Direct methods for vision-based robot control : application and implementation

This thesis deals with Visual Servoing and its strictly connected disciplines like projective geometry, image processing, robotics and non-linear control. More specifically the work addresses the problem to control a robotic manipulator through one of the largely used Visual Servoing techniques: the Image Based Visual Servoing (IBVS). In Image Based Visual Servoing the robot is driven by on-line performing a feedback control loop that is closed directly in the 2D space of the camera sensor. The work considers the case of a monocular system with the only camera mounted on the robot end effector (eye in hand configuration). Through IBVS the system can be positioned with respect to a 3D fixed target by minimizing the differences between its initial view and its goal view, corresponding respectively to the initial and the goal system configurations: the robot Cartesian Motion is thus generated only by means of visual informations. However, the execution of a positioning control task by IBVS is not straightforward because singularity problems may occur and local minima may be reached where the reached image is very close to the target one but the 3D positioning task is far from being fulfilled: this happens in particular for large camera displacements, when the the initial and the goal target views are noticeably different. To overcame singularity and local minima drawbacks, maintaining the good properties of IBVS robustness with respect to modeling and camera calibration errors, an opportune image path planning can be exploited. This work deals with the problem of generating opportune image plane trajectories for tracked points of the servoing control scheme (a trajectory is made of a path plus a time law). The generated image plane paths must be feasible i.e. they must be compliant with rigid body motion of the camera with respect to the object so as to avoid image jacobian singularities and local minima problems. In addition, the image planned trajectories must generate camera velocity screws which are smooth and within the allowed bounds of the robot. We will show that a scaled 3D motion planning algorithm can be devised in order to generate feasible image plane trajectories. Since the paths in the image are off-line generated it is also possible to tune the planning parameters so as to maintain the target inside the camera field of view even if, in some unfortunate cases, the feature target points would leave the camera images due to 3D robot motions. To test the validity of the proposed approach some both experiments and simulations results have been reported taking also into account the influence of noise in the path planning strategy. The experiments have been realized with a 6DOF anthropomorphic manipulator with a fire-wire camera installed on its end effector: the results demonstrate the good performances and the feasibility of the proposed approach.

Wheeled Mobile robots (WMRs) are very popular because they are appropriate for applications with relatively low mechanical complexity and energy consumption. WMRs are used in many military and civilian applications. These applications include surveillance, transportation, inspection, exploration, and many other applications either in structured or unstructured environments. This paper provides a review on the developments of the wheeled mobile robots (WMRs) control. It reviews the kinematic and dynamic modeling of the robots, the path following and tracking, the avoiding obstacles control algorithms, and finally a review on robot vision and visual servoing control system is presented including the main challenges in each field of research.

Bidirectional Human-Robot Learning: Imitation and Skill Improvement

Ideas about retrieval of memories are based largely on localization of function inferred after brain damage. On the other hand, communication in the cerebral cortex occurs through an intricate network of connections, suggesting highly distributed circuits. Loss of memory after lesions thus affects not only the damaged area, but also its rich inputs and outputs. The discussion here is based on a structural model of neural communication that can help explain the process of recruiting highly distributed relevant signals for conceptual and sensory recognition and memory. Discussion of cortical communication has traditionally been sensory-centric, based on the flow of information from the periphery to primary sensory areas and beyond, or in the reverse direction. However, visual signals often do not activate areas along a sequential route as predicted by a hierarchical model, exemplified by the fact that some visual signals activate prefrontal areas earlier than some visual association areas. On the other hand, the structural model for connections is based on the fundamental principle of systematic variation in laminar architecture of the cortex and applies to all cortical systems. The structural model is relational: it predicts that laminar-specific connections in the cortex depend on the structural (laminar) relationship between linked areas. Thus, connections that link an area with simpler laminar structure originate mainly – but not exclusively – in the deep layers and terminate mainly in the upper layers of an area with more elaborate laminar structure. In the opposite direction, connections originate in the upper layers and terminate in the middle-deep layers. Because the laminar architecture across cortical areas is graded, so is the distribution of connections within layers. The structural model holds for connections in all cortical systems and diverse species. Moreover, laminar-specific connections as predicted by the structural model terminate within inhibitory microenvironments that also differ by the distribution of functionally distinct classes of inhibitory neurons that can dynamically shift activity among layers and change cortical rhythms. The relational nature of the structural model and predictions of graded patterns of connections thus suggest that distributed areas and layers are recruited dynamically to meet task demands, including recollection of concepts and events.

BackgroundControl of reaching movements for manual work, vehicle operation, or interactions with manual interfaces requires concurrent gaze control for visual guidance of the hand. We hypothesize that reaching movements are based on negotiated strategies to resolve possible conflicting demands placed on body segments shared by the visual (gaze) and manual (hand) control systems. Further, we hypothesize that a multiplicity of possible spatial configurations (redundancy) in a movement system enables a resolution of conflicting demands that does not require sacrificing the goals of the two systems.MethodsThe simultaneous control of manual reach and gaze during seated reaching movements was simulated by solving an inverse kinematics model wherein joint trajectories were estimated from a set of recorded hand and head movements. A secondary objective function, termed negotiation function, was introduced to describe a means for the manual reach and gaze directing systems to balance independent goals against (possibly competing) demands for shared resources, namely the torso movement. For both systems, the trade-off may be resolved without sacrificing goal achievement by taking advantage of redundant degrees of freedom. Estimated joint trajectories were then compared to joint movement recordings from ten participants. Joint angles were predicted with and without the negotiation function in place, and model accuracy was determined using the root-mean-square errors (RMSEs) and differences between estimated and recorded joint angles.ResultsThe prediction accuracy was generally improved when negotiation was included: the negotiated control reduced RMSE by 16% and 30% on average when compared to the systems with only manual or visual control, respectively. Furthermore, the RMSE in the negotiated control system tended to improve with torso movement amplitude.ConclusionsThe proposed model describes how multiple systems cooperate to perform goal-directed human movements when those movements draw upon shared resources. Allocation of shared resources can be undertaken by a negotiation process that is aware of redundancies and the existence of multiple solutions within the individual systems.

Natural and Anthropogenic Drivers of Water Quality in the Normanby Basin and Princess Charlotte Bay, Cape York Peninsula, Australia

This thesis presents research on scheduling in an uncertain environment, which forms a part of the rolling stock life cycle logistics applied research and development program funded by Dutch railway industry companies. The focus therefore lies on scheduling of maintenance operations on rolling stock in the railway industry. The first chapter describes some of the history of the Dutch railways, focusing on the rolling stock used, and it introduces the context in which NedTrain, a major Dutch rolling stock maintenance company, operates. While maintenance of rolling stock is not a new problem, recent changes in the field are identified in this chapter which introduce new challenges: the first is the declining involvement of maintenance experts in the procurement of new rolling stock, the second is the on-going demand for increasing efficiency. Based on this discussion, the goal of this thesis is to devise a method with which schedules can be created based on a flexible process. The schedules resulting from this method are to be both flexible, meaning that they can be adapted easily, and robust, meaning that they are resilient to the effects of uncertain events. Lastly, our goal is to create the schedules in such a way that the effects of disturbances in one team have little or no effect on any other teams in the same workshop. Chapter 2 examines existing research in the context of the formulated research goals. It is shown that the well-known Resource-Constrained Project Scheduling Problem (RCPSP) can be mapped to the problem of scheduling maintenance tasks on trains in the NedTrain workshops. The most important methods to solve the RCPSP are discussed, among which are many exact methods. Both the scale of the problem and the requirement of having a flexible schedule instead of a fixed time assignment for the tasks prohibit using these methods. Among the heuristic methods discussed, the Precedence Constraint Posting method turns out to be an especially good fit: it is able to solve large-sized problems very quickly, and it uses a Simple Temporal Network (STN) as a way of representing many different solutions to the original problem, offering a lot of flexibility. Based on the research discussed, research questions are formulated to answer the research goals. - The first question asks how to define criteria encapsulating the notion of flexibility as posed by our research goals, and how to measure the flexibility of a schedule. - The second question asks how we can make the best use of available flexibility to attain robustness. - The third question asks if there is also a way to extend the concept of temporal flexibility to include a form of sequential flexibility. - The last question asks how we can use our work on robustness to decouple schedules, ensuring that disturbances in the schedule for one agent have no impact on any other agents. The third chapter investigates criteria corresponding to our intuitive notion of flexibility in scheduling. The concept of interval schedules serves as a basis for this investigation: the idea is that each task in the schedule is assigned an interval from which its start time can be picked, and the length of all these intervals serves as a representation for the flexibility of the schedule. We show that flexibility measures in existing research lead to an over-estimation of the total available flexibility, because dependencies between tasks are not taken into account properly: simply adding the interval sizes for two dependent tasks causes over-estimation, because the picked time for one task can reduce the interval size for another task. Our first major contribution is an independence requirement for the creation of these intervals: we state that intervals should be constructed in such a way that any choice in one interval has no impact on the available choices in any of the other intervals. Our second major contribution is a method to actually construct such intervals, which results in a valid interval schedule for an \stn, allowing efficient schedule execution under dynamic circumstances. Additionally, we note that an interval schedule can also serve as a mechanism to decouple a schedule, giving a partial answer to the last research question. In the fourth chapter, we investigate how to make use of the available flexibility to ensure that schedules are robust. We show that having a maximal total amount of flexibility does not imply that the schedule is as robust as possible: maximization might lead to a very skewed distribution of flexibility over the schedule. Different ways of distributing flexibility over a schedule are proposed and analyzed in an experimental setting. From the experiments it can be clearly concluded that sacrificing some of the total flexibility to improve the distribution of flexibility can have a positive influence on the robustness of the schedule. We also propose a different maximization strategy: one which maximizes the minimum of the interval sizes for the schedule. This strategy is shown to work especially well for small delays. So far we only discussed strategies concerning temporal flexibility, i.e., those in which start times remain flexible instead of fixed. Chapter 5 introduces a novel representation of solutions for planning problems in which not all ordering decisions needed to avoid resource contention are made in advance. This is achieved by grouping tasks which need to be executed sequentially together, in such a way that either ordering of the tasks remains feasible during execution time. An analysis using simulation experiments shows that this technique results in higher robustness, at the cost of somewhat lower resource utilization. The proposed algorithm to construct such schedules offers limited control over the size of the grouped tasks. The experiments show a slightly counter-intuitive results: schedules with larger groups offer lower robustness than those with smaller groups. A plausible explanation is the fact that delays can occur at any place in a schedule. Having multiple smaller groups instead of a few very large ones increases the chance that such a delay can be compensated for using a task group.

Enhancing Humans Trust in Robots through Explanations

Modelling life trajectories and mode choice using Bayesian belief networks

knowledge base that the individual is endowed with, and which can grow by the influence of culture. This component does not show decline until very old age. The ‘mechanics’ component is the means to gain knowledge, while ‘pragmatics’ component is the abstract knowledge gained, or that is being 'crystallized'. Evidence for this two-component model comes from multiple sources (see Baltes et al., 2006; p. 598600.), including longitudinal studies (e.g. Seattle Longitudinal Study) indicating that cognitive abilities related with the ‘mechanic’ component (i.e. reasoning, memory, perceptual speed, spatial orientation) follow a nearly linear decline throughout adulthood, while cognitive abilities related with the ‘pragmatics’ component (i.e. verbal ability, numeric ability) does not show decline until very old age (see Figure 1). Figure 1 Cross-sectional data from the Seattle Longitudinal Study (image from Hedden & Gabrieli, 2004) Causes of Age-Related Changes in Cognition In this section different approaches to explain age-related changes in cognition will be outlined, along with empirical findings. The first of these is the 'general slowing' hypothesis, which mainly explains age-related changes in intellectual fluid abilities as a 9 function of a general slowing of processes (T. Salthouse, 1996). According to this theory, the slowing down of executing processes leads to impaired cognitive performance in most cognitive tasks. While the general slowing hypothesis is still an integral theory for explaining age-related decline in many areas, it is not adequate to explain age-differences in conditions with high cognitive control demands (Mayr & Kliegl, 1993; Verhaeghen, Steitz, Sliwinski, & Cerella, 2003). Age-related changes in intellectual abilities can also be explained as a function of changes in inhibitory processes. An early study (Kramer, Humphrey, Larish, & Logan, 1994) for instance, found little evidence for age-related deficits in inhibitory processes, however older adults had more difficulty than younger adults to abort an action (in a stop signal paradigm where participants are asked to respond to a visual stimuli unless they hear an auditory signal), and it also proved more difficult for older adults to learn new rules in a categorization task. Also, comparing younger and older adults, inhibition related processes were found to be delayed in the elderly as evidenced by delayed inhibition-related ERPs (event related potentials, recorded with electroencephalogram) (Falkenstein, Hoormann, & Hohnsbein, 2002). It is also possible to attribute age-related differences in intellectual abilities to changes in working-memory performance. Recent studies have demonstrated that older adults show deficits in suppressing task-irrelevant information during working-memory performance whereas their ability to enhance task-relevant information is intact (Gazzaley, Cooney, Rissman, & D’Esposito, 2005). Furthermore, it has been shown that an interruption during task maintenance distracts both younger and older adults, however while younger adults quickly disengage from the task-irrelevant disruptor and reestablish task-relevant functional networks, older adults are impaired in dynamic shifting between competing representations (Clapp, Rubens, Sabharwal, & Gazzaley, 2011). Similar 10 findings came also from a study employing an inhibition of return paradigm (Wascher, Falkenstein, & Wild-Wall, 2011) and another study using a dual task paradigm (Hahn, Wild-Wall, & Falkenstein, 2011) suggesting that older adults process irrelevant information similarly as relevant information, and this leads to impairments in workingmemory performance. We can also approach to explain age-related differences in intellectual abilities in terms of biological changes in the nervous system. Some decades ago large scale general neuron loss was suggested to be the main reason for age-related cognitive decline (Ball, 1977; Coleman & Flood, 1987). However, recently it became evident that general neuron loss in most areas does not contribute significantly to age-related cognitive decline (Burke & Barnes, 2006) but rather there are specific areas such as the prefrontal cortex (PFC) in which neuron loss is related to cognitive decline. A study, for example, with aged monkeys have found an approximately 30 percent reduction of neurons compared to younger animals in the dorsolateral PFC area 8A, and this reduction correlated with impaired performance in a working-memory task, but at the same time other areas related to working-memory (such as area 46) were well preserved (Smith, Rapp, McKay, Roberts, & Tuszynski, 2004). Less age-related change has been reported for many electrophysiological properties of neurons in the hippocampus and the PFC, such as resting membrane potential, membrane time constant, threshold to elicit an action potential, and rise time and duration of an action potential (Burke & Barnes, 2006). However, reduced synapse number in older animals might be related with cognitive decline. It has been found in rats, that synapse number per neuron diminishes during aging in the dentate gyrus, which is a hippocampal sub-region that is important in spatial memory (Geinisman, Toledo-Morrell, & Morrell, 1986).

Fil: Roberti, Flavio. Consejo Nacional de Investigaciones Cientificas y Tecnicas. Centro Cientifico Tecnologico Conicet - San Juan. Instituto de Automatica. Universidad Nacional de San Juan. Facultad de Ingenieria. Instituto de Automatica; Argentina

In this paper, a new visual feedback control system for a swirling-flow melting furnace is proposed. The method controls the furnace combustion by manipulating the burner fuel based on image data, which shows the melting condition inside the furnace. The feature of this control system is that a tracking property is realized by using an adaptive control algorithm, which changes the control gain based on image evaluation as the melting point varies. Through actual experiments, it is clear that this visual feedback control system can control its combustion stably.

In this paper, the autonomous navigation of six-crawler machine is studied, and a visual tracking control method based on machine vision for fuzzy proportional–integral–derivative control of six-crawler machine is proposed. The steering principle of the six-crawler machine and the matching relationship between the steering angle and the speed of each crawler are introduced, and the control system is described in detail. Besides, the mathematical model for the unsteady steering is introduced to analyze the influence of deflection angle on the steering trajectory of the six-crawler machine. The image processing algorithm is programmed by LabVIEW software. After the image is fitted by graying, binary, filtering, edge detection, and least square method, the navigation line-fitting curve is obtained. The fuzzy proportional–integral–derivative control algorithm is programmed in the control system to control the six-crawler machine to drive along the navigation line. In order to obtain reasonable control parameters, a virtual prototype model of a six-crawler machine is established. In the CoLink module, the control algorithm of a six-crawler machine is established, and the co-simulation is carried out. By analyzing the simulation results, the control parameters of the fuzzy proportional–integral–derivative controller of the six-crawler machine are established. In order to verify the control effect of the visual tracking control system of the six-crawler machine, a physical prototype of the six-crawler machine is constructed and tested. The results show that the visual tracking control system of the six-crawler machine can complete the preset functions.

The European Union Reference Laboratory for Genetically Modified Food and Feed (EU-RL GMFF), accredited under ISO/IEC 17043, organised a comparative testing (CT) round for National Reference Laboratories (NRLs) nominated under Regulation (EC) No 882/2004, with voluntary participation of other official control laboratories.\nThe test items consisted of rice noodles and commercial soybeans spiked with ground powder of soybean event DP-356043-5 in two different concentrations (Level 1 and 2). Participants were required to perform species identification and test for the presence of any GM event in the two test items. Any event detected then had to be quantified. Participants could report the results in mass/mass % or copy/copy % and the EU-RL GMFF calculated the robust means (R) for Level 1 and 2 test items accordingly. The target standard deviation for CT was fixed by the Advisory Board for Comparative Testing at 0.2 for the event, based on the experience of previous CT rounds. The robust means and target standard deviation were used to derive z-scores for the participants’ results. \nEighty-eight laboratories from 42 countries registered for this CT round, of which 85 laboratories from 41 countries returned at least qualitative test results. \nWhen performing species identification, almost all laboratories correctly identified soybean and rice in both test items, and a few laboratories also detected maize and/or oilseed rape. In total 71 laboratories reported the presence of GM material in the test items, but 14 laboratories failed in this task. All of the 71 laboratories, except eight, correctly identified soybean event DP-356043-5 in the test items.\nResults of the quantitative evaluation of the GM content were satisfactory for both measurement units, with only two NRLs appointed under Regulation (EC) No 1981/2006 (one measuring in m/m % and one in cp/cp %) obtaining unsatisfactory z-scores (|z| ≥ 2.0) for both test items. \nDespite the overall satisfactory outcome of this CT round, only 58 % of participants provided information on measurement uncertainty in a complete and consistent manner, and further improvement in this crucial area is needed.

شناخت آسیب پذیری جمعیت و گروه های انسانی به عنوان اقدامی ابزاری در راستای پیشگیری از زایش فاجعه مبتنی بر زیستن در نواحی پر خطر مطرح است. تحلیل آسیب پذیری کانون‌های جمعیتی واقع در عرصه های خطر اقدامی علمی برای فراهم سازی زیر ساخت اطلاعاتی مناسب از درجه و میزان خطرپذیری از رخداد مخاطره است. بر اساس چنین دیدگاهی مسئله تحلیل آسیب پذیری ناشی از مخاطرات ژئومورفولوژیک فرونشست در جنوب شهر تهران مورد توجه قرار گرفته است. بدین منظور در ابتدا با استفاده از تکنیک تداخل سنجی راداری نرخ و دامنه فرونشست در محدوده مورد مطالعه استخراج گردید. سپس برای تولید نقشه فضایی عرصه های پر مخاطره نرخ و دامنه فرونشست به عنوان فاکتور عامل در زایش مخاطره به تحلیل گر سیستم در محیط GIS وارد گردید. نقشه خروجی معرف سلول‌های واجد ارزش متفاوت از فاکتور زایش مخاطره است. استخراج عرصه متعلق پهنه در معرض مخاطره نشان داد که از مجموع مساحت محدوده 7/12 درصد در پهنه های بسیار پرخطر، 7/14 در عرصه پرخطر،5/15 در عرصه نسبتاً پرخطر بیش از نیمی از محدوده (57درصد) در عرصه نسبتاً کم خطر قرار گرفته‌اند. برای ارزیابی میزان خطرپذیری، الگوی توزیع فضایی تأسیسات و ساخت‌وساز های موجود با نقشه عرصه های در معرض خطر هم نهاد گردید. مبتنی بر این اطلاعات و با استفاده از پیمایش میدانی اقدام به بررسی آسیب پذیری جمعیت ،مساکن و تأسیسات موجود در محدوده مورد مطالعه صورت پذیرفت. یافته های تحقیق موید آسیب پذیر بودن اکثر سکونت گاه‌ها و جمعیت ساکن درآنهاو تأسیسات موجود در محدوده می‌باشد. همچنین بر اساس پایش های میدانی صورت گرفته مشخص گردید که اکثر ساخت و سازهای موجود در منطقه غیر اصولی بوده که این امر باعث فراهم آوری زیرساخت آسیب پذیری برای مستحدثات در شرایط خطرپذیر است.