Since the beginning of mobile robots the developers considered safety aspects. The challenge was - and still is - to identify hazards and to find appropriate manners to react on them. This paper presents aspects concerning the safety of wall-climbing robots and discusses different types of hazards (especially those which could cause a drop-off), their impact and how they can be handled. The main challenge is to perform a safety analysis without neither sufficient knowledge about the system nor significant statistical data for evaluation. Here, a fault tree analysis of a wall-climbing robot is performed to identify bottle necks related to safety and to specify requirements to the robot's hard- and software. Finally, some suitable measures are highlighted which improve its safety tremendously.
This work suggests a new approach of robot system to shelve and retrieve imprecisely placed objects with the help of vision and force feedback according to their alphabetic/numeric codification system. A practical example of shelving and retrieving tasks is automation of libraries. Previously, automation of libraries was based on the complicated transport mechanisms and also on robotic systems without using the vision or force/torque system. Hence, the system will bring additional drawbacks which can be eliminated by using vision system; these drawbacks cause the waste of energy and need more maintenance. In this work the proposed vision system detects objects, position/orientation, characterizes and classifies the objects, identify the codes assigned to objects (SIFT features), and provides automatically decisions which define the art of the control (when vision or force feedback should be used). The proposed control can perform position control only or combination of force-vision control depending on the conditions of the task and the environment.
This paper introduces the hard- and software of an educational mobile robot as well as exemplary parts of the framework design to discuss the requirements of such a system. The framework design relies heavily on abstraction and encapsulation while featuring Design Pattern to allow an easy access to the system and to keep it extensible. The system itself is in use since 2009 and has received positive feedback from students/users who have worked with it. This includes feedback on hand-on exercises and tutorials as well as larger projects. Exemplary results of these projects are presented in this paper. Topic: educational robots and tools Keywords: educational mobile robot, framework, abstraction
Mobile manipulators are of high interest to industry because of the increased flexibility and effectiveness they offer. The combination and coordination of the mobility provided by a mobile platform and of the manipulation capabilities provided by a robot arm leads to complex analytical problems for research. These problems can be studied very well on the KUKA youBot, a mobile manipulator designed for education and research applications. Issues still open in research include solving the inverse kinematics problem for the unified kinematics of the mobile manipulator, including handling the kinematic redundancy introduced by the holonomic platform of the KUKA youBot. As the KUKA youBot arm has only 5 degrees of freedom, a unified platform and manipulator system is needed to compensate for the missing degree of freedom. We present the KUKA youBot as an 8 degree of freedom serial kinematic chain, suggest appropriate redundancy parameters, and solve the inverse kinematics for the 8 degrees of freedom. This enables us to perform manipulation tasks more efficiently. We discuss implementation issues, present example applications and some preliminary experimental evaluation along with discussion about redundancies.
We present the Kinematic Continuous Collision Detection Library (KCCD), a new library for real-time continuous collision detection for humanoid and industrial robots. The C++ library operates on joint state intervals. Such, it does not only test a single or N intermediate configurations but assures safety of a whole movement. The library uses sphere swept convex hulls (SSCH) as its volume representation, computes swept volumes and their distances for all body pairs of a robot, and provides an operation set that allows for a trade-off between accuracy and computation time. The paper gives an overview of the applied algorithm and describes basic features of the library, additionally provided configuration and visualization tools, as well as existing applications.
Fast manipulation of unknown objects is a basic ability on the way to the robotic co-worker, because it is mostly impractical to have a model of each object in the environment. This paper addresses the problem of grasping and delivering unknown objects by an industrial robot from a table using one single camera image. The delivery planning for each object is done with respect to the packing problem, with the additional constraint that the objects must be regraspable after their delivery. The aim is to build up a system, which has for example the ability to tidy up a table. For this aim, we present two planning algorithms, one for the grasp planning and one of the delivery planning. The performance and robustness of the system is validated by test runs with several objects.
In this paper, a new automated optimum path and trajectory generation system for robotic painting process is presented. Usually, a direct self-learning manual method is usually adopted, i.e. the operator drives the robot manually through a complete spraying cycle. Recently, in order to-avoid this time-consuming procedure, CAD-based or acquire-comparerecognize methods have been developed. Here, a general technique that avoids the need for manual programming or CAD drawings and allows to obtain an optimal path and trajectory by using the graph theory and operative research methods is developed. After an image acquisition phase, a partitioner splits the object into a set of primitives, then the proposed algorithm is run on the graph in order to generate the optimal path; then, a optimum trajectory planning phase is performed. Kinematic and dynamic simulators of the CMA Robotics robots have been developed and validated in order to allow a correct planning and evaluation of the results. The new path planning algorithm has been implemented in Matlab and Visual-Studio.NET environments. Practical applications are presented.
Three-dimensional realtime characterization of the robot??s workspace offers good perspectives for efficient motion planning. In our approach, the robot??s environment is observed by Photonic Mixer Device cameras, providing depth information for each pixel. On that basis a 3-dimensional environment model is generated and optimized for path planning and obstacle avoidance. The path is continuously adapted to changes in the workspace to generate a smooth and safe motion towards the target position. All static and dynamic obstacles on its path are automatically avoided, which is a step towards the challenging objective to enable safe presence of humans in the work cell. Key technology and system design aspects are presented as well as a first characterisation of the achievable performance. Keywords: path planning, obstacle avoidance, human-robot-cooperation, workspace sensing, 3D-cameras
Vertical Take-Off and Landing (VTOL) capable Mini Unmanned Aerial Vehicles (MUAV) are an economic alternative to real helicopters if local aerial pictures or other aerial measurements are needed. With electronic stabilization, GPS based position-hold and autonomous way-point functions these systems are easy to operate. While autonomous take off and flying are done routinely, this is not true for autonomous landing. GPS based position stabilization, in conjunction with the normally used cheap and small IMUs, is not accurate enough to avoid dangerous moves in the very last moment of the landing operation. While sufficient for an emergency landing this is not suitable as routine operation. In this paper a solution is presented which uses optical flow techniques for position stabilization in the moment just before the touch down to avoid the inaccuracies from the GPS at this critical moment.
We introduce an approach of multi-view based surveillance of human/robot workspaces. Unknown objects within a robot workcell as e.g. humans are to be detected to avoid accidents with the robot. Therefore, two aspects are to be discussed: First, the reconstruction of regions within the robot workcell which are potentially occupied by unknown objects based on sensor data and additional plausibility criterions. Second, since the quality of reconstruction mainly depends on the placement of the sensors, objective functions and optimal values for this purpose are presented.
We propose a simple but efficient control strategy to manipulate objects of unknown shape, weight, and friction properties – prerequisites which are necessary for classical offline grasping and manipulation methods. With this strategy, the object can be manipulated in hand in a large scale,(eg. to rotate the object 360 degree) regardless whether there is rolling or sliding motion between the fingertips and object. The proposed control strategy employs estimated contact point locations, which can be obtained from modern tactile sensors with good spatial resolution. The feasibility of the strategy is proven in simulation experiments employing a physics engine providing exact contact information. However, to motivate the applicability in real world scenarios, where only coarse and noisy contact information will be available, we also evaluated the performance of the approach when adding artificial noise.
This paper presents a method for constructing high fidelity geometrical models of objects in a point cloud scene with the goal of grasp planning and evaluation. We consider objects whose volume can be modeled as a solid of revolution which is extracted from the scene in a sample consensus framework. The approach also provides a natural strategy for segmenting object features such has cup-handles that are an exception to this shape assumption.
During the last years, in laser material processing an enormous leap in development has happened due to various innovation boosts coming from laser technology. New and further developments in systems engineering, laser beam sources and laser beam processes have enlarged the field of laser material processing and brought along many new areas of application such as scanner welding, for instance. The newest generation of laser beam sources, such as the disc laser, allow through their outstanding laser beam quality for applications with very high processing speeds and minimal downtimes. Through the increasing processing speeds, also the generally high requirements concerning the construction of components, the welding bevels, the clamping method and the processing sensor system grow.
How can robots effectively assist humans in the packaging industry, when it comes to handling large components, such as a LCD television? This is the question that scientists and industry partners of the research project CustomPacker [1] are seeking to answer. The aims of the project are to free the human worker of tedious, unchallenging work, whilst simultaneously reducing packaging costs.
Recent research progress enables mobile ground robots and UAVs to create maps of larger and less restricted environments. However, since these maps build upon landmarks designed for computers, like image keypoints, the benefit for human operators is limited. To close this gap, we present mapping results based on automatic extraction and matching of image features that consider the natural perceptual experience of human operators. The extraction mechanism is borrowed from the early human visual system and finds image areas (called proto-objects) that are likely to provoke bottom-up visual attention of a human observer. This work builds upon the neuroscientific Itti-Koch model. We introduce a new approximation of the normalization procedure of this model to achieve real-time performance. For proto-object matching, we propose a combined descriptor of color, texture and intermediate results of the saliency computation process. To evaluate the new landmarks, results in the areas of object-recognition and simultaneous localization and mapping are presented. Keywords Mobile Robotics, Mapping, Saliency, Proto-Objects, Visual Landmarks, SLAM
Motion in our macroscopic world is ruled by mass-related, volumetric forces, mostly provoked by inertia and gravity. In minor dimensions, surface forces gain growing importance. Downscaling of robots thus offers new opportunities for design. Volumetric forces minimized by light-weight construction combined with maximized forces in the contact area to the environment using micro-scale phenomena allow the realization of new functions in meso scale, which are not accessible by standard macroscopic construction approaches. To illustrate this strategy, we introduce as well an example for the construction of a machine element (an adhesion module usable as a gripper, on the one hand for manipulation, on the other for locomotion) as that of a complete climbing robot.
In this paper we present our approach to network robotics using a flying robot. The robot is an autonomous multicopter equipped with a computational module and a wireless mesh node. We also present the wireless environment in which algorithms are developed and evaluated. We argue that sensorimotor interaction, i.e. shaping the sensory inputs by specific movement, is able to cope with the dynamics of wireless networks. Additionally, we present prototypical experiments on navigation strategies in the wireless environment. Keywords: Network Robotics, Wireless Network, UAV, Multicopter
Modular satellite systems offer significant benefits for future space activities: They facilitate the commercial production of satellite components and improve life expectancy by providing an easy way to replace individual modules in orbit or even to reconfigure a system completely to suit a new mission. In this paper, we discuss a new concept for a completely modular satellite system based on reusable building blocks with standardised hardware and software interfaces. We outline the on-board IT infrastructure requirements and present software components which detect the satellite system??s physical configuration and collect status reports from all satellite modules during runtime. Finally, we explore scenarios for servicing and reconfiguring such modular satellite systems in orbit using robotic manipulators.
The contribution proposes a control architecture, which enables a multi-link-flexible robot arm under gravitational influence to catch multiple balls sequentially thrown by a human. A net at the end-effector is utilized to intercept the balls when they pass the vertically oriented robot plane of motion. The ball detection, tracking as well as the prediction of the ball intercept location is based on a wall-mounted Kinect RGB-D sensor. Previously caught balls represent a varying payload, which induces also dynamic disturbances due to the pendulum motion. These disturbances as well as the coupled flexible-link vibrations are damped with a model free independent joint controller. The inverse kinematics approach is based on neural networks and augmented by an online payload estimation.
In vision data processing often the positions of detected objects have to be determined. This can be done by many sophisticated algorithms using several characteristic optical properties of each object. The position of each object can then be fed into additional state estimation algorithms. Thus it is possible to use multiple successive frames to calculate a precise estimation of the state of each object containing position, orientation, velocity and its trajectory. If however many objects are detected but all objects look the same (no characteristic optical properties can be determined), tracking of objects becomes more difficult. The main problem that arises is to assign the detected object positions to the predicted position estimations which can be calculated from the state estimation for each object. In literature, this topic is referred to as multi target tracking. This paper describes a simple solution for a multi target tracking problem that arises, if multiple small scale transportation vehicles are operated in a camera monitored area. While the state estimation for each vehicle and obstacle is performed by a Kalman filter algorithm, the assignment of the detected object positions to the position predictions of all objects is done by the Hungarian method with some extension to avoid assignments of objects that are located too far apart from each other.
