Motivation:
Interaction with the physical world through our machines becomes increasingly important as we expand the frontiers of computer science, the Internet, and information technology. We are faced with new challenges as we design engineering systems to solve problems such as environmental monitoring, security sweeping and surveillance, automated transportation, and even household maintenance.
In such cyber-physical systems, information arrives directly from sensors. Furthermore, physical devices may be commanded to move around to improve information gathering capabilities or accomplish physical tasks. Successful design and deployment of such systems depends on understanding the interplay between sensing, actuation, and computation. By considering these together, one central theme emerges: Understand the information requirements of a particular task so that the complexity of the overall system can be reduced. This offers many benefits, such as improving robustness and reducing costs.
This course provides fundamental concepts for reasoning about the information spaces that arise in these systems. Inferences are made from sensing and actuation information in combination with geometric, topological, or statistical models. This leads to approaches that do not require full state estimation. They instead work by directly manipulating information states. In many cases, the sensing requirements may be dramatically reduced. This leads sensor-based algorithms and even theoretical computation models that depart from standard Turing machines.
Topics:
Sensor models, visibility sensors, sensor networks,
inference problems, information spaces, actuation models, minimalist
planning, visual sweeps, searching with limited information,
pursuit-evasion games, sensor-based navigation tasks, coordinate-free
models, stochastic models, nontraditional communication models,
sensor-centric models of computation, decidability and complexity for
actuated sensor systems.
Daily Schedule (Mon-Fri):
09:15-10:30 : Lecture (Part 1)
10:45-12:00 : Lecture (Part 2)
15:15-16:30 : Discussion
Notes:
Lecture 1
Lecture 2
Lecture 3
Lecture 4
Lecture 5
Lecture 6
Lecture 7
Lecture 8
Lecture 9
Lecture 10
Lecture 11
Lecture 12
Lecture 13
Lecture 14
Lecture 15
Lecture 16
Lecture 17
Discussion 1
Discussion 2
Discussion 3
Discussion 4
Discussion 5
Discussion 6
Discussion 7
Discussion 8
Discussion 9
Lectures: (tentative)
Date | Lecture Topics | Sources |
Week 1 | ||
  6 Aug  | Overview, physical sensors |   |
  7 Aug  | Abstract sensing models, configuration spaces, state spaces |   |
  8 Aug  | Time, sensor histories, inference problems |   |
  9 Aug  | Inference problems, introducing actuation |   |
10 Aug  | Exploring grids and graphs |   |
Week 2 | ||
13 Aug  | General I-space concepts |   |
14 Aug  | Visibility problems, pursuit-evasion |   |
15 Aug  | Landmark-based navigation, on-line problems |   |
16 Aug  | Decidability, dominance, complexity |   |
17 Aug  | Probabilistic I-spaces |   |
Related Textbook: Planning Algorithms, S. M. LaValle,
Cambridge University Press, 2006. Also available for free download at
http://planning.cs.uiuc.edu/.