|
|
|||
|
|
|
||
| Research | Architecture of Autonomous Systems | Autonomous Systems | Artificial Intelligence | Education | ||
PUBLICATION LIST, including abstracts
Publication list, without abstracts.
| An omnidirectional camera has properties which are highly beneficial for navigation of a mobile robot. In this paper this is demonstrated with a new application; a visual map of the environment created from bird-eye view perspective. This visual map can be created on-line, which allows an operator to navigate the robot through an unknown environment based on this visual map. The quality of the map is tested in a self-localization experiment. |
Arnoud Visser and Julian de Hoog, "Amsterdam Oxford Joint Rescue Forces - Realistic Simulations to aid research and education in advanced Robot Control algorithms", Proceedings of the Scientific ICT Research Event Netherlands (SIREN 2008), page 22, Amsterdam, The Netherlands, September 2008. (Portable Document Format 671 Kb)
| A white paper with the shared research portfolio of advanced Robot Control algorithms at the laboratories in Amsterdam and Oxford. |
Arnoud Visser and Bayu A. Slamet, "Including communication success in the estimation of information gain for multi-robot exploration", Proceedings of the International Workshop on Wireless Multihop Communications i n Networked Robotics, Berlin, Germany, April 4, 2008. (Portable Document Form of 2.8 Mb).
| This article investigates the effect of incorporating knowledge about the communication possibilities in an exploration algorithm used to map an unknown environment. The mission is to explore a hypothetical disaster site with a small team of robots. The challenge faced by the robot team is to coordinate their actions such that they efficiently explore the environment in their search for victims. The coordination can only be optimal when the robots share the same map. With a limited communication range the map cannot be shared in all circumstances. This article concentrates on the effect of a distributed map, where each robot has only has knowledge of a part of the global map and has no guaranteed connection to the other robot or the operator. |
Arnoud Visser and Bayu A. Slamet, "Balancing the information gain against the movement cost for multi-robot frontier exploration", accepted for the 2nd European Robotics Symposium 2008, Prague, Czech Republic March 26-27, 2008. (Portable Document Form of 761 Kb)
|
This article investigates the scenario where a small team of robots needs
to explore a hypothetical disaster site. The challenge faced by the robot-team is to
coordinate their actions such that they efficiently explore the environment in their
search for victims.
A popular paradigm for the exploration problem is based on the notion of frontiers: the boundaries of the current map from where robots can enter yet unexplored area. Coordinating multiple robots is then about intelligently assigning frontiers to robots. Typically, the assignment of a particular frontier to a particular robot is governed by a cost measure, e.g. the movement costs for the robot to reach the frontier. In more recent approaches these costs are traded off with the potential gain in information if the frontier would be explored by the robot. In this paper we will further investigate the effect of balancing movement costs with information gains while assigning frontiers to robots. In our experiments we will illustrate how various choices for this balance can have a significant impact on the exploratory behavior exposed by the robot team. |
Stephen Balakirsky, Stefano Carpin, Alexander Kleiner, Michael Lewis, Arnoud Visser, Jijun Wang and Vitto rio Amos Ziparo, "Towards heterogeneous robot teams for disaster mitigation: Results and Performance Metrics from RoboCup Rescue", Journal of Field Robotics, volume 24(11-12):pp. 943-967, November 2007. (Portable Document Format 4.4 Mb)
| Urban Search and Rescue is a growing area of robotic research. The RoboCup Feder ation has recognized this, and has created a new virtual robots competition to c omplement its existing physical robot and agent competitions. In order to succes sfully compete in this competition, teams needed to field multi-robot solutions that cooperatively explored and mapped an environment while searching for victim s. This paper will present the results of the first annual RoboCup Rescue Virtua l League competition. It will present details on the metrics used to judge the c ompetition as well as summaries of the algorithms used by the top four teams. Th is will allow readers to compare and contrast these successful approaches. Furt hermore, the simulation engine itself will be examined and real-world validation results on the engine and algorithms will be presented. |
Arnoud Visser, "Measurement-Driven Simulation of Complex Engineering Systems", PhD thesis, Universiteit van Amsterdam, 23 October 2007 (Portable Document Format 102 Mb) .
|
he steadily increasing amount of traffic in the vicinity of their economical centers imposes great difficulties for most western countries. To reduce this steady increase, road pricing has proven to be an effective countermeasure. It has been introduced in different countries throughout the world, forcing people to consider alternative means of traveling. Since, for such measures to be effective, human behavior and social structures are deeply influenced, emotional political discussions have arisen.
Considering the importance of such social changes as well as the complexity of techniques involved in a project like "Rekening Rijden", the academic community, having public responsibility, should respond when asked for advice. So, the advice given was to initiate an evaluation study that would include modeling & simulation as well as extensive validation. Cooperating closely with our university, the Dutch government performed a study including all these components, resulting in a thorough overview of the reliability of the techniques, needed to build a system like "Rekening Rijden". The university remained involved in the project until the end, developing the models & methods necessary to carry out the study. If in the future road pricing systems will be introduced, the results & analyses from this research, as well as the models & methods developed, will still be of full importance. |
J. Sturm, P. van Rossum, A. Visser, "Panoramic Localization in the 4-Legged League: Removing the dependence on artificial landmarks" , in "RoboCup 2006: Robot Soccer World Cup X", (edited by G. Lakemeyer, E. Sklar, D. Sorrenti and T. Takahas hi), Lecture Notes on Artificial Intelligence series, volume 4434, p. 387-394, Springer, Berlin Heidelberg New York, October 2007. (Portable Document Format 727 Kb)
|
The abilities of mobile robots depend greatly on the performance of basic skills such as vision and localization. Although great progress has been made in the 4-Legged league in the past years, the performance of many of those approaches completely depends on the artificial environment conditions established on a 4-Legged soccer field.
In this article, an algorithm is introduced that can provide localization information based on the natural appearance of the surroundings of the field. The algorithm starts making a scan of the surroundings by turning head and body of the robot on a certain spot. The robot learns the appearance of the surroundings at that spot by storing color transitions at different angles in a panoramic index. The stored panoramic appearance can be used to determine the rotation (including a confidence value) relative to the learned spot for other points on the field. The applicability of this kind of localization for more natural environments is demonstrated in two environments other than the official 4-Legged league field. |
Arnoud Visser, Bayu Slamet, Tijn Schmits, Luis A. González Jaime, and Aksel Ethembabaoglu "Design decisions of the UvA Rescue 2007 Team on the Challenges of the Virtual Robot competition", Proc. of the Fourth International Workshop on Synthetic Simulation and Robotics to Mitigate Earthquake Disaster, July 2007, Atlanta, GA (pp. 20-26). (Portable Document Format 1.5 Mb)
| This discussion paper highlights the design decisions of the UvA Rescue Team on the challenges imposed by the 2007 competition. Consistent with the approach of the previous year, the UvA Rescue Team research focus is dedicated to perceptional issues, to allow later extensions towards complex multi-agent exploration. The Simultaneous Localization and Mapping (SLAM) approach is extended with additional scan matching techniques. Victim detection is made possible with Histogram based Skin Detection. |
Max Pfingsthorn, Bayu Slamet and Arnoud Visser, "A Scalable Hybrid Multi-Robot SLAM Method for Highly Detailed Maps", Proceedings CD of the 11th RoboCup I nternational Symposium, July 2007, Atlanta, USA. To be published in the Lecture Notes on Artificial Intelligence series. (Portable Document Format 1.8 Mb)
| Recent successful SLAM methods employ hybrid map representations combining the strengths of topological maps and occupancy grids. Such representations often facilitate multi-agent mapping. In this paper, a successful SLAM method is presented, which is inspired by the \emph{manifold} data structure by Howard et al. This method maintains a graph with sensor observations stored in vertices and pose differences including uncertainty information stored in edges. Through its graph structure, updates are local and can be efficiently communicated to peers. The graph links represent known traversable space, and facilitate tasks like path planning. We demonstrate that our SLAM method produces very detailed maps without sacrificing scalability. The presented method was used by the UvA Rescue Virtual Robots team, which won the Best Mapping Award in the RoboCup Rescue Virtual Robots competition in 2006. |
Arnoud Visser, Xingrui-Ji, Merlijn van Ittersum, Luis A. González Jaime and Laurentiu A. Stancu, "Beyond frontier exploration", Proceedings CD of the 11th RoboCup International Symposium, July 2007, Atla nta, USA. To be published in the Lecture Notes on Artificial Intelligence series. (Portable Document Format 2.2 Mb)
| This article investigates the prerequisites for a global exploration strategy in an unknown environment on a virtual disaster site. Assume that a robot equipped with a laser range scanner can build a detailed map of a previous unknown environment. The remaining question is how to use this information on this map for further exploration. On a map several interesting locations can be present where the exploration can be continued, referred as exploration frontiers. Typically, a greedy algorithm is used for the decision which frontier to explore next. Such a greedy algorithm only considers interesting locations locally, focused to reduce the movement costs. More sophisticated algorithms also take into account the information that can be gained along each frontier. This shifts the problem to estimate the amount of unexplored area behind the frontiers on the global map. Our algorithm exploits the long range of current laser scanners. Typically, during the previous exploration a small number of laser rays already passed the frontier, but this number is too low to have major impact on the generated map. Yet, the few rays through a frontier can be used to estimate the potential information gain from unexplored area beyond the frontier. |
A. Visser, J. Sturm, P. van Rossum, J. Westra and Th. Bink "Dutch Aibo Team: Technical Report RoboCup 2006", December 2006, (Portable Document Format 102 Kb)
| A method is developed which automatically learns the colors of the objects on the soccer-field under a wide variety of lighting conditions. The methods applies an expectation-maximization algorithm to cluster the colors in the YUV-colorspace, and identifies the colors clusters by applying a heuristic that makes use of the relative position of the cluster centers in the HSI-colorspace. Experimental results show that the absolute position of the clusters centers in the HSI-space show a large variety, but the relative position of the clusters is a robust criterion. |
Arnoud Visser, Jurgen Sturm and Frans Groen, "Robot companion localization at home and in the office", Proc. 18th Dutch-Belgian Artificial Intelligence Conference, BNAIC'06}, October 2006, Namen, Belgium (pp. 347-354). (Portable Document Format 629 Kb)
| The abilities of mobile robots depend greatly on the performance of basic skills such as vision and localization. Although great progress has been made to explore and map extensive public areas with large holonomic robots on wheels, less attention is paid on the localization of a small robot companion in a confined environment as a room in office or at home. In this article, a localization algorithm for the popular Sony entertainment robot Aibo inside a room is worked out. This algorithm can provide localization information based on the natural appearance of the walls of the room. The algorithm starts making a scan of the surroundings by turning the head and the body of the robot on a certain spot. The robot learns the appearance of the surroundings at that spot by storing color transitions at different angles in a panoramic index. The stored panoramic appearance is used to determine the orientation (including a confidence value) relative to the learned spot for other points in the room. When multiple spots are learned, an absolute position estimate can be made. The applicability of this kind of localization is demonstrated in two environments: at home and in an office. |
Max Pfingsthorn, Bayu Slamet, Arnoud Visser, and Nikos Vlassis, "UvA Rescue Team 2006; RoboCup Rescue - Simulation League", Team description paper for the 10th RoboCup International Competition, June 2006, Bremen, Germany. (Portable Document Format 288 Kb)
| The UvA Rescue Team will participate in the RoboCup Rescue Simulation League by employing a Simultaneous Planning, Localization and Mapping (SPLAM) approach. This approach augments traditional SLAM approaches with Planning in order to accomplish the set tasks. In this approach SLAM is modeled as one of the potential subgoals that may enable completion of the overall task. The benefit of this approach is that the quality of Localization and/or Mapping is regarded with respect to the tasks to achieve, which will gauge the optimization policy towards completing the task as a whole. Since victim detection is considered part of the exploration problem, we consider an approach based on the SPLAM paradigm as complete. |
A. Visser, P. van Rossum, J. Westra, J. Sturm, D.A. van Soest and M. de Greef, "Dutch AIBO Team at RoboCup 2006", Team description paper for the 10th RoboCup International Competition, June 2006, Bremen, Germany. (Portable Document Format 406 Kb)
| The Dutch AIBO Team is a multi-institute team which competes in the 4-legged robot league of RoboCup since 2004. This team description paper briefly outlines the approach taken for the Passing and Open Challenge of the RoboCup 2006. |
F.A. Oliehoek, A. Visser, "A hierarchical model for decentralized fighting of large scale urban fires" , Hierarchical Autonomous Agents and Multi-Agent Systems H-AAMAS'06, May 2006, Hakodate, Japan. (Portable Document Format 143 Kb)
| In this article we present a hierarchical model for planning and coordinating firefighting in situations as RoboCup Rescue, where an urban earthquake is simulated. We show that a hierarchical approach brings leverage to the planning process. By using formal decision theoretic models, we also present a more formal analysis of the RoboCup Rescue domain. Finally we discuss how our model could be applied and treat further abstractions that may be necessary. |
J. Sturm, A. Visser and N. Wijngaards, "Dutch Aibo Team: Technical Report RoboCup 2005", Technical Report, Dutch Aibo Team, October 2005. (Portable Document Form at 602 Kb)
| In this report we describe the algorithms implemented by the Dutch Team in the Sony 4-Legged League. In particular, we describe the new modules on the Aibo side (self-localization, role switching policies, vision) as well as some useful tools we created for the developer side (storing binary snapshots, improved color table creation and performance measurements). Our approach in the variable lighting challenge is explained, which allows to compensate for a wide range of varying lighting conditions. The range of lighting conditions during the challenge was too small to test our algorithm to the extreme, we expect that with this algorithm we can play soccer both indoors and outdoors. This component is still under ongoing development and we hope that next year we can demonstrate fully autonomous self-configuration under arbitrary lighting conditions. |
B. Slamet, A. Visser, "Purposeful perception by attention-steered robots" , Proceedings of the 17th Dutch-Belgian Artificial Intelligence Conference, BNAIC'05, October 2005, Brussels . (Portable Document Format 181 Kb)
| Designers of autonomous systems, embodied in an uncertain environment, have the tendency to build up a world model from everything that can be perceived. In contrast to this view, psychological researchers find for humans a selective interpretation of a scene, with phenomena like inattentional blindness [1]. Objects remain unseen if they are not central to the current behavior, even while they are clearly within view. Research in the Dutch Aibo Team [2] has proven that also for robots behavior-specific image processing can be very beneficial. In this article we design an experiment where we can not only indicate the appropriate moments to limit the perception to the objects relevant to the task, but also indicate the appropriate moments to release those limitations and to increase the overall situation. |
Niek Wijngaards, F. Dignum, P. Jonker, T. de Ridder, A. Visser, S. Leijnen, J. Sturm, S. Weers, "The Dutch AIBO Team at RoboCup 2005" , in Proceedings of the RoboCup 2005, July 2005, Osaka. (Portable Document Format 68 Kb)
| The Dutch AIBO Team is a multi-institute team which competes in the 4-legged robot league of RoboCup since 2004. Our team combines serious research with serious fun: collaborative autonomous (intelligent) systems are applied in Soccer, our shared application domain. This team description paper briefly outlines our current activities for RoboCup 2005. |
S. Oomes, P. Jonker, M. Poel, A. Visser, M. Wiering "The Dutch AIBO Team 2004" , in 8th International Workshop on RoboCup 2004, Lecture Notes on Artificial Intelligence, Springer Verlag, Berlin (to appear). (Portable Document Format 42 Kb)
| Several institutes in the Netherlands have joined forces and formed The Dutch AIBO Team. The team consists of a group of researchers and students from the universities of Amsterdam, Delft, Twente and Utrecht, and the DECIS Lab. Our goal is to stimulate research, teaching, and applications in the fields of artificial intelligence and collaborative robotics in the Netherlands by joining the international RoboCup community. |
A. Visser, G. Pavlin, S.P. van Gosliga, M. Maris. "Self-organization of multi-agent systems", Proc. of the International workshop Military Applications of Agent Technology in ICT and Robotics, The Hague, the Netherlands, 23-24 November 2004. (Portable Document Format 101 Kb)
| When one has to design multi-agent systems for realistic world applications one needs a certain level of self-organization to be able to cope with the dynamic environment. The self-organization can manifest itself in different aspects. It depends on the application on which aspect one should focus the research effort. We have identified which aspects we consider important, and translated this into research questions to be addressed in running and upcoming national and international projects. |
F.P. Terpstra, G.M. Meijer, A. Visser "Intelligent Adaptive Traffic Forecasting System using Data Assimilation for use in Travel Information Systems", submitted to the 1th IFIP International Conference on Artificial Intelligence Application and Innovations (AIAI 2004), Toulouse, France, 22-27 August, 2004. (Compressed PostScript 115 Kb) (Portable Document Format 321 Kb) (citeseer)
| In this paper we describe a novel application for data assimilation, a technique commonly used for atmospheric forecasting, in the area Traveler Information Systems. Our view is that the next generation of these systems will provide travel advise that employs not only the current traffic situation but also the future state of traffic. We show with actual traffic data that the impact of external influences (such as weather and large public events) are key to achieve accurate predictions. We propose an approach to model these influences in detail. Furthermore we describe how these detailed models can be combined with existing traffic models, using Data Assimilation. The results is an adaptive real-time system that can be used for intelligent Traveler Information Systems. We show our development and experimentation environment which offers solutions to the problems encountered in the development of the data assimilation system. Finally we discuss the progress and the challenges ahead. |
A. Visser, J.G. Zoetebier, H.H. Yakali, L.O. Hertzberger "An application for the Virtual Traffic Laboratory: calibrating speed dependence on heavy traffic", Workshop on Dynamic Data Driven Application Systems, Krakow, Poland, 6-9 June, 2004. (Compressed PostScript 159 Kb) (Portable Document Format 167 Kb) (citeseer)
| iN This paper we introduce an application for the Virtual Traffic Laboratory. We have seamlessly integrated the analyses of aggregated information from simulation and measurements in a Matlab environment, in which one can concentrate on finding the dependencies of the different parameters, select subsets in the measurements, and extrapolate the measurements via simulation. Available aggregated information is directly displayed and new aggregate information, produced in the background, is displayed as soon as it is available. |
A. Visser, J.G. Zoetebier, H.H. Yakali, L.O. Hertzberger "The measurement architecture of the Virtual Traffic Laboratory", International Conference on Computer Science (ICCS 2004), Krakow, Poland, 6-9 June, 2004. (Compressed PostScript 159 Kb) (Portable Document Format 167 Kb) (citeseer)
| In this paper we introduce the measurement architecuture of an application for the Virtual Traffic Laboratory. We have seamlessly integrated the analyses of aggregated information from simulation and measurements in a Matlab environment in which one can concentrate on finding the dependencies of the different parameters, select subsets in the measurements, and extrapolate the measurements via simulation. Available aggregated in formation is directly displayed and new aggregate information, produced in the background, is displayed as soon as it is available. |
A. Visser, S.B.M. Post, M.L. Fassaert, "The communication approach of the 'UvA Rescue C2003'-team" , in D. Polani, B. Browning, A. Bonarini, K. Yoshida (Eds.), RoboCup 2003, Lecture Notes on Artificial Intelligence, Springer Verlag, Berlin. (Portable Document Format 119 Kb)
| The 'UvA Rescue C2003'-team has been participating in the 'Rescue^M Simulation League' in the RoboCup competition in Padova. In this^M article we will explain our approach to cope with the limited^M communication between the centers and the agents in the disaster^M area. Our approach has two components. On the one hand, we^M introduce high-level communication: the agents will update the^M center with summaries of the current situation, the center^M aggregates the summaries to a global picture, synchronizes this^M picture with other centers, and redistributes this information^M back to the agents. On the other hand, each agent will build a^M common knowledge model together with the agents in its^M neighborhood. Based on the common knowledge, the agents are able^M to predict the behavior of their teammates, which enables the^M possibility to cooperate without explicit communicating their^M intentions. |
S.B.M. Post, M.L. Fassaert, A. Visser "The high-level communication model for multiagent coordination in the RoboCupRescue Simulator", in D. Polani, B. Browning, A. Bonarini, K. Yoshida (Eds.), RoboCup 2003, Lecture Notes on Artificial Intelligence, Springer Verlag, Berlin. (Portable Document Format 119 Kb)
| In this article we will concentrate on the communication problems in a multi-agent system, operating within the 'RoboCupRescue' Simulator system. To cope with the limited communication between the center and the agents in the field, we separate the communication in two layers that focus on synchronizing world models with different levels of detail, responsiveness and range. In this article we will explain the requirements and methods used in the high-level communication that distributes summaries of the current situation in different sectors of the map. |
M. Fassaert, S.Post, A. Visser "The common knowledge model of a team of rescue agents", 1th International Workshop on Synthetic Simulation and Robotics to Mitigate Earthquake Disaster, Padova, Italy, 6 July 2003. (Portable Document Format 119 Kb) (citeseer)
| In this article we will concentrate on the underlying knowledge model used in the design of a multiagent system, operating within the 'RoboCupRescue' Simulator system. To cope with the limitited communication between the center and the agents in the field, we introduce teams with an information manager. One of the agents, the information manager, gets the task to update its center with summaries, enriching its own model with information from other teams. Based on the common knowledge, the agents are able to predict the behavior of their teamates, which enables the possibillity to cooperate without explicit communication. As basis of the common knowledge model we use sectors, a represantation that is worked out in more detail in this article. |
F.P. Terpstra, G.M. Meijer, A. Visser "Information discovery and combination from divergent dat sources for Travel Information Systems", submitted to the 7th European Conference on Principles and Practice of Knowledge Discovery in Databases (ECML/PKKD-2003), Cavtat-Dubrovnik, Croatia, 22-26 September 2003. (Compressed PostScript 115 Kb) (Portable Document Format 321 Kb) (citeseer)
| In this paper we describe the development and use of knowledge discovery processes in the area Traveler Information Systems. Our view is that the next generation of these systems will provide travel advise that employs not only the current traffic situation but also on the future state of traffic. We show with actual traffic data that the impact of external influences (such as weather and large public events) are key to achieve accurate predictions. We propose an approach to model these influences in detail and describe an information discovery and combination process supporting this approach. It will combine real-time and historical information on traffic and its influencing factors by means of correlation of data clustered in both space and time. We show our development and experimentation environment which offers solutions to the problems encountered in the development of the information discovery and combination process. Finally we discuss the progress and the challenges ahead. |
S.B.M. Post, M.L. Fassaert, A. Visser "Reducing the communication for multiagent coordination in the RoboCupRescue Simulator", 7th RoboCup International Symposium, Padua, Italy, 10-11 July 2003. (Portable Document Format 119 Kb) (citeseer)
| In this article we will concentrate on two problems that are important for the design of a multiagent system, operating within the 'RoboCupRescue' Simulator system. To cope with the limitited communication between the center and the agents in the field, we introduce teams with an information manager. The team will build a common knowledge model. The manager will update the center with summaries, enriching its own model with information from other teams. Based on the common knowledge, the agents are able to predict the behavior of their teamates, which enables the possibillity to cooperate without explicit communication. |
A. Visser, A. van Inge, F.C.A. Groen "Experiences with the innovation of an Autonomous Systems Course", 9th International Conference of European University Information Systems, Amsterdam, The Netherlands, 2-4 July 2003. (Compressed PostScript 98 Kb) (Portable Document Format 98 Kb) (citeseer)
| The course on the Organisation and Design of Autonomous Systems is given at the University of Amsterdam since 1995. Every year there were innovations, which could be on the content, the pedagogic approach, or the on-line support. This year we had two novelties: the migration of the on-line support to the Blackboard environment and the introduction of a project assignment instead of a practicum and a writing assignment. In this article we will summarize our experiences. |
A. Visser, I.E.M. Breuker, S.J. Huiberts "Advanced Logistics Information Exchange - 'Hoe verbeteren we de dynamiek bij het vergelijken van logistieke diensten?'", Vervoerslogistieke Werkdagen, Corsendonk, Belgium, November 2002. (Compressed PostScript 457 Kb) (Portable Document Format 585 Kb)
| In het project 'Advanced Logistics Information Exchange' (ALIE) worden de internetsites van logistieke bedrijven met geavanceerde en flexibele webtechnologie met elkaar verbonden en vergeleken. ALIE is een omgeving waar gemakkelijk de vertaling gemaakt kan worden van logistieke diensten op de ene site naar de omschrijving van logistieke diensten op een andere site. Voor vertaalslagen met oplopende complexiteit kan men daarbij denken aan hulpmiddelen als stylesheets, java-applets, web-services en intelligent agents. Een van de uitdagingen van het project, en een aardige discussie voor de logistieke werkdagen, is de vraag hoe generiek deze vertaalslagen en hulpmiddelen zijn. |
A. Visser, H. Kok, M. Bergman, H.H. Yakali, L.O. Hertzberger, "Calibration of a traffic generator for high-density traffic, using the data collected during a road pricing project", paper 4052 to the 9th World congress on Intelligent Transport Systems, Chicago, Illinois, October 2002. (Compressed PostScript 484 Kb) (Portable Document Format 529 Kb)
| Microscopic traffic simulation is often used to investigate phenomena that can not be measured in real traffic. In our case, the Dutch Government has employed such a simulation to investigate the performance of several prototype road pricing systems before they were actually implemented on a road site. In the end, only a few number of prototypes were actually tested on real traffic. During these tests, a separate system recorded not only their outputs, but also the measurements of a reference system. This was an inductive loop system, which is the standard source of most validation sets for traffic simulation. It gave the opportunity to combine two independent measurements that resulted in a data set unique in its reliability and detail. |
A. Visser, H.H. Yakali, A.J. van der Wees, M. Oud, G.A. van der Spek, L.O. Hertzberger, "The reliability of a DSRC-link for ETC applications modelled on 5 levels of detail", IEEE 5th International Conference on Intelligent Transportation Systems, September 3-6 2002, Singapore. (PostScript 2.9 Mb) (Portable Document Format 1.4 Mb)
| For electronic payments, the communication link has to be reliable. Dedicated short range communication is a proposed solution for the automatic debiting of vehicles without disturbing the traffic flow. The requirements on the reliability of such a system are high, which implies that only large scale simulations with a lot of detail are effective to analyse an occasional error. In this article, an hierarchical approach is worked out that allows such simulations of the communication link with a 80% reduction of the computational effort compared to simulation with full detail. |
A. Visser, H.H. Yakali, A.J. van der Wees, M. Oud, G.A. van der Spek, L.O. Hertzberger, "An hierarchical view on modelling the reliability of a DSRC-link for ETC applications", IEEE Transactions on Intelligent Transportation Systems Vol. 3: No. 2, June 2002, (PostScript 26 Mb) (Portable Document Format 1.4 Mb) (citeseer)
| For electronic payments, the communication link has to be reliable. Dedicated short range communication is a proposed solution for the automatic debiting of vehicles without disturbing the traffic flow. The requirements on the reliability of such a system are high, which implies that only large scale simulations with a lot of detail are effective to analyse an occasional error. In this article, an hierarchical approach is worked out that allows such simulations of the communication link with a 80% reduction of the computational effort. |
F. Terpstra, A. Visser, L.O. Hertzberger, "An On-Line Planner for MARIE", Proceedings of the 12th Irish Conference on Artificial Intelligence and Cognitiv e Science (AICS2001), NUI Maynooth, September 2001, p. 199-209. (Compressed PostScript 1.9 Mb) (Portable Document Format 401 Kb) (citeseer)
| In this paper we introduce an on-line planner implemented on MARIE, a real autonomous robot. Planning occurs at a high level, the planner controls the other modules such as the pathplanner which contain planners themselves. Plans are represented as a tree structure. The on-line planner has a library of partial trees to choose from when replanning is needed. We think that our research is unique in applying on-line planning on such high level on a real robot. The on-line planner uses a simple algorithm based on elements of POP in combination with random choice. The random choice is essential for real world applications, with their information errors, bu preventing the robot to take repetitively the 'best' choice. Despite being a simple algorithm it does perform very well in practice. Tests show that the on-line planner gives a major increase in robustness and reliability while not being significantly slower than a perfect algorithm would be. |
H. Afsarmanesh, R. Belleman, A. Belloum, A. Benabdelkader, J.F.J. van den Brand, T.M. Breit, H. Bussemaker, G. Eijkel, A. Frenkel, C. Garita, D.L. Groep, A.W. van Halderen, R.M.A. Heeren, Z.W. Hendrikse, L.O. Hertzberger, J. Kaandorp, E.C. Kaletas, V. Klos, P. Sloot, R.D. Vis, A. Visser, H.H. Yakali, "VLAM-G: A Grid-Based Application Environment", Scientific Progamming 10 (2002) 173-181. (Portable Document Format 110 Kb) (citeseer)
| The VLAM (Virtual Laboratiory AMsterdam) provides a science portal for distributed analysis in applied scientific research. It offers scientists a familiar environment to carry out their experiments, where content and data are clearly separated. Emphasis is put on the development and use of open standards and seamless integration of external devices. In order to setup a testbed, we have applied several recent technologies and simulation techniques. In this paper the proposed design and its prototype implementation are presented. Moreover, some domain specific case studies are described in more detail. |
M. Hommes, A. Visser, L.O. Hertzberger, "VriendenOV, de gebruiker als bron van dynamische reisinformatie", Verkeerskunde 3/2001 (ANWB, Den Haag), March 2001. (PostScript 9.8 Mb) (Portable Document Format 5.3 Mb)
| This article (in Dutch), describes an dynamic travelinformation system we have designed and build. The unique feature of this system is that the dynamic information is not delivered by the travelcompanies, but by the travellers themselves. The traveller only needs a WAP-enabled telephone, although a system with a GPS is much user-friendly. |
A. Visser, A.J. van der Wees, and L.O. Hertzberger, "Discrete Event Modelling Methodology for Intelligent Transport Systems", paper 2016 of the proceedings of the World Congress on Intelligent Transport Systems, Torino, Italy, November 2000. (PostScript 428 Kb) (Portable Document Format 67 Kb) (citeseer)
| Motivated by work on Electronic Fee Collection, we formulate a general simulation concept to evaluate the performance of Intelligent Transport Systems, and introduce the virtual sensor concept as a concept for sensor system modelling. We demand that interaction between elements of the system is well defined, and that their state changes at pre-definable time intervals. As a result we can model the total system as a discrete event model. |
A. Visser, H.H. Yakali, A.J. van der Wees, M. Oud, G.A. van der Spek, L.O. Hertzberger, "An hierarchical view on modelling the reliability of a DSRC-link for ETC applications", Technical Report CS-99-02, IEEE Transactions on Intelligent Transportation Systems Vol. 3: No. 2, June 2002, (PostScript 26 Mb) (Portable Document Format 1.4 Mb) (citeseer)
| For electronic payments, the communication link has to be reliable. Dedicated short range communication is a proposed solution for the automatic debiting of vehicles without disturbing the traffic flow. The requirements on the reliability of such a system are high, which implies that only large scale simulations with a lot of detail are effective to analyse an occasional error. In this article, an hierarchical approach is worked out that allows such simulations of the communication link with a 80% reduction of the computational effort. |
L. Dorst, A. Hoekstra, J.M. van den Akker, J. Breeman, F.C.A. Groen, J. Lagerberg, A. Visser, H. Yakali, L.O. Hertzberger, "Evaluating Automatic Debiting Systems by modelling and simulation of virtual sensors", IEEE Instrumentation and Measurement Magazine, June 1998, vol.1, no.2, pp. 18-25. (PostScript 318 Kb) (Portable Document Format 154 Kb) (citeseer)
|
An ADS (Automatic Debiting System) is an electronic fee collection (EFC)
system on a freeway, which interacts with a traansponder in each car
and subtracts a fee from a credit card.
Non-payers are to be photographed and fined.
In requirements on such systems in The Netherlands,
privacy laws demand separation between the financial transaction
and the registration; the system should not be usable for tracking citizens.
There are also very strict demands on performance, with
error rates of the order of 10^-5 to minimize the number of incorrect
enforcements. Both demands combined require complex ADS systems.
This paper describes modelling and simulation of the proposed ADS's of contractors, to be used for their quantitative evaluation. It is highly important to choose the proper level for the abstraction modelling of the simulated system, both to keep the simulation within reasonable cost and time, and also to be able to validate that the simulated system indeeds represents the actual proposal. In this paper, we show how a consistent application of the concept of virtual sensors can be applied to clarify the choice of the proper level of abstraction for the ADS proposals, and its implementation. It leads to a simulation that is faster than a full implementation of the ADS, but no less specific in its quantitative results. |
A.G. Hoekstra, L. Dorst, M. Bergman, J. Lagerberg, A. Visser, H. Yakali, F. Groen and L.O. Hertzberger, "Modelling and simulation of automatic debiting systems for electronic collection on motor highways", proceedings IASTED International Conference Applied Modelling and Simulation, Ed M.H. Hamzam pp 104-108, 1997. (PostScript 1.3 Mb) (Portable Document Format 270 Kb) (citeseer)
| We have developed a Modelling and Simulation platform for technical evaluation of Automatic Debiting Systems (ADS) for Electronic Toll Collection on Motor Highways. This platform is used in a project of the Dutch government to assess the technical feasibility of Toll Collection systems proposed by industry. The modelling approach and the simulation environment (ADS-SIM) will be presented. As a case study, an example of modelling and simulation of a fictitious ADS will be shown. Finally, the need for High Performance Computing to carry out the simulations is discussed. |
L. Dorst, F.C.A. Groen, J.M. van den Akker, J. Breeman, A. Hoekstra, J. Lagerberg, A. Visser, H. Yakali, L.O. Hertzberger, "The virtual sensor concept applied to simulation: the evaluation of Automatic Debiting Systems", Proceedings of 1997 Workshop on Emergent Technologies & Virtual Systems for Instrumentation and Measurements, Niagara Falls, Ontario, Canada, May 15-17, pages 118-124, 1997. (PostScript 358 Kb) (Portable Document Format 175 Kb)
|
An ADS (Automatic Debiting System) is a toll-collecting system on a freeway,
which interacts with a transponder in each car, and subtracts a fee from a credi
t card. Non-payers are to be photographed and fined.
In requirements on such systems in The Netherlands,
privacy laws demand separation between the financial transaction and the registr
ation; the system should not be usable for tracking citizens.
There are also very strict demands on performance, with
error rates of the order of 10^-5 to minimize the number of incorrect enforc
ements. Both demands combined require complex ADS systems.
This paper describes a simulation of the proposed ADS's of bidders, to be used for their quantitative evaluation. It is highly important to chose the proper level for the abstraction of the simulated system, both to keep the simulation within reasonable cost and time, a nd also to establish that the simulated system indeeds represents the actual pro posal. In this paper, we show how a consistent application of the concept of {\em virtu al sensors} can be applied to clarify the choice of the proper level of abstract ion for the ADS proposals, and its implementation. It leads to a simulation that is faster, but no less specific in its quantitativ e results than a full implementation of the ADS. |
| We have developed a Modelling and Simulation platform for technical evaluation of Electronic Toll collection on Motor Highways. This platform is used in a project of the Dutch government to assess the technical feasibility of Toll Collection system proposed by industry. Motivated by this work we introduce the concept of High Performance Discrete Event Simulations to evaluate the performance of Complex Industrial Systems. We assume that the elements have well-known behaviour, and demand that the interaction between the elements is well-known and their state changes at pre-definable time intervals. As a result we can model the total complex system as a discrete event model. We present the case of Automatic Debiting Systems for Electronic Toll Collection and introduce the modelling approach, the simulation environment (ADS-SIM), and discuss the need for HPC to carry out such simulations. |
| In autonomous control systems the sensor-actuator control loop plays a dominant role. It is essential to give the system the capability to react to changes in its environment. A flexible autonomous system needs to reconfigure its sensor-actuator control loops depending on its surroundings, the task at hand, and the available sources of information. In the system architecture of the MARIE vehicle ther is a clear difference between control and data flow. In this article we describe our general approach to realise the data interfaces between the several sensing- and actuation-modules. Information users can dynamically select their sources and monitoring is transparently supported. Also we describe how the modules can be controlled (initialised, parameterised, activated, suspended) from a higher abstraction layer through the so called elementary operation interface. |
|
Currently, much production time is lost in teaching and re-teaching
robot programs. New techniques that avoid this, such as off-line
programming, require an accurate kinematics model of the robot.
To construct this model, measurements of the robot are required. At the University of Amsterdam a self-calibrating measuring system has been developed, using a camera mounted in the robothand and a reference object. Analysis of a single image of the reference object allows us to obtain an estimate of the position and orientation of the viewing point. A much better estimate is obtained by analysing a series of images, as the essential geometrical and optical camera parameters can be derived from the redundancy in the measurements. Our experimental results show that we can compute the pose of the camera for the different images with an accuracy in the order of 0.15 mm and 1.5 arc minute. This is usually adequate for robot calibration. |
|
To maintain robot accuracy, calibration equipment is needed. In this
paper we present a self-calibrating measuring system based on a camera
in the robot hand plus a known reference object in the robot workspace.
A collection of images of the reference object is obtained. Using
image-processing, image-recognition and photogrammetric techniques,
the positions and orientations of the camera are computed. The
essential geometrical and optical camera parameters can be
derived from the redundancy in the measurements. From each
image the positions of markers on the reference object are
extracted and the individual markers are identified. The camera
positions for all images plus the parameters of the camera are
solved together in a non-linear least-squares fitting procedure. Experimental results for this low-cost measuring system are presented. |
|
At the University of Amsterdam a measuring system has been built that can
be used for calibration of robots, using a camera mounted in the robot
hand. The system uses the fact that an object looks different from a
different viewing angles. Analysis of an image of a known reference
plate from a certain viewpoint can result in an estimate of the location
of that viewing point. But for an accurate estimate a precise model of
the imaging process has to be made. In this presentation we will make a short tour inside the camera, and will show how a marker on the reference plate is imaged onto the light-sensitive elements of the camera. With the help of the redundant information available in a series of images we can make precise estimates of the camera properties, and of the locations of the different viewpoints. At the end we will show that we can predict the position of a marker in the image with sub-pixel accuracy, in spite of several imaging errors greater than a pixel. The location of the camera for the different images is then known with an accuracy in the order of 0.1 mm and 1 arcminute which is adequate for robot calibration. The work reported here was done for the ESPRIT project "CAR". |
| The exception-handling framework described in this paper creates an opening to compare different exception-handling approaches in a structured way. This comparison is made, linkin taxonomies of different research-groups together. Concurrently the framework specifies a general data-structure to store knowledge about exception-handling, which makes it easier to adapt the proposed taxonomy in the implementation of existing and impeding work-cell controllers and production planners. |
|
In this article we have presented a low-cost method, based on photogrammetry,
to obtain measurements for the calibration of robot systems. The
measuring system is self-calibrating. The method has been implemented and tested and provides promising results for practical application. The components used are relatively inexpensive, and can easily be combined to yield a portable system. As most of the data processing has been highly automated, such a system will be usable by non-expert personnel. By combining the video camera with a fast frame grabber and recording system, dynamic measurements should be obtainable. The relative locations and orientations of two robots in a work cell can be found by placing the reference plate between the robots and calibrating both robots with that common reference. |
|
Measurement and calibration playan increasingly important role in
robot production as well as in robot implementation and operation
within computer-integrated manufacturing or assembly systems.
The production, implementation and operation of and with robots and
multi-axis machine tools are all areas where robot-calibration results
can lead to significant accuracy-improvements and/or cost-saving
opportunities. Presented are robot calibration tools for robot modelling, model parameter identification and robot measurement which can be used for evaluating quality control of robot production accuracy as required for use of off-line programming. |
|
In the frame of the ESPRIT project CIM-PLATO a toolbox of planning
tools for CIM applications has been developed. The planning tasks
addressed by the various tools span a wide range of CIM activities
for the design, planning and installation of Flexible Manufacturing
Systems (FMS) and Flexible Assembly Systems (FAS) in a CIM
environment. By selecting individual tools, a customized toolbox
can be configured for a specific CIM problem. The tools
in such a toolbox can be integrated via an Information Management
System. In the development of the tools and the Information System
existing and forthcoming standards have been taken into account:
STEP, the standard for exchange of product model data, EXPRESS,
the specification language used in STEP, and OSF/MOTIF for the
design of the user interfaces. Apart from the development
of the individual tools an important goal of the project is
to define a framework for an open toolbox concept in which
new tools can be easily integrated. This paper presents O->M, one of the application specific toolboxes implemented by the CIM-PLATO consortium. O->M addresses the main planning activities for product configuration and manufacturing of a computer system selected from the Bull catalog. It consists of five tools from the CIM-PLATO toolbox and an Information System. O->M demonstrates the adaptability of the tools involved to specific applications; their integratability, which is achieved by the use of the common Information System based on the standards for data exchange, and the flexibility of the CIM-PLATO toolbox approach which yields flexible solutions with different levels of automation. |
|
In this report the ORCA scheme to control an A&R system on a high level
was adapted to the Columbus Automation & Robotic Testbed.
A scheme, consisting of an Environment Model, a Robot Program Model,
a Control Model and a Exception Handling Model, existed for REXSY
testbed at the ERNO laboratory in Bremen. Adaptations
were necessary in the Environment Model to represent relationships
between locations, and to cope with the large amount of active
components in the C.A.Testbed. The Robot Program Model is made
flexible for the number of abstraction levels and the level
on which executable activities are sent to the devices. Further
modefications were made in the Control Model, to distribute
the authority to prepare and execute activities between different
users. In the Exception Handling Model it was high-lighted that the
activities of the CAT-robot were compensatory, so that backward
recovery was possible as long as the samples were not modified
in the payloads. The result is a FAIR functional architecture, in which the control levels of too high detail and too low abstraction can be assigned to experienced users or autonomous control layers, so that humans without profound knowledge of robotics can use a robot to assist them with their experiences in space. |
|
In this document the requirements for reliable, low-cost measuring
systems for partial calibration. Such systems should be portable,
robust and easy to operate. A selection of sensing techniques
is described. The selected measuring device designs for the CAR project are discussed: a camera-based system and the four-beam system. A system based on a camera and space resection has the advantage that it is easily applicable and precise positioning of the camera is not needed. The disadvantage, however, is that the required accuracy can obly be approached. The four-beam system has the advantage that a high accuracy can be attained and that the system does not have expensive components. For the four-beam system, however, the beams have to be centered very precisely on the PSD-s, which means initial calibration is needed to determine the model parameters with suitable accuracy. A technique based on the singular value decomposition is used to analyze the geometry of the measuring systems, This technique appears very helpful for the evaluation of the measuring systems. It can be used to find out which combinations of system parameters can be identified and to what accuracy, given a particular set of measurements. It also shows which combinations of system parameters cannot be determined and in this way also helps in selecting additional measurements and calibration measurements for the measuring system itself. Initial experiments results are presented for a camera-based system. The results are promising in view of the quality of the prototype. The demonstration will be extended to fully automized test-bed, which show the feasability of this approach as ease to use and low cost calibration system. |
|
In the study IRAS the design of a general control system for symbiotic,
interactive remote A&R servicing has been investigated and developed.
The architecture of the IRAS control concept supports a flexible
allocation of responsibilities between man and machine, where
the machine provides the sensory and manipulative capabilities as well
as some planning functions, while the human can contribute with his
superior reasoning capabilities. The interaction between the
user and IRAS can either be initiated by the user or by IRAS. If
a particular problem has to be solved, IRAS will activate the appropiate
planning function, which will either be carried out completely
automatically or with the support of the operator. If the operator
is involved th planning functions will ensure that all informations
which are relevant for the operation will be presented to the user.
Under all circumstances the operator keeps complete control over
the system and can switch at any time from automatic to step
by step execution of activities. This concept of interactive control has been established as a prototype software system. The implemenation includes all different planning methods for the operation preparation as scheduling from precedence relations to sequence of TASKS as well as ACTION pattern selectin for TASK decomposition or parameter planning for the transformation of the implicite defined ACTIONS to executable elementary operation commands. The implemented concept of exception handling incorporates as envisaged autonomous system provided monitoring of activity specific sensor primitive conditions. In case of violation of conditions the systems prompts immediate response by stopping the further execution, adequate information of the operator and, if desired, support in diagnosis of the cause of the anomaly. Recovery to nominal conditions is performed either by predefined recovery activities or by online planned, verified and executed recovery operation sequences. The established prototype proves well as a platform for further interactive remote operations planning and execution investigations. The system allows detailed investigations for the allocation of responsibilities between man and machine even respecting different potential user groups. |
|
This document describes an Orthogonal Robot Control Architecture, ORCA,
for modeling and controlling the execution of robot activities. It is
called orthogonal since it supports the orthogonal design of a complex
robot control system, that can include various forms of automatic
planning, operator controlled intervention and reactive behavior
to contingencies. To use this architecture for space robotics applications one has to define the environment model, the elementary actions that a robot can execute, the functions (if possible, otherwise it will depend on human intelligence) that relate parameters of tasks with parameters of subtasks, possible exceptions and their relationships to observations, and recovery tasks. |
| 1 visitors in February 2012 | Created in November 1993 | arnoud@science.uva.nl |