Loading...

Algorithms for Tracking Moving Points

Mansouri, Hamid | 2010

479 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 40575 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Ghodsi, Mohammad
  7. Abstract:
  8. Nowadays binary sensors such as infrared sensors are vastly employed in industrial, commercial and general applications due to their simplicity and low price. Beam sensors are common type of these sensors which can detect objects when they cross its beam. Sequence of sensors which are crossed in a time order is called sensor word. Consider an object, produced sensor word shows the order which it has crossed the beams. If we have more than one object then sensor word is combination of all sensor words produced by all object. We need to separate each sensor word. In this research, sensor word is developed to Fixed-Charge Minimum Cost Maximum Flow Network type.In this case, the consequent network flow is solvable by nonlinear programming solver tools. also both of these problems - original problem and proposed network flow problem - are classified as NP problems, solving proposed network flow by nonlinear programming solver, reduced problem solving time significantly. One of the problems usually arises in binary sensor network is Detection Error. This error include two major types which are False Detection and Miss-detection. In π-order fault tolerant approach, by receiving aggregated sensor word, the valid sensor words for object is generated which their combination is different at most in π places in comparison with aggregated one. In this research proposed network flow is modified to generate fault tolerant version that tolerate miss-detection error.
  9. Keywords:
  10. Fault Tolerance ; Binary Beam Sensors ; Tracking Moving Points ; Path Separation ; Fixed-Charge Minimum Cost Maximum Flow Problem

 Digital Object List

 Bookmark

No TOC