sorting and displacement operation. The bionic principle of the fishs tail fin finds repeated use as a so-called 3D Fin Ray structure adapted to

three-dimensional space. This structure makes for an energy-efficient lightweight design. The tripod is rotated at 90 from the conventional configuration and can thus be described as a horizontally arranged tripod. The integrated adaptive gripper, the so-called FinGripper, and a kind of ball joint based on the function of the human wrist provide maximum flexibility for picking up objects from a working surface and depositing them at different levels. The adaptive FinGripper constitutes the interface between the object and the actuator system. It consists of a pneumatic actuator in the form of a bellows and three gripping fingers, which are designed as an adaptive structure with Fin Ray effect. The resulting pliability and flexibility make the gripper ideal for tasks in human-machine interaction, e.g. in agriculture for sorting fruit and vegetables, or in sorting materials for recycling in industrial application.

A third hand that can pass a screwdriver, a wrench or a component for installation is also an ideal helper for all kinds of assembly tasks. EGC electric linear axis units and EMMS electric drive ensure precise control and displacement of the tripod structure. The entire system is controlled by CMXR robotic control software. In the hand axis by analogy with the function of the human wrist three actuators are arranged around a ball joint in such a way that three further degrees of freedom are attained, thereby allowing deflection by up to 30 With Modular Lightweight . Handling a light-design articulated-arm kinematical system can be built up. Weighing only 4 kg, and with a payload of up to 800 g and six degrees of freedom, it is ideally suited as a learning system. Thanks to its simple, cost-efficient design, it can be used in pick and-place applications, in the automation of laboratory processes or as a third hand in the field of service robotics. Its low weight allows the use of kinematics without a protective cage, thus paving the way for interactive human-machine cooperation. Modular Lightweight Handling is as changeable as a chameleon. From single-axis to six-axis kinematics, all variants can be modified or even subsequently retrofitted with a minimum of effort. This yields a variety of possible configurations. The principle of a modular, lightweight construction kit is continued at the front end, for example with the gripping tool interface. Modular Lightweight Handling autonomously selects two- and three-finger grippers, parallel and angle grippers, vacuum grippers and Bernoulli grippers from its storage unit. Adaptive gripper fingers based on the Fin Ray Effect also allow complex, irregularly shaped products to be grasped. Through the use of actuators with precision gearing, Modular Lightweight Handling can position objects precisely. Due to de-centralized control units in the joints, all the motors operate at low voltage and are addressed by a CMXR robot control unit via a CAN-based network.

www.festo.com

************************

RTC-CANopen
RTC-CANopen has following key features:
Lightweight: Robustness: Real-time: Reusability: Flexibility: Implements light RT middleware providing only necessary functions for embedded devices. Improves system robustness introducing monitoring methods for operations on native buses. Implement fast and reliable communication between components with the aid of advantages of native buses. Improves reusability of each element, since both hardware and software can be dealt as components. It is possible to change system configurations flexibly, switching parts or changing combinations.

RTC-CANopen is a device profile for RT (robotic technology) components, which consists of modularized robotics functional elements, to run on a CANopen environment. Traditionally robots often were called industrial robots, mainly used in production and manufacturing. But in a changed social environment with advanced robotics technologies, practical applications of servicerobots, which perform their operations nearer to human beings, will become acceptable. Non-manufacturing business such as medical and healthcare fields, and personal fields such as home automation are the targets for such service robots. However, environment for human life is very complicated, so service robots need to recognize various objects stably and respond accordingly to dynamically changing environments or situations in order to work at places near humans. Furthermore, high safety is also required not to cause adjusting people harm. Required functions will be expected to be of great variety, since targeted places and users range much wider. Because of the background described above, it is considered that traditional development methods to make out complete robots individually will meet a limitation when applied for service robots. Therefore a new approach for robotics development has been proposed. In this approach, robots are divided into various robotics elements as components and those elements are combined to be a complete robotics system.

RT middleware is also being developed as a platform on which RT functional elements (RT components) run. Since RT middleware enables reuse and switch of RT components, it becomes possible to respond to various requirements from users by combination of RT components. And it also becomes possible to develop parts individually and make development projects smaller. In the result, each vendor can concentrate on those parts they are good at. It is defined that an RT component is a robotics component, which provides some meaningful functionality. Thus not only devices such as servomotors, various sensors, cameras, and so on are called RT components, but also combinations of these devices such as motor-driven carts and arms. In addition, RT components are not limited to modules embedded in hardware, but software modules such as different control algorithms become RT components as well. What RT components should provide, for example interface specifications and component models, is defined in the standard specification (Robotic Technology Component Specification) at OMG (Object Management Group). OMG is an international nonprofit standardization consortium standardizing object oriented technologies. However, this standard specification defines only functionality and state transitions, which RT components should provide. Details about communication techniques are not described. RTC-CANopen is a profile for using CANopen in order to communicate among RT components. This profile describes necessary information to make existing CANopen devices work as RT components. RTC-CANopen enables mutual collaboration between various devices connected to native buses set up for CANopen, and RT components running on general-purpose networks. It becomes easy to combine them and build a robotic system. A system built on RTC-CANopen consists of DeviceRTCs running on an embedded MPU (microprocessor unit), ProxyRTCs managing DeviceRTCs on RT middleware, and RTCCANopen servers managing the whole system. ProxyRTCs and DeviceRTCs have correspondence relations. Using such relations, when a ProxyRTC connects to another ProxyRTC, their corresponding DeviceRTCs transmit data directly via a native bus. RTC-CANopen profile is defined by adding necessary information for devices to run as RT components, to the existing CANopen profile. Therefore, it is compatible with existing CANopen compliant products, and makes extension for robots easier. A set of tools is also provided, which generates robotics system definition files from definition files of CANopen (DCF files) automatically. RTC-CANopen enables improvement of robotic system development efficiency and realization of various requirements from users. robotics.omg.org www.hri.ee.shibaura-it.ac.jp

************************ Mobile robots for cooperative machine operation

omrobs (cooperating mobile robots Braunschweig) are small robots based on the platform of the robot Helios developed by the Fredt student group at the Technical University in Braunschweig (Germany). The Comrobs are 350 mm wide, 600 mm long and have 40 mm of ground clearance. Weighting 20 kg, they may transport 20 kg additionally.

The vehicles are powered by 24-V NiMH batteries allowing maximal four-hours movement at up to 20 km/h. The electronic system consists of two levels (high and low) interconnected via CAN between each other. The processing in the high level is fulfilled by the Micro-Autobox from Dspace and an Intel-based Mini-PC (1,6 GHz, 2-GiB RAM, 32-GiB SSD). The Micro-Autobox is directly connected to CAN via one of the four provided CAN interfaces, while the Mini-PC uses the USB-to-CAN adapter. The standard vehicle equipment incorporates different sensors. Sensors with a high need for computing capacity, such as a Firewire camera, are analyzed directly by the Mini-PC. A 2-D positiondetermination system by system, based on the runtime measurement of radar signals. It delivers the XY-coordinate, orientation and speed information of the vehicle. A GPS receiver by U-Blox connected via EIA-232 to the Mini-PC provides the absolute vehicle position. For the two dimensional orientation a laser scanner from Sick is mounted on the vehicle and connected via EIA-232 to the Micro-Autobox. By means of the scanner, barriers as well as relative position to other vehicles are detected. On the low level the self-developed devices are interconnected via CAN. These fulfill the control and evaluation of the steering servo-drive, wireless emergency-stop function, measurement of acceleration and velocity as well as Zigbee-communication between different Comrobs and with one remote controller. Meanwhile several of Comrobs exist at the institute as a platform to deliver students a practical insight into up-to-date technologies of sensors, positioning, communication as well as construction and layout of controlled systems. The small robots collect experiences in a simple way on the small vehicles and transfer these to large machines in the end, thus offering an interim step between simulation and the practical experiments. In ongoing projects, the Comrobs are used for development of communication structures as well as algorithms for the cooperative machine operation i.e. the interactive control of more than two machines. In agricultural applications these are, for example, the automation of the loading process between field chopper and transport units. The robotconcept provides also potential for joint research in such disciplines as robotics, automotive engineering (e.g. car-tocar communication) or aerospace technology. It is supposed to develop an aerospace application, in which the earthbound units cooperate with air units for scouting (e. g. in disaster areas).

www.tu-braunschweig.de/ilf/lehre/fredt

************************ Open framework for wheelchair-based robots

The Robochair project develops a mobile robotics open platform for wheelchair-based robots. It has been developed as add-on module for commercial wheelchairs. Its design is modular and based on open Interfaces for easy extension, robust and lowcost enough to be affordable to most of the people with disabilities.

Environment sensing and information processing are done by means of low cost sensors and small form factor single board computer (SBC). The coupling between environment constraints (detected obstacles and target to reach) and wheelchairs power module is implemented by using CANopen networks.The Robochair system is developed around the Player robotics framework and implements all the devices needed to be Player compatible. This way, third party Player developers source code could be run in Robochair. In addition, and from hardware developers point of view, Robochair defines a low-level communication interface based on CANopen, that allows the interoperability and exchangeability among third party robotics devices. As shown in Figure 1, the distributed control system has a main device for global control and configuration tasks. The other devices perform only local tasks, so the minimum necessary data will be sent to the main module. This way, the manner each local task is solved is transparent to the main module, and consequently to any high-level application running in it. When a device is connected to the system, the main module detects its features and checks what kind of services it offers to the system. Robochair uses Can festival, which is an open source implementation of the CiA 301 CANopen application layer. The sensor selection is a critical task and defines how the system could sense the environment. There are several types of sensors used to perceive the surroundings of the robot; ultrasonic, infrared and laser range finders are the most popular. There are a lot of different models with a variable list of features and cost from a huge variety of manufactures. Among the low cost ones, SRF08 ultrasonic range finders by Devantech and GP2D12 infrared range finders by Sharp were selected due to their price/features relation.

Several others were analyzed and ruled out by some undesirable features. Laser range finders were also ruled out due to their high cost. A force feedback joystick has been selected as input/output method. orce feedback or haptic refers to a technology, which interfaces the user via the sense of touch by applying forces, vibrations and motions to the user. This mechanical stimulation is used, for example, to create virtual objects, which allow the user feel their shape. In the other hand, and regarding to wheelchairs, it could also be used to inform the user about the system status in a no intrusive manner. The second device selected was a touch LCD screen in order to provide the user with graphical information. Some soft button and a virtual joystick were developed to control the navigation of the wheelchair. Software components The high-level software interfaces defined by Robochair are based in the Player framework and are called Player interfaces. Player extensively uses the software interface concept. Different drivers, that control different devices with the same functionality (or almost), must provide the same Player interface. Player defines several interfaces for the control of well-known robotics devices. As shown in the figure, Robochair uses 2-dimensional position, sonar, infrared (IR), and power Player interfaces for the remaining devices. The 2Dposition interface is used to move the robot and to retrieve its pose (the robot position and orientation according to a given reference system). Sonar and IR interfaces are used to receive distance reading from ultrasonic and infrared sensor, respectively. The power interface is used to read the battery status and the remaining capacity. For their specification refer to layer documentation. Each I/O device is controlled by a low-cost microcontroller that runs the software needed to interface the device in one hand, and the CANopen stack in the other. Each of these components embeds a CANopen NMT slave device, which handles the protocol issues regarding to the master communication. The CANopen NMT master device is running in the other side of the CAN network in the Main Control module, a 32-bit single-board computer (SBC),in which Player is also running. The Robochair Player driver, which runs within the Player server, allows applications (Player clients) access the hardware in a

Fig. 2: In the Robochair project several sensors are used to detect obstacles; they are connected to the CANopen network by means of I/O modules

transparent manner through 2D-position, sonar and IR interfaces. Software platforms, toolkits and APIs The Player/Stage project creates Free Software that enables research in robot and sensor systems. The Player robot server is probably the most widely used robot control interface in the world. Its simulation backends; Stage and Gazebo, are also very widely used. Released under the GNU General Public License, all code from the Player/Stage project is free to use, distribute and modify. Player/Stage is developed by an international team of robotics researchers and used at labs around the world. Player provides a network interface to a variety of robot and sensor hardware. Players client/server model allows robot control programs to be written in any programming language and to run on any computer with a network connection to the robot. Player supports multiple concurrent client connections to devices, creating new possibilities for distributed and collaborative sensing and control. Player supports a wide variety of mobile robots and accessories. Stage simulates a population of mobile robots moving in and sensing a two-dimensional bitmapped environment. Various sensor models are provided, including sonar, scanning laser rangefinder, pan-tiltzoom camera with color blob detection and odometer function. Stage devices present a standard Player interface so few or no changes are required to move between simulation and hardware. Many controllers designed in Stage have been demonstrated to work on real robots. Gazebo is a multi-robot simulator for outdoor environments. Like Stage, it is capable of simulating a population of robots, sensors and objects, but does so in a three-dimensional world. It generates both realistic sensor feedback and physically plausible interactions between objects (it includes an accurate simulation of rigid-body physics). Gazebo presents a standard Player interface in addition to its own native interface. Controllers written for the Stage simulator can generally be used with Gazebo without modification (and vise-versa). Gazebo runs under Linux and other operating systems. www.irobotics.org .

************************