You can get different views of this calendar at this page, or you can download or subscribe to it from your calendar application (http, webCal, 14KB).

Summer School Schedule
When Su, 17th Mo, 18th Tu, 19th We, 20th Th, 21st Fr, 22nd
8.00 - 9.00 Arrival Breakfast
9.00 - 9.15 Opening Address Talk #5:
N. Neogi,
Integrating UAVs into Civilan Airspace

Reading Groups #9:
P. Santi
Topology Control

L. Parker
Dynamic Task Allocation in Heterogeneous Multi-Robot Teams

Talk #11:
C. Woolsey
Using Flight Vehicles for Control Volume Sampling

Talk #16:
C. Woolsey
Energy Shaping for Mechanical Systems

Talk #1:
L. Parker,
Performance Evaluation and Benchmarking in Multi-Robot Teams

10.30 - 11.00 -break-
Talk #2:
N. Neogi
Fault Detection and Diagnosis Techniques for Networked UAVs

Talk #6:
A. Redfern
Pervasive Computing in the Every Day World

Talk #10:
D. Stilwell
Control and Estimation Over Networks, Part 1

Talk #12:
W. Burgard
Coordinated Multi-Robot Exploration

Talk #17:
F. Mattern
Towards the Internet of Things

Tutorial #3:
P. Santi
Topology Control in Wireless Multi-hop Networks, Part 1

Tutorial #7:
L. Parker
Dynamic Task Allocation in Heterogeneous Multi-Robot Teams

Social Event:

Tutorial #13:
A. Redfern
Pervasive Computing Application Development, Part 1

Talk #18:
D. Stilwell
Control and Estimation Over Networks, Part 2

15.30 - 16.00 -break- -break- Closing Remarks
Tutorial #4:
P. Santi
Topology Control in Wireless Multi-hop Networks, Part 2

Students' Session #8:
Sports Event

Tutorial #14:
A. Redfern
Pervasive Computing Application Development, Part 2

17.30 - 18.00 -break- -break-
Dinner Talk #15:
F. Mattern:
Wonderful Future

19.00 - 19.30 Barbecue
Last evening party

Talk about
the Castle
Social Event:
Students' Session #8:
Posters and Demos
Social Event:



Pervasive Computing in the Every Day World (Andrew Redfern, ~1.5h)

Firefighting is an extremely demanding occupation set in a chaotic environment where lives and buildings are at stake. Firefighters must make quick decisions with little information and divide their attention between many events, making it difficult to efficiently and effectively complete critical tasks such as building search and rescue. The Fire Information and Rescue Equipment (FIRE) project at UC Berkeley addresses these challenges by applying and designing new technologies, such as wireless sensor networks (WSNs) and small head-mounted displays (HMDs), to firefighting. In this talk, I will highlight three core components of FIRE: (1)SmokeNet, (2)FireEye, and (3)eICS. The design of each of these components will be discussed and I will demonstrate how user interviews, observations of firefighter tactics and prototypes have help to refine these three main subsystems of FIRE. FIRE is just one example of pervasive computing, in addition to the FIRE system I will introduce a few other pervasive computing applications including some industrial monitoring systems, interactive toys, and another application of people tracking.

Control and Estimation Over Networks: Part 1 (Dan Stillwell, ~1h)

Graph-theoretic tools have proven useful for analysis and design of systems that operate over networks. In this talk, we introduce several basic graph-theoretic concepts and discuss how they apply to analysis of control and estimation problems over networks. We specifically address the case of time-varying network topologies and consider the case of both deterministic and stochastic switching between topologies. Network analysis concepts are illustrated via analysis of oscillator synchronization, consensus algorithms, and data fusion algorithms.

Control and Estimation Over Networks: Part 2 (Dan Stillwell, ~1h)

In this talk, we discuss control and data fusion architectures for mobile sensor networks that are based on either information states or full-state observers. We describe a general architecture and show that no Kalman separation principal exists. Thus we conclude, in general, that data fusion and control algorithms are coupled. By stating the problem in general, we are able to recognize specific cases for which there exists no coupling, and we explore two cases in detail. At Virginia Tech, we are using this general framework for developing data fusion and control algorithms that enable a team of autonomous underwater vehicles to cooperatively find an acoustic target. We use this example throughout the talk for motivation and illustration.

Coordinated Multi-Robot Exploration (Wolfram Burgard, ~1.5h)

In this presentation we consider the problem of exploring an unknown environment by a team of robots. As in single-robot exploration the goal is to minimize the overall exploration time. The key problem to be solved therefore is to choose appropriate target points for the individual robots so that they simultaneously explore different regions of their environment. We present an approach for the coordination of multiple robots which simultaneously takes into account the costs of reaching a target point and the utility of target points. The utility of target points is given by the size of the unexplored area that a robot can cover with its sensors upon reaching a target position. Whenever a target point is assigned to a specific robot, the utility of the unexplored area visible from this target position is reduced for the other robots. This way, a team of multiple robots assigns different target points to the individual robots. The technique has been implemented and tested extensively in real-world experiments and simulation runs. We will present experimental results which demonstrate that our coordination technique significantly reduces the exploration time compared to previous approaches. We also will describe different extensions of this approach towards limited communication and limited bandwidth of the communication channel. Additionally, we will describe how to utilize high-level information about the type of places to improve the exploration process.

Towards the Internet of Things (Friedemann Mattern, ~2h)

The term "Internet of Things" has come to describe a number of technologies that enable the Internet to reach out into the real world of physical objects. In fact, the trend of information and communication technology seems to be clear: The ongoing miniaturization of electronic devices enables processors and sensors being embedded into more and more everyday things - not only electrical devices, but also very mundane things such as key chains or even clothes. Many of these devices will then be interwoven and connected together by wireless networks.

A world full of smart things that may communicate with each other and that interact with global services sounds fascinating, giving rise to many new applications and business opportunities. But how realistic are the promises? To approach this question, we will first elaborate the vision of the "Internet of Things" and summarize its enabling technologies. We take a broad view and identify long-term trends which, by extrapolation, give us some hints on what to expect in the future. We then discuss the main challenges and analyze what progress is necessary to overcome current technological hurdles.

In our talk we will also present several applications and prototypes of cooperating real-world objects that have been realized at ETH Zurich, and we shall briefly discuss the social and economic consequences and challenges of a future world pervaded by invisible computers.

Wonderful Future - Early Predictions of (Information) Technology (Friedemann Mattern, ~1h)

As is well known, making predictions is a bit tricky. A hundred years ago, a man took a cautious view of our time and predicted something wonderful: the mobile phone. This would allow not only monarchs and chancellors to run their businesses from a distance, but also the happy time for love would begin - because the couples would always know what the partner would be doing. The past technology forecasts promised still many other fantastic things - teaching machines replace teachers, color fax machines and screen phones to be found in every home, and household robots doing the dishes and serving coffee. Only the Web, E-commerce, search engines, SMS, game consoles, blogs, Ebay, camera phones,… that is, all the blessings of the information age which did not exist 15 years ago, whose name had not even been invented, but without which we could barely live today, would virtually seem not having being predicted by anyone! Or did the unrecognized prophets of the digital age exist?

Performance Evaluation and Benchmarking in Multi-Robot Teams (Lynne Parker, ~1h)

A common challenge when comparing and contrasting alternative multi-robot solutions is the selection of appropriate metrics and application domains that enable the generation of meaningful comparisons. Many metrics are application-specific, and, while useful for a narrow application domain, are difficult to generalize across other domains. Ideally, we would like to be able to identify more general metrics that can provide some predictive performance when applying a multi-robot solution to a new task, or in a new domain. Perhaps the use of benchmarks can help standardize the comparisons of robotic solutions. For example, one benchmark popularly used in the multi-robot domain is robotic soccer, where the measure of effectiveness of the robotic solution is the win-loss record of the robot team. However, a single benchmark is unlikely to adequately evaluate robot systems along all the characteristics of interest. This talk will explore these issues of performance evaluation and benchmarking in the context of multi-robot teams. We will discuss many techniques used to date, along with new developments that may be needed to provide better evaluation capabilities for multi-robot systems.

Using Flight Vehicles for Control Volume Sampling (Craig Woolsey, ~1h)

Unmanned air vehicles (UAVs) and autonomous underwater vehicles (AUVs) are often used to assess phenomena that occur within a sample volume by measuring the flux of related quantities or properties across the volume’s boundary. This presentation will describe example applications in this class of sampling tasks and related work on vehicle control and coordination for autonomous underwater gliders and UAVs.

Energy Shaping for Mechanical Systems (Craig Woolsey, ~1h)

The operating envelope for an autonomous vehicle can often be expanded through clever control design that respects and exploits the vehicle’s natural, nonlinear dynamics. Energy-based control design can provide controllers which are energy efficient, robust to model uncertainty, and perform well over a large operating envelope. This presentation will discuss methods for controlling mechanical systems using feedback that reshapes the system’s total energy and its dynamic structure.

Fault Detection and Diagnosis Techniques for Networked UAVs (Natasha Neogi, ~1h)

The seminar outlines a software oriented approach for the real-time detection of faults and byzantine behaviour in multi-modal hybrid systems. Online model generation of continuous variables enables the detection of multiple faults, while managing computational complexity. As an example, the aileron control loop and propulsive system of a UAV using a Piccolo Plus autopilot are examined. A theoretical upper bound for the detection threshold for the ”stuck aileron” fault is derived. This enables the dynamic selection of the optimal threshold, in terms of false alarms, over a range of maneuvers in an online manner. The development of several independent models for the calculation of the propulsion system residual that use diverse and redundant data sources, allows for the diagnosis of several distinct faults via multiple thresholds on a single residual in a concurrent fashion. A conflict detection and positioning algorithm for a network of UAVs is outlined, in the presence of byzantine agents, and bounds for k-quality consensus are developed.

Integrating UAVs into Civilan Airspace: Security of Networked Autonomous Agents (Natasha Neogi, ~1h)

A review and analysis of uninhabited aerial vehicle (UAV) accident data has been conducted to identify important human factors issues related to their use. Classification of accident data into primal causative factors, denoted by categories such as human factors, maintenance, aircraft integrity, as well as unknown factors, can be used as a first pass for analysis. Each category can be further elaborated upon by examining the accidents in detail, and creating a set of likely subdivisions, for example, accidents caused due to human factors related issues can be related to display design, procedural errors, skill-based errors etc. A critical issue lies in the vast array of differences in the currently fielded platforms. Since UAVs span a range of sizes and weights, from microvehicles to fully functional combat fighter aircraft, the related accidents and incidents are inextricably intertwined with their functional missions and payloads. Furthermore, failure modes and path-to-failure are also conditioned upon these key factors. Nonetheless, many of these accident failure modes can be anticipated by a throrough analysis of user interfaces employed and procedures implemented for their use. This creates a basis for the qualities inherently necessary for certifiable UAV platforms to enable their integration into civilian airspace.


Hands-on Introduction to Pervasive Computing Application Development (Andrew Redfern, ~3h)

Building embedded application has been traditionally difficult because different toolsets must be used for various aspects of the system. Sentilla has solved this problem by developing a system that enables a developer to build pervasive computing applications using Java. In this lab, the participants will be introduced to the Sentilla technology with a demonstration of how Java enables rapid sensor network development. Examples of design wins include logistics, transportation, security, health care, agriculture and green technology. After introducing the technology there will be hands-on lab in which the participants will be introduced to the fundamentals of building pervasive computing applications. The tutorial will cover tools, debugging techniques and architecture schemes for building applications using Java on a resource constrained device. Basic sensor implementation will be demonstrated through simple acceleration and temperature measurements, and a more advanced sensor implementaiton will follow if time permits. After completing this course the participants will have a rudimentary understanding of the Sentilla technology and how they might be able to apply it in their field.

Dynamic Task Allocation in Heterogeneous Multi-Robot Teams (Lynne Parker, ~1h)

In a heterogeneous team of robots, the morphological and/or behavioral capabilities of the team members can vary significantly from robot to robot. These differences lead to very different abilities of robots to perform the tasks within an application. A key challenge in these problems is to determine the proper allocation of tasks to robots, to attempt to optimize some criteria of performance. Unfortunately, this problem is similar to the set-covering problem, which is an NP-hard problem. Thus, researchers have developed a variety of alternative approximation approaches to task allocation to address this problem. In this tutorial, we will define the general multi-robot task allocation problem (MRTA), and discuss some alternative techniques for addressing this problem. Approaches to be discussed include behavior-based approaches, such as ALLIANCE and BLE, as well as market-based approaches, such as M+, MURDOCH, TraderBots, and Hoplites. The intent of this tutorial is to outline the general strategies that have been proposed, along with comparisons of alternatives.

Topology Control in Wireless Multi-hop Networks (Paolo Santi, ~3h)

Forthcoming wireless multi-hop networks such as wireless sensor networks will allow network nodes to control the communication topology by choosing their transmit power and, consequently, transmission range. Informally speaking, topology control (TC) is the art of co-ordinating nodes' decisions regarding their transmission range in order to generate a network with the desired features. Building an optimized network topology helps surpass the prevalent scalability and capacity problems.

In this tutorial, we first make the case for TC for what concerns both energy consumption and network capacity. We then introduce the several TC techniques presented in the literature, ranging from the simplest form of TC (optimally setting a common transmission range), to more complex topology optimization problems aimed at reducing nodes' energy consumption. Special care will be devoted to describe fully distributed TC approaches. Finally, we will consider more recent TC approaches, in which the optimization goal is no longer energy, but reducing wireless interference between nodes.

Reading Groups

Dynamic Task Allocation in Heterogeneous Multi-Robot Teams (Lynne Parker, ~1h)

The objective of this reading group is for the participants to gain a deeper understanding of techniques for dynamic task allocation in heterogeneous multi-robot teams. We will discuss papers on dynamic task allocation, as introduced in the corresponding tutorial session. We will compare and contrast the approaches, and discuss the advantages and disadvantages of the alternative techniques. We will examine some hypothetical case studies to motivate the need for alternative approaches.

  1. Gerkey and Mataric, "A Formal Analysis and Taxonomy of Task Allocation in Multi-Robot Systems", Int'l. J. of Robotics Research, 23 (9), 2004: 939-954.
  2. Parker, "ALLIANCE: An Architecture for Fault Tolerant Multi-Robot Cooperation", IEEE Trans. on Robotics and Automation, 14 (2), 1998: 220-240.
  3. Dias and Stentz, "A Free Market Architecture for Distributed Control of a Multirobot System", 6th International Conference on Intelligent Autonomous Systems (IAS-6), July, 2000: 115-122.

Topology Control (Paolo Santi, ~1h)

The goal of this reading group is to let the participants gain a view of state-of-the-art implementations of TC techniques. In particular, the participants will gain a deeper understanding of the notable technological challenges that must be solved in order to apply even very simple TC techniques into a real wireless multi-hop network.

  1. V. Kawadia, P.R. Kumar, "Power Control and Clustering in Ad Hoc Networks", Proc. IEEE Infocom, 2003.
  2. D. Son, B. Krishnamachari, J. Heidemann, "Experimental Study of the Effects of Transmission Power Control and Blacklisting in Wireless Sensor Networks", Proc. IEEE Secon, 2004.
  3. S. Lin, J. Zhang, G. Zhou, L. Gu, J.A. Stankovic, T. He, "ATPC: Adaptive Transmission Power Control for Wireless Sensor Networks", Proc. ACM SenSys, 2006.
  4. G. Hackmann, O. Chipara, C. Lu, "Robust Topology Control for Indoor Wireless Sensor Networks", Proc. ACM SenSys, 2008 (to appear).

Title (Name, length)

Reading material yet to be made available.


Title (Name, length)

Reading material yet to be made available.