Skip to main content

Über dieses Buch

The two volume set LNAI 7101 and 7102 constitute the refereed proceedings of the 4th International Conference on Intelligent Robotics and Applications, ICIRA 2011, held in Aachen, Germany, in November 2011.

The 122 revised full papers presented were thoroughly reviewed and selected from numerous submissions. They are organized in topical sections on progress in indoor UAV, robotics intelligence, industrial robots, rehabilitation robotics, mechanisms and their applications, multi robot systems, robot mechanism and design, parallel kinematics, parallel kinematics machines and parallel robotics, handling and manipulation, tangibility in human-machine interaction, navigation and localization of mobile robot, a body for the brain: embodied intelligence in bio-inspired robotics, intelligent visual systems, self-optimising production systems, computational intelligence, robot control systems, human-robot interaction, manipulators and applications, stability, dynamics and interpolation, evolutionary robotics, bio-inspired robotics, and image-processing applications.



A Body for the Brain: Embodied Intelligence in Bio-inspired Robotics

Biomechatronics for Embodied Intelligence of an Insectoid Robot

In this paper, the design and development of the new hexapod robot HECTOR is described. To benefit from bio-inspired control approaches for walking, it is fundamental to identify the most important morphological and biomechanical aspects and to associate them with biological control approaches whose function principles rely on those special body features. In a second step, these pairs can be transferred to the robot to lay the foundation for embodied intelligence. According to this idea, the main characteristics of HECTOR as presented here are the muscle-like elasticity in the self-contained joint drives with integrated sensor processing capabilities, actuated 2D body segment drives, the layout and orientations of the legs and joint-axes and a lean bus system for onboard communication.

Axel Schneider, Jan Paskarbeit, Mattias Schäffersmann, Josef Schmitz

Novel Approaches for Bio-inspired Mechano-Sensors

In this paper, we present novel approaches for building tactile-array sensors for use in robotic grippers inspired from biology. We start by describing the sense of touch for humans and we continue by proposing different methods to build sensors that mimic this behaviour. For the static tactile sense we describe the principles of piezoresistive materials, and continue by outlining how to build a flexible tactile-sensor array using conductive thread electrodes. An alternative sensor is further described, with conductive polymer electrodes instead. For the dynamic tactile sense, we describe the principles of PVDF piezoelectric thin films and how can they be used for sensing. The data acquisition system to process the information from the tactile arrays is covered further. We validate the proposed approaches by a number of applications: classifying a number of fruits and vegetables using only the haptic feedback during their palpation, recognizing objects based on their contact profile and detecting gentle contact and vibrations using the piezoelectric sensor. We conclude by showing what needs to be improved and addressed further to achieve human-like tactile sensing for robots.

Alin Drimus, Arne Bilberg

Helping a Bio-inspired Tactile Sensor System to Focus on the Essential

Insects use their antennae (feelers) as near-range sensors for orientation, object localization and communication. This paper presents further developments for an approach for an active tactile sensor system. This includes a hardware construction as well as a software implementation for interpreting the sensor readings. The discussed tactile sensor is able to detect an obstacle and its location. Furthermore the material properties of the obstacles are classified by application of neural networks. The focus of this paper lies in the development of a method which allows to determine automatically the part of the input data which is actually needed to fulfill the classification task. For that, non-negative matrix factorization is evaluated by quantifying the trade-off between classification accuracy and input (and network) dimension.

Sven Hellbach, Marc Otto, Volker Dürr

Robust Dataglove Mapping for Recording Human Hand Postures

We present a novel dataglove mapping technique based on parameterisable models that handle both the cross coupled sensors of the fingers and thumb, and the under-specified abduction sensors for the fingers. Our focus is on realistically reproducing the posture of the hand as a whole, rather than on accurate fingertip positions. The method proposed in this paper is a vision-free, object free, data glove mapping and calibration method that has been successfully used in robot manipulation tasks.

Jan Steffen, Jonathan Maycock, Helge Ritter

Software/Hardware Issues in Modelling Insect Brain Architecture

The concept of cognitive abilities is commonly associated to humans and animals like mammals, birds and others. Nevertheless, in the last years several research groups have intensified the studies on insects that posses a much simpler brain structure even if they are able to show interesting memory and learning capabilities. In this paper a survey on some key results obtained in a joint research activity among Engineers and Neurogeneticians is reported. They were focussed toward the design and implementation of a model of the insect brain inspired by the

Drosophila melanogaster

. Particular attention was paid to the main neural centers the Mushroom Bodies and the Central Complex. Moreover a Software/Hardware framework, where the model could be tested and evaluated by using both simulated and real robots, is described. This research activity aims at introducing an insect brain to act as a controller for very smart and sophisticated insectoid body structures, to give rise to a new generation of novel embodied intelligent machines.

Paolo Arena, Luca Patané, Pietro Savio Termini, Alessandra Vitanza, Roland Strauss

Higher Brain Centers for Intelligent Motor Control in Insects

The higher control of orientation, walking and gap climbing behavior in the fruit fly


is studied by neurogenetic means. An insect brain model is presented for the control of object approaches. The model comprises learning abilities of flies at two different time scales. A short-term orientation memory allows for the continued approach of objects that disappeared from sight. Flies can come back to the still invisible object even after a detour to a distracter object. A long-term memory allows for the storage of experience with particular types of objects in order to trigger avoidance behavior in the future instead of the default approach behavior. Moreover, we provide evidence that the highly adaptive and successful locomotion of flies relies also on short-term integrators, motor learning, body size representation and adaptive termination of behavior.

Roland Strauss, Tammo Krause, Christian Berg, Bianca Zäpf

An Insect-Inspired, Decentralized Memory for Robot Navigation

Navigation in animals is often discussed to require a ‘cognitive map’. Here we propose an artificial neural system that consists of a network allowing for both path integration and landmark guidance. This network is able to describe experiments with desert ants and honey bees, the latter eventually interpreted as to show the existence of a cognitive map. In contrast, our network represents a decentralized system containing procedural memory elements and a motivation network, but no “central control room” or “global neural workspace”. Its output can directly be used to control the forward movement of a robot.

Holk Cruse, Rüdiger Wehner

Models of Visually Guided Routes in Ants: Embodiment Simplifies Route Acquisition

It is known that ants learn long visually-guided routes through complex terrain. However, the mechanisms by which visual information is first learnt and then used to control a route direction are not well understood. In this paper we investigate whether a simple approach, involving scanning the environment and moving in the direction that appears most familiar, can provide a model of visually guided route learning in ants. The specific embodiment of an ant’s visual system means that movement and viewing direction are tightly coupled, a familiar view specifies a familiar direction of viewing and thus a familiar movement to make. We show the feasibility of our approach as a model of ant-like route acquisition by learning non-trivial routes through a simulated environment firstly using the complete set of views experienced during learning and secondly using an approximation to the distribution of these views.

Bart Baddeley, Paul Graham, Andrew Philippides, Philip Husbands

Intelligent Visual Systems

Robust Object Tracking for Resource-Limited Hardware Systems

Resource-limited hardware systems often generate LFR (low frame rate) videos in many real-world robot vision applications. Most existing approaches treat LFR video tracking as an abrupt motion tracking problem. However, in LFR video tracking applications, LFR not only causes abrupt motions, and also large appearance changes of objects because the objects’ poses and illumination may undergo large changes from one frame to the next. This adds extra difficulties to LFR video tracking. In this paper, we propose a robust and general tracking system for LFR videos. The tracking system consists of four major parts: dominant color-spatial based object representation, cross bin-ratio based similarity measure, annealed PSO (particle swarm optimization) based searching, integral image of model parameters. The first two parts are combined to provide a good solution to the appearance changes, and the abrupt motion is effectively captured by the annealed PSO based searching. Moreover, an integral image of model parameters is constructed, which provides a look-up table for evaluation, and this greatly reduces the computational load. Experimental results demonstrate that the proposed tracking system can effectively tackle the difficulties caused by LFR.

Xiaoqin Zhang, Li Zhao, Shengyong Chen, Lixin Gao

Adaptive Rank Transform for Stereo Matching

Window selection is the main challenge for local stereo matching methods based on the rank transform and it involves two aspects : the rank window selection and the match window selection. Most recent methods only focus on how to select the match window but pay little attention to the selection of the rank window. In this paper, we propose a novel matching method based on adaptive rank transform. Differing with the existing rank-based matching methods, the proposed method can deal with the rank and match window selection at the same time. The experimental results are evaluated on the Middlebury dataset as well as real images, showing that our method performs better than the recent rank-based stereo matching methods.

Ge Zhao, Yingkui Du, Yandong Tang

Real Time Vision Based Multi-person Tracking for Mobile Robotics and Intelligent Vehicles

In this paper, we present a real-time vision-based multi-person tracking system working in crowded urban environments. Our approach combines stereo visual odometry estimation, HOG pedestrian detection, and multi-hypothesis tracking-by-detection to a robust tracking framework that runs on a single laptop with a CUDA-enabled graphics card. Through shifting the expensive computations to the GPU and making extensive use of scene geometry constraints we could build up a mobile system that runs with 10Hz. We experimentally demonstrate on several challenging sequences that our approach achieves competitive tracking performance.

Dennis Mitzel, Georgios Floros, Patrick Sudowe, Benito van der Zander, Bastian Leibe

A Method for Wandering Trajectory Detection in Video Monitor

For video monitor system, wandering trajectory detection of moving targets is one of key problems needed to be solved, as it is an effective method to discover wandering behavior and prevent the potential harmful behavior. In view of the weakness of the existing algorithms,a method for wandering trajectory detection based on angle is proposed in this paper. In this method, wandering trajectory curves are divided into three kinds, which are closed curve, spiral curve and S curve. The occurring of wandering can be analyzed and judged through the angle of trajectory changing. The experimental results demonstrate that the method can detect the wandering trajectory accurately in video sequences, without any training samples. So, the method can improve the real-time performance of the monitor system.

Ruohong Huan, Zhehu Wang, Xiaomei Tang, Yun Pan

The Design of a Vision-Based Motion Performance System

This paper presents the structure of our real time vision-based motion performance system. The system requires user to wear markers with a certain color. Several novel algorithms in the system are introduced including algorithms for feature detection and feature tracking under occlusion. Feature Detection takes advantages of four properties of markers to avoid the interference from non-markers regions. Besides, we propose a simple but effective method to track these features and handle occlusion by estimating velocity of missing features based on prior, smoothness and fitness term. These algorithms are to ensure the accuracy and low computation cost of reconstruction of 3D points of the markers. At run time, the system automatically scans, identifies, tracks and finally reconstructs the markers to 3D points. We test the ability of our system by having user perform walking, running and jumping.

Cheng Ren, Shuai Ye, Xin Wang

Window Function for EEG Power Density Estimation and Its Application in SSVEP Based BCIs

A high quality power density estimation for certain frequency components in a short time is of key importance in Steady-State Visual Evoked Potentials (SSVEP) based Brain Computer Interface (BCI). In this paper, the effect of the window functions in SSVEP based BCIs is discussed. EEG signal is a typical color noise with a high energy of the low frequency component. The main findings are that (1) The spectral leakage for EEG signals has some regular patterns. An obvious oscillation with the corresponding frequency can be observed. The amplitude of the oscillation decreases with the growth of the frequency. A short analysis is also given for the leakage. (2) The leakage from the low frequency component can be effectively suppressed by the using of some windows, such as Hamming, Hann and triangle window; (3) By removing the influence of the leakage from the low frequency component with high pass filter, the classification results are mainly determined by the width of the main lobe. The rectangle window would have a better accuracy than Hamming, Hann and triangle window. Some windows constructed with a narrower main lobe width have a potential use in SSVEP based BCIs.

Gan Huang, Jianjun Meng, Dingguo Zhang, Xiangyang Zhu

Efficient Multi-resolution Plane Segmentation of 3D Point Clouds

We present an efficient multi-resolution approach to segment a 3D point cloud into planar components. In order to gain efficiency, we process large point clouds iteratively from coarse to fine 3D resolutions: At each resolution, we rapidly extract surface normals to describe surface elements (surfels). We group surfels that cannot be associated with planes from coarser resolutions into co-planar clusters with the Hough transform. We then extract connected components on these clusters and determine a best plane fit through RANSAC. Finally, we merge plane segments and refine the segmentation on the finest resolution. In experiments, we demonstrate the efficiency and quality of our method and compare it to other state-of-the-art approaches.

Bastian Oehler, Joerg Stueckler, Jochen Welle, Dirk Schulz, Sven Behnke

3D Body Pose Estimation Using an Adaptive Person Model for Articulated ICP

The perception of persons is an important capability of today’s robots that work closely together with humans. An operator may use, for example, gestures to refer to an object in the environment. In order to perceive such gestures, the robot has to estimate the body pose of the operator.

We focus on the marker-less motion capture of a human body by means of an

Iterative Closest Point

(ICP) algorithm for articulated structures. An articulated upper body model is aligned with the depth measurements of an RGB-D camera. Due to the variability of the human body, we propose an adaptive body model that is aligned within the sensor data and iteratively adjusted to the person’s body dimensions. Additionally, we preserve consistency with respect to self-collisions. Besides that, we use an inverse data assignment, that is particularly utile for articulated models.

Experiments with measurements of a Microsoft Kinect camera show the advantage of the approach compared to the standard articulated ICP algorithm in terms of the

root mean squared

(RMS) error and the number of iterations the algorithm needs to converge. In addition, we show that our consistency checks enable to recover from situations where the standard algorithm fails.

David Droeschel, Sven Behnke

Self-optimising Production Systems

Artificial Cognition in Autonomous Assembly Planning Systems

Cognition is of great interest in several scientific disciplines. The issue is to transfer human cognitive capabilities to technical systems and so generate artificial cognition. But while robots are learning to communicate or behave socially only a few examples for applications in production engineering and especially in assembly planning exist. In this field cognitive systems can achieve a technological advance by means of self-optimization and the associated autonomous adaption of the system’s behavior to external goal states. In this paper cognitive technical systems and their software architectures in general are discussed as well as several assembly planning systems. A precise autonomous assembly planning system and its implementation of cognitive capabilities is presented in detail.

Christian Buescher, Marcel Mayer, Daniel Schilberg, Sabina Jeschke

Self-optimization as an Enabler for Flexible and Reconfigurable Assembly Systems

In the face of continuously increasing cost pressure, a wide range of product versions and shorter innovation cycles, the demand for more versatile assembly and handling systems is steadily growing. Co-operating robots represent a suitable approach for this purpose. However, reconfiguring a multi-device robot cell usually involves a certain programming effort and unfavorable down times. By integrating self-optimizing functions, the complex task of reconfiguration is substantially simplified in order to make economic use not only of the referenced co-operating robotic systems. Therefore, several self-optimizing functions for different stages of production have been developed and applied to various production tasks. The implemented functions comprise self-optimizing planning and commissioning as well as a self-optimizing joining process. Based on the experience gained from these examples, the self-optimizing functions will be similarly applicable to various cases with relatively small additional effort.

Rainer Müller, Christian Brecher, Burkhard Corves, Martin Esser, Martin Riedel, Sebastian Haag, Matthias Vette

Flexible Assembly Robotics for Self-optimizing Production

This paper provides an overview of the research results on self-optimizing production systems. Self-optimization strategies developed for assembly systems will be presented focusing on the enhancement of flexibility of assembly processes through a holistic approach regarding product-process-interdependencies. Key elements of the research like automation-friendly product and process design as well as highly-flexible automation equipment and control will be pointed out. This paper then draws a conclusion from that work and derives future research topics for making self-optimizing assembly systems a technology ready to be transferred to industry. The authors identified cooperation technologies, sensor-integration and sensor-guidance as well as meta-level task specification as relevant enablers for self-optimization in assembly systems as they further increase flexibility, autonomy, and cognition – the pre-requisites for self-optimization. Concept approaches will be described.

Sebastian Haag, Nicolas Pyschny, Christian Brecher

Meta-modeling for Manufacturing Processes

Meta-modeling for manufacturing processes describes a procedure to create reduced numeric surrogates that describe cause-effect relationships between setting parameters as input and product quality variables as output for manufacturing processes. Within this method, expert knowledge, empiric data and physical process models are transformed such that machine readable, reduced models describe the behavior of the process with sufficient precision. Three phases comprising definition, generation of data and creation of the model are suggested and used iteratively to improve the model until a required model quality is reached. In manufacturing systems, such models allow the generation of starting values for setting parameters based on the manufacturing task and the requested product quality. In-process, such reduced models can be used to determine the operating point and to search for alternative setting parameters in order to optimize the objectives of the manufacturing process, the product quality. This opens up the path to self-optimization of manufacturing processes. The method is explained exemplarily at the gas metal arc welding process.

Thomas Auerbach, Marion Beckers, Guido Buchholz, Urs Eppelt, Yves-Simon Gloy, Peter Fritz, Toufik Al Khawli, Stephan Kratz, Juliane Lose, Thomas Molitor, Axel Reßmann, Ulrich Thombansen, Dražen Veselovac, Konrad Willms, Thomas Gries, Walter Michaeli, Christian Hopmann, Uwe Reisgen, Robert Schmitt, Fritz Klocke

Computational Intelligence

Control Architecture for Human Friendly Robots Based on Interacting with Human

This paper discusses a control architecture for human friendly robot. Recently, robot middleware is developed for intelligent robot software platform. The robot system can be constructed by making the program of each functional module and connecting each other module. However, software architecture for connecting various modules is not provided now, therefore general versatility of a software module is deteriorated. In particular, a human friendly robot is very complicated relationship between software modules, because it should consider various situations such as human interaction, communication and safety. Therefore, we propose a control software architecture for a human friendly robot. We verify that availability of our proposal through experiments of clearing a table. Moreover, we add some intelligent modules on the proposed architecture to discuss availability of our proposed software architecture.

Hiroyuki Masuta, Eriko Hiwada, Naoyuki Kubota

Multi-modal Communication Interface for Elderly People in Informationally Structured Space

This paper proposes a universal remote controller using iPhone for elderly people. First, we discuss system configuration of universal remote controllers for elderly people in informationally structured space. The developed system is composed of database management server PC, physical robot partners, environmental systems, and human interface systems. Next, we explain human interface of universal remote controllers using accelerometer and compass. Finally, we discuss the usability of the developed universal remote controller through experimental results.

Rikako Komatsu, Dalai Tang, Takenori Obo, Naoyuki Kubota

Motion Control Strategies for Humanoids Based on Ergonomics

Due to the body analogy and their ability to communicate also via body language, humanoid robots are discussed as being mostly fit for applications in service robotics. An essential precondition of their deployment in human environments is the generation and control of human-like and thus predictable motions. This contribution introduces new motion control strategies which are based on the “Rapid Upper Limb Assessment” (RULA), a method for the analysis of ergonomic conditions at manual workplaces. We have previously adapted RULA to work with the Virtual Human, a simulated anthropomorphic multiagent system for the analysis of human motions and manipulations. Enabled by the control framework behind the Virtual Human, we here transfer RULA to humanoids in general and propose it as a well-defined and transparent heuristic for the control of human-like motions.

Christian Schlette, Jürgen Rossmann

Fuzzy Representations and Control for Domestic Service Robots in Golog

In the


domestic service robot competition, complex tasks such as “get the cup from the kitchen and bring it to the living room” or “find me this and that object in the apartment” have to be accomplished. At these competitions the robots may only be instructed by natural language. As humans use qualitative concepts such as “near” or “far”, the robot needs to cope with them, too. For our domestic robot, we use the robot programming and plan language Readylog, our variant of Golog. In previous work we extended the action language Golog, which was developed for the high-level control of agents and robots, with fuzzy concepts and showed how to embed fuzzy controllers in Golog. In this paper, we demonstrate how these notions can be fruitfully applied to two domestic service robotic scenarios. In the first application, we demonstrate how qualitative fluents based on a fuzzy set semantics can be deployed. In the second program, we show an example of a fuzzy controller for a follow-a-person task.

Stefan Schiffer, Alexander Ferrein, Gerhard Lakemeyer

Robot Control Systems

Minimum Jerk-Based Control for a Three Dimensional Bipedal Robot

In this paper, an optimized gait pattern generation is produced for a three dimensional bipedal robot using a Minimum Jerk criterion in the single support phase. Three approaches are introduced and compared in this framework. The Minimum Jerk based control approaches are the point-to-point, the Via-point and the shape function trajectory. Simulation results show that the point-to-point Minimum Jerk-based control cannot be satisfactory since the supposed swinging leg of the bipedal robot doesn´t lift off the ground. However, a bipedal stable human-like movement is guaranteed using both last approaches.

Amira Aloulou, Olfa Boubaker

Development of a Smart Motion Control Card with an IEEE-1394 Interface

IEEE 1394 is a high-efficiency communication network to guarantee timely data transmission and perform excellent network interconnection. In this paper, an IEEE-1394-based smart motion control card is presented, which is constructed by the hardware structure of the combination of a digital signal processor (DSP) and a field-programmable gate array (FPGA). The former DSP module implements an IEEE-1394 controller, a servo controller and memory mapping for FPGA access, while the FPGA module is utilized to achieve the logical functions containing quadrature-encoder-pulse (QEP) circuit, feedback count, direction decoder, addressing mapping, DAC pre-processing circuit and I/O interface. For real-time communication, an ISA/IEEE-1394 interface board for the host is designed and the Ardence Real-time Extension (RTX) is adopted for deterministic control of Windows XP-based systems. As a meaningful attempt, an experimental platform is established to evaluate the communication performance of the IEEE-1394 interface. The experimental results show excellent real-time communication performance, which demonstrates the feasible application of the IEEE 1394 interface for distributed motion control systems.

Guo-Ying Gu, LiMin Zhu, Ying Feng

Control System by Observer for a Hyper-redundant Robot

The paper focuses on the control of a class of hyper-redundant arms with continuum elements, with boundary measuring and control. First, the dynamic model of the continuum arm is presented. The measuring systems are based on the film sensors that are placed at the terminal sub-regions of the arm. The observers are proposed in order to reconstruct the full state of the arm. A back-stepping method is used to design a boundary control algorithm. Numerical simulations of the arm motion toward an imposed position are presented. An experimental platform shows the effectiveness of the proposed methods.

Mircea Ivanescu, Nirvana Popescu, Mihaela Florescu

Towards a Multi-peclet Number Pollution Monitoring Algorithm

Environments can range from low peclet numbers in which diffusion is predominant to high peclet numbers in which turbulence and advection occur. Control algorithms deployed on robotic platforms to monitor spatiotemporal distributions are often very specific to a particular peclet number environment and suffer reduction in efficiency when used in another peclet number environment. This paper investigates this issue and proposes the development of a pollution monitoring controller that can be used in various environments possessing different peclet numbers. A diffusion based controller and a controller that uses velocity flow information present in the environment are used as candidates for investigation. Even though the diffusion based controller lacks the ability to find a pollution source in a high turbulent environment, it still possess a desirable characteristic that could be used to map a pollution plume in a seaport environment.

John Oyekan, Dongbing Gu, Huosheng Hu

Self-balancing Controllable Robots in Education: A Practical Course for Bachelor Students

We present a framework for a programming course for undergraduate computer science students. The technical motivation is to implement a two-wheeled self-balancing controllable robot. Advanced requirements make it a full-grown software project. The emphasis of this course is on one hand to teach basic concepts of software programming. The students work in groups of five and each student is assigned a role, which is typical for the software development process. On the other hand, the course is intended to give some basic hands-on experience in control theory.

Paul Hänsch, John Schommer, Stefan Kowalewski

Intelligent Control Design in Robotics and Rehabilitation

For control purposes in robotics or rehabilitation, we may use properly simplified dynamic models with a reduced number of degrees of freedom. First, we define a set of variables that best characterize its dynamic performance in the required motion task. Second, driving forces/torques are properly assigned in order to achieve the required dynamic performance in an efficient way. The usual performance requirements are for positioning accuracy, movement execution time, and energy expenditure. We consider complex biomechatronic systems (BMS) like human with active orthosis or robotic arm that have to perform two main types of motion tasks: goal-directed movements and motion/posture stabilization. We propose new design concepts and criteria for BMS based on necessary and sufficient conditions for their robust controllability. Using simplified, yet realistic, models, we give several important examples in robotics and rehabilitation to illustrate the main features and advantages of our approach.

Petko Kiriazov, Gergana Nikolova, Ivanka Veneva

Human-Robot Interaction

Behavior Based Approach for Robot Navigation and Chemical Anomaly Tracking

We present a system for detecting chemical anomaly and to track the anomaly to its source. The chemical sensors are mounted on mobile robots. We describe navigating through medium and close proximity environments through a behavior-based approach, which can be represented through subsumption architecture approach. In this paradigm different robot behaviors compete for the actuators that make the robot act in the world. Further, these behaviors are driven by sensory data received through sensors and analyzed. In our case we support three competing behaviors that will be arbitrated: 1) navigation of indoor spaces through visual landmark recognition, 2) confined spaces navigation through close proximity sensors, and 3) airborne chemical and / or ground trail following.

Sambit Bhattacharya, Bogdan Czejdo, Shubo Han, Mohammad Siddique

Detection of Lounging People with a Mobile Robot Companion

This paper deals with the task of searching for people in home environments with a mobile robot. The robust estimation of the user’s position is an important prerequisite for human robot interaction. While detecting people in an upright pose is mainly solved, most of the user’s various poses in living environments are hard to detect. We present a visual approach for the detection of people resting at previously known seating places in arbitrary poses, e.g. lying on a sofa. The method utilizes color and gradient models of the environment and a color model of the user’s appearance. Evaluation is done on real-world experiments with the robot searching for the user at different places.

Michael Volkhardt, Steffen Müller, Christof Schröter, Horst-Michael Groß

Autonomous Control for Human-Robot Interaction on Complex Rough Terrain

Cooperation control between human’s and mobile manipulators has received big interest in the last few years. Mobile manipulators (MMs) have been suggested for various applications such as tasks involving hazardous environments, explosive handling, waste management, outdoor exploration and space operations. This paper describes a novel control algorithm for human-MM cooperation executing tasks in rough outdoor terrains. The proposed approach uses inexpensive common MM sensors such as wrist force/torque, IMU and joint/wheel encoders to achieve minimal human effort. The paper describes a control mechanism that enables ground mobile manipulators to execute complex tasks in cooperation with humans or other autonomous robots when working in unknown, dynamic heterogeneous outdoor rough terrains/environments. Simulation tests using a detailed SIMMECHANICS/ SIMULINK model of the employed MM are presented to illustrate and show the performance of the developed mechanisms.

Mahmoud Mustafa, Alex Ramirez-Serrano

A Modular Approach to Gesture Recognition for Interaction with a Domestic Service Robot

In this paper, we propose a system for robust and flexible visual gesture recognition on a mobile robot for domestic service robotics applications. This adds a simple yet powerful mode of interaction, especially for the targeted user group of laymen and elderly or disabled people in home environments. Existing approaches often use a monolithic design, are computationally expensive, rely on previously learned (static) color models, or a specific initialization procedure to start gesture recognition. We propose a multi-step modular approach where we iteratively reduce the search space while retaining flexibility and extensibility. Building on a set of existing approaches, we integrate an on-line color calibration and adaptation mechanism for hand detection followed by feature-based posture recognition. Finally, after tracking the hand over time we adopt a simple yet effective gesture recognition method that does not require any training.

Stefan Schiffer, Tobias Baumgartner, Gerhard Lakemeyer

Safety System and Navigation for Orthopaedic Robot (OTOROB)

OTOROB is a telemedicine mobile robot for orthopaedic surgeons that have remote presence capability to diagnose patients in remote area. As a telemedicine robot that interacts with human being, it is required to have extensive safety system. This paper presents a Fuzzy Logic based Danger Monitoring System (DMS) and Fail-Safe and Auto Recovery System (FSARS) that monitors the robot’s operation. It is incorporated to the onboard flexible robotic arm vision system and the robot’s navigation system. It monitors the robot surrounding and internal systems for danger or failures and takes precaution measures to overcome it. The system is tested by a set of experiments and found to be demonstrating an acceptable performance.

Muralindran Mariappan, Thayabaren Ganesan, Vigneswaran Ramu, Muhammad Iftikhar

Approaching a Person in a Socially Acceptable Manner Using a Fast Marching Planner

In real world scenarios for mobile robots, socially acceptable navigation is a key component to interact naturally with other persons. On the one hand this enables a robot to behave more human-like, and on the other hand it increases the acceptance of the user towards the robot as an interaction partner. As part of this research field, we present in this paper a strategy of approaching a person in a socially acceptable manner. Therefore, we use the theory of “personal space” and present a method of modeling this space to enable a mobile robot to approach a person from the front. We use a standard Dynamic Window Approach to control the robot motion and, since the personal space model could not be used directly, a Fast Marching planner is used to plan an optimal path to approach the person. Additionally, we give a proof of concept with first preliminary experiments.

Jens Kessler, Christof Schroeter, Horst-Michael Gross

Manipulators and Applications

Stiffness Identification for Serial Robot Manipulator Based on Uncertainty Approach

Stiffness of the robot manipulator plays a crucial role in improving welding accuracy. Due to the relatively low stiffness, the robot can hardly achieve the specified accuracy under the loading condition. Hence, identifying the joint stiffness and compensating the displacement in Cartesian space is an intuitive work to do in welding industry. Although substantial works had been done on stiffness modeling and identification, the uncertainties existed in transmission and manufacturing were neglected which may affect the manipulator’s stiffness and accuracy to a certain extent, in practice. In this paper, we propose an uncertain approach to identify the stiffness of a welding manipulator KUKA-KR16. Firstly, the uncertainties of the D-H parameters are considered and simulated by Monte-Carlo method. Then, the Cartesian stiffness is identified through an experiment which is composed of API laser tracker and a cable pulley system with deadweights. Finally, combining the previous enhanced stiffness model, we obtain the distribution of the joint stiffness, based on which the compensation results are proved to be effective in Cartesian space.

Xiaoping Zhang, Wenyu Yang, Xuegang Cheng, YuShan Chen

Serial Comanipulation in Beating Heart Surgery Using a LWPR-Model Based Predictive Force Control Approach

Compensation of cardiac motion during robot-assisted surgical procedures is needed to ensure better quality stabilization. Serial Comanipulation to actively compensate physiological motion is one alternative to common used Teleoperation techniques. In this paper, a 1 DOF hand-held force controlled prototype is presented. The active part of the instrument moves in synchronism with the heart motion in order to guarantee that the contact is maintained thanks to the application of a controlled force, while the surgeon’s hand is in charge to perform the surgical task.

It then focuses on a crucial control aspect: there is a lack of a parametric model describing the interaction between the surgical instrument and the heart that would provide enough precision for prediction. Namely, the robot low level controller and the beating heart are modeled thanks to a Locally Weighted Projection Regression (LWPR). The paper discusses how this technique can be used in the context of predictive force control and shows conclusive simulation results.

Juan Manuel Florez, Delphine Bellot, Jérôme Szewczyk, Guillaume Morel

Application of Wavelet Networks to Adaptive Control of Robotic Manipulators

In this paper, a wavelet-based adaptive control is proposed for a class of robotic manipulators, which consist of nonlinearities for friction effects and uncertain terms as disturbances. The controller is calculated by using a mixed of feedback linearization technique, supervisory control and H


control. In addition, the parameter adaptive laws of the wavelet network are developed using a Lyapunov-based design. It is also shown that both system tracking stability and convergence of the error estimation can be guaranteed in the closed-loop system. Simulation results on a three-link robot manipulator show the satisfactory performance of the proposed control schemes even in the presence of large modeling uncertainties and external disturbances.

Hamid Reza Karimi

Development of a Deep Ocean Master-Slave Electric Manipulator Control System

Underwater vehicles and manipulators play an important role in underwater tasks such as salvage, maintenance and so on. Underwater environment, especially deep Ocean is a hazard place for human to explore for its high water pressure. It is illustrated a development course of a deep ocean Master-slave manipulator for the demand of emergency oil exploration system. This paper describes a development of master-slave manipulator feedback control system. The underwater control system is based network and consisted of three embedded PC/104 computers which are used for servo control, task plan and target sensor respectively. The sensor control system and strategy of the master-slave manipulator control system are discussed in three parts include structure, modeling and experiment. Finally; establish the two-port network mode based scaling condition and analysis the stability of the overall manipulator control system.

Xiong Shen, GuoHua Xu, Kun Yu, Guoyuan Tang, Xiaolong Xu

Modelling of Flexible Link Manipulators

The present work describes a method for generating the dynamic equations within the Hamiltonian formalism of flexible robots with open-chain linkage mechanisms. Rotations are presented through vectors as elements of a Lie group with a smart composition law. The exact treatment of the flexible robots leads to partial differential equations which describe the nature of the elasticity. In spite of the fact that examples for control laws obtained from such direct approach exist, the common practice is to work with finite–dimensional approximations. Both approaches are considered and some conclusions are made. An example of single-link flexible manipulator is given.

Clementina Mladenova

An Outline for an Intelligent System Performing Peg-in-Hole Actions with Flexible Objects

We describe the outline of an adaptable system which is able to perform grasping and peg-in-hole actions with flexible objects. The system makes use of visual tracking and shape reconstruction, physical modeling of flexible material and learning based on a kernel density approach. We show results for the different sub-modules in simulation as well as real world data.

Andreas Jordt, Andreas R. Fugl, Leon Bodenhagen, Morten Willatzen, Reinhard Koch, Henrik G. Petersen, Knud A. Andersen, Martin M. Olsen, Norbert Krüger

Stability, Dynamics and Interpolation

Framework for Use of Generalized Force and Torque Data in Transitional Levels of Autonomy

Manipulation of hazardous materials requires the use of robotics to limit exposure of human operators to the danger. In order to improve manipulator effectiveness while ensuring reliability and redundancy, layers of control are implemented in increments. Each level of autonomy is established such that should a fault occur, the control system can be operated at a lower level of autonomy. Force and torque data can be used as both a structural element of a level of autonomy and for fault detection. This paper presents a framework for the use of generalized force and torque data for improving manipulator safety, operational effectiveness, and world model augmentation. The framework is applied to a demonstration of the automated door opening.

Kyle Schroeder, Mitch Pryor

A New Solution for Stability Prediction in Flexible Part Milling

The machining instability (chatter phenomenon) easily occurs when flexible parts are machined. This paper predicts the milling stability domain of the flexible part with multiple structure mode interaction induced by the cutting force. First, the dynamic milling process of a flexible part is modeled as a multiple modal degree-of-freedom mechanical model. Then, the full-discretization method is employed to calculate the stability boundary and the cases with different factors are compared, the simulation shows that the cutting position has the dominating effect on the stability analysis of the flexible part milling. The chatter experiment verifies the validity of the proposed method. The proposed method can be used to improve the machining efficiency of flexible parts in aerospace and power industries.

XiaoJian Zhang, Caihua Xiong, Ye Ding

A Practical Continuous-Curvature Bézier Transition Algorithm for High-Speed Machining of Linear Tool Path

A continuous-curvature smoothing algorithm is developed to approximate the linear tool path for high speed machining. The new tool path composed of cubic Bézier curves and lines, which is everywhere



continuous, is obtained to replace the conventional linear tool path. Both the tangency and curvature discontinuities at the segment junctions of the linear tool path are avoided. The feed motion will be more stable since the discontinuities are the most important source of feed fluctuation. The algorithm is based upon the transition cubic Bézier curve that has closed-form expression. The approximation error at the segment junction can be accurately guaranteed. The maximal curvature in the transition curve, which is critical for velocity planning, is analytically computed and optimized. The curvature radii of all transition Bézier curves are also globally optimized to pursue the high feed speed by a linear programm model. Therefore, the algorithm is easy to implement and can be integrated into a post-process system.

Qingzhen Bi, Yuhan Wang, Limin Zhu, Han Ding

Design of a FPGA-Based NURBS Interpolator

In this paper, a NURBS hardware interpolator based on FPGA is designed to perform the feedrate profile scheduling, de-Boor Cox calculation and second-order Talor expansion to realize real-time interpolation. Look-ahead algorithm including curve-scanning, feedrate adjustment and acceleration/deceleration planning is implemented in the computer to release the computational load of the interpolator, whereas a motion control card with DSP+FPGA architecture receives the pre-processed results from the look-ahead circuit through PCI bus, and sequently performs the interpolation task in the FPGA and position servo control in the DSP. Experiments are carried out to verify the feasibility of this interpolator. The results imply the FPGA can finish the interpolation within 0.5ms, meanwhile its resource utilization and the calculation speed can compromise to satisfy the practical application.

Huan Zhao, Limin Zhu, Zhenhua Xiong, Han Ding

Iso-scallop Trajectory Generation for the 5-Axis Machining of an Impeller

Iso-scallop trajectories, tool orientations and a tool swept volume for the 5-axis machining of an impeller have been developed in the paper. As the merits of ball end tools, the planning work has been done on the cutter location surface. Firstly, a boundary curve is selected as the master trajectory, and cutting intervals are calculated from the geodesic curvature on the blade surface to keep a constant scallop height. Secondly, tool orientation determinations are discussed to promise the trajectories interference-free. The determined tool orientations are starting from the trajectory point and pointing to a guide line located in the mid-plane between two neighbor blades. Finally, measurement results from a machining simulation and a produced tool swept volume are given to illustrate the feasibility and validity of the proposed method.

Xubing Chen, Jinbo Wang, Youlun Xiong

Evolutionary Robotics

Swarm Robot Flocking: An Empirical Study

Robots can be used in exploration or investigation of unknown terrains especially if the environment is dangerous. It is customary to employ a sophisticated robot for such task. However, this approach is vulnerable since a failure of the robot means, failure of the entire mission. An emerging approach in robotics research is to employ many simple robots that can collectively achieve a demanding task. Even the failure of some robots should not affect the overall mission. Maneuvering such large systems poses new challenges in controlling them. In our earlier work, a control strategy, namely triangular formation algorithm (TFA), was developed and tested using simulation tools. The TFA is a local interaction strategy which basically makes three neighboring robots to form a regular triangular lattice. Simulation results show that swarm behaviors such as aggregation, flocking and obstacle avoidance can be achieved successfully. Here, we are concerned with implementing the algorithm in practice with real robots. We have developed a swarm of five robots and tested the performance of the algorithm in practice. This paper presents our initial findings.

M. Fikret Ercan, Xiang Li

Self-reconfiguration Path Planning Design for M-Lattice Robot Based on Genetic Algorithm

M-Lattice is a kind of lattice modular robot, which can finish self-reconfiguration in three-dimensional plane. How to substitute the broken modules effectively is a critical question for modular robot system. In order to solve it, we introduce the topology structure of M-Lattice system and math representation for the reconfiguration question. An energy factor to illustrate the relationship between energy cost and moving path is defined. The non-real time path planning based on genetic algorithm is also given. From the results of simulation, the reliability and feasibility of the planning is demonstrated.

Enguang Guan, Zhuang Fu, Weixin Yan, Dongsheng Jiang, Yanzheng Zhao

Mobile Robot Controller Design by Evolutionary Multiobjective Optimization in Multiagent Environments

Evolutionary computation has been often used for the design of mobile robot controllers thanks to its flexibility and global search ability. A lot of studies have been done based on single-objective functions including weighted-sum scalarizing objective functions. For an example of mobile robot navigation, at least the minimization of the arrival time to the target and the minimization of dangerous situations should be considered. In this case, a weighted-sum of two objectives is always minimized. It is, however, difficult to specify an appropriate weight vector beforehand. This paper demonstrates the application of evolutionary multiobjective optimization to mobile robot navigation in order to optimize the conflicting objective simultaneously. We analyze the obtained non-dominated controllers through simulation experiments in multiagent environments. We also show the utilization of the obtained non-dominated controllers for situation change.

Yusuke Nojima, Hisao Ishibuchi

Learning Intelligent Controllers for Path-Following Skills on Snake-Like Robots

Multi-link wheeled robots provide interesting opportunities within many areas such as inspection and maintenance of pipes or vents. A key functionality in order to perform such operations, is that the robot can follow a predefined path fast and accurately. In this paper we present an algorithm to learn the path-following behavior for a set of motion primitives. These primitives could then be used by a planner in order to construct longer paths. The algorithm is divided into two steps: an example-based stage for controller learning, and a controller tuning stage, based on an objective function and simulations of the path-following process. The path-following controllers have been tested with a simulator of a multi-link robot in several complex paths, showing an excellent performance.

Francisco Javier Marín, Jorge Casillas, Manuel Mucientes, Aksel Andreas Transeth, Sigurd Aksnes Fjerdingen, Ingrid Schjølberg

Bio-inspired Robotics

Modular Behavior Controller for Underwater Robot Teams: A Biologically Inspired Concept for Advanced Tasks

In ambition to give subaqueous robot groups more robustness and behavioral flexibility for real applications, this paper proposes a modularized behavior control architecture. Schools of naval mammals provide the proof that also individual members of the group can achieve higher leveled intelligence independent of the simplicity of their collective behavior. Due to their structurally and functionally modularized brain organization, dolphins are capable of language based communication and learning complex motions by human training. Inspired by dolphins, 3 modules for the behavior controller can be conceptualized. The swarming module optimized by evolutionary methods represents the basic behavior given in the natural environment. The mission module includes extendable sets of behavior primitives that can be structured by reinforcement learning. A knowledge based sensing module can be implemented separately to increase the information reliability. With this approach, subaqueous robot schools can be expected to perform more advanced tasks than just moving as a swarm.

Dong-Uck Kong, Jinung An

BioMotionBot – A New 3D Robotic Manipulandum with End-Point Force Control

In this paper we present the design of a new 3D robotic manipulandum that will be used in human motor-control research and additionally enables physiotherapists to design tailor-made robotic therapies. Moreover, it offers the opportunity to develop completely new types of movement-specific coordination and condition training programs in sports. The presented manipulandum has a special designed 3D kinematics that allows movements in 3D space while maintaining its orientation. The paper contains an overview of the mechanical design, the electronic components, the user interface, the design of the control system as well as a first performance test.

Volker Bartenbach, Klaus Wilging, Wolfgang Burger, Thorsten Stein

Adaptive Control Scheme with Parameter Adaptation - From Human Motor Control to Humanoid Robot Locomotion Control

As the origin intention of humanoid robot is showing the possibility of the biped walking and explaining the principle, there are many common issues between human motor control and humanoid robot locomotion. This paper considers two major common issues of the two researches. First is modeling. Both in human dynamics simplification and humanoid dynamics modeling, we actively or passively choose parts of the variable states because of dynamics simplification and unmodeled dynamics. In these cases, it is questionable that the dynamics represented by the partial variables states still corresponds to a physical system. In this paper, we discuss this problem and prove that the partial dynamics satisfies the conditions of a physical system, which is the basis of control scheme design. Second is control. To tolerate all the errors or perturbations, we design a control scheme which is composed of variable state control and parameter adaptation. The former can tolerate modeling error; the latter can identify the dynamic system in real time. Finally, we apply the proposed control scheme into a humanoid robot control case, which shows the effectiveness of the proposed control scheme.

Haiwei Dong, Zhiwei Luo

Online Walking Gait Generation with Predefined Variable Height of the Center of Mass

For biped robots one main issue is the generation of stable trajectories for the center of mass (CoM). Several different approaches based on the zero moment point (ZMP) scheme have been presented in the past. Due to the complex dynamic structure of bipedal robots, most of the considered algorithms use a simplified time invariant linear model to approximate the dynamics of the system. This model is extended to a time variant one and then used to generate stable CoM trajectories with variable predefined CoM height. This allows to generate trajectories online for walking underneath obstacles with more accuracy. It is shown that using this extended scheme it is possible to overcome some kinematic limits as joint speed in the knee or the maximum step length for common walking.

Johannes Mayr, Hubert Gattringer, Hartmut Bremer

Image-Processing Applications

Visual Control of a Remote Vehicle

We present a method for locating and determining the six degrees of freedom through a simple algorithm based on artificial vision. This algorithm can estimate the relative orientation of the camera with respect to a precise figure, it gives, roll, pitch and yaw, as well as the distance to the figure. We make use of the Euler number of a set of figures with a given distribution. We use the system to drive a small radio controlled car with the only assistant of the information gathered by a standard web-cam.

David Sanchez-Benitez, Jesus M. de la Cruz, Gonzalo Pajares, Dawei Gu

Longitudinal and Lateral Control in Automated Highway Systems: Their Past, Present and Future

Due to the increase in road transportation by 35% over the last years in Europe it is essential to find solutions to optimize highway traffic. Therefore, several projects involving automated highway systems were initiated. In these systems, the longitudinal and lateral controls enable (with the help of other components) vehicles to be coupled electronically to form a platoon. Here, just the first vehicle is driven actively and the following vehicles are controlled automatically. Several projects were initiated to develop systems for different environments (i.e. Urban, Motorway). However, the developed techniques still are limited in their application range and e.g. cannot be applied in unstructured environment (i.e. rural or dirty areas). Furthermore, they were not tested for many different heterogeneous vehicles like trucks or passenger cars. This paper presents the past and present of automated highway systems and discusses solutions for future developments, e.g. how existing technologies can be adapted for a wider application range.

Mohammad Alfraheed, Alicia Dröge, Max Klingender, Daniel Schilberg, Sabina Jeschke

Surface Defects Classification Using Artificial Neural Networks in Vision Based Polishing Robot

One of the highly skilled tasks in manufacturing is the polishing process. The purpose of polishing is to get uniform surface roughness. In order to reduce the polishing time and to cope with the shortage of skilled workers, robotic polishing technology has been investigated. This paper proposes a vision system to measure surface defects that have been classified to some level of surface roughness. Artificial neural networks are used to classify surface defects and to give a decision in order to drive the actuator of the arm robot. Force and rotation time have been chosen as output parameters of artificial neural networks. The results show that although there is a considerable change in both parameter values acquired from vision data compared to real data, it is still possible to obtain surface defects classification using a vision sensor to a certain limit of accuracy. The overall results of this research would encourage further developments in this area to achieve robust computer vision based surface measurement systems for industrial robotics, especially in the polishing process.

Anton Satria Prabuwono, Adnan Rachmat Anom Besari, Ruzaidi Zamri, Md Dan Md Palil, Taufik

Efficient Skin Detection under Severe Illumination Changes and Shadows

This paper presents an efficient method for human skin color detection with a mobile platform. The proposed method is based on modeling the skin distribution in a log-chromaticity color space which shows good invariance properties to changing illumination. The method is easy to implement and can cope with the requirements of real-world tasks such as illumination variations, shadows and moving camera. Extensive experiments show the good performance of the proposed method and its robustness against abrupt changes of illumination and shadows.

Bishesh Khanal, Désiré Sidibé


Weitere Informationen

Premium Partner