Novel visual sensor deployment algorithm in occluded wireless visual sensor networks

Research output: Contribution to journalJournal Article peer-review

Abstract

In this paper, the visual sensor coverage and deployment problem in occluded wireless visual sensor network (WVSN) is studied. A new occlusion probability model is proposed to compute the occlusion probability from the fixed obstacles and the moving obstacles. Based on this occlusion probability model, an optimization formulation is developed, and then, a novel heuristic algorithm (FoVIC) is devised. The basic idea of the FoVIC algorithm is to deploy a sensor, one at a time, that can cover the largest number of uncovered objects with highest nonoccluded probability. In identifying a new sensor to be deployed, the FoVIC algorithm removes the removable sensors deployed at the earlier stages. The algorithm repeats until all the objects of interest are covered. From the computational experiments, it shows that, in occluded WVSN, the larger span angle can help to reduce the number of deployed sensors than the sensing range. Moreover, the FoVIC algorithm performs well in both random occlusion probability scenario and obstacle-location-dependent occlusion probability scenario. When considering the sensor deployment cost, the FoVIC intelligently expands the sensing range of existing sensors in high-occluded WVSN and deploys a new sensor in low-occluded WVSN to minimize the total cost.
Original languageAmerican English
JournalIEEE Systems Journal
VolumePP
Issue number99
DOIs
StatePublished - 2015

Keywords

  • Field of view (FoV)
  • Occlusion
  • Sensor coverage
  • Sensor deployment
  • Wireless visual sensor networks (WVSNs)
  • Computational experiment
  • Deployment problems
  • Location dependents
  • Moving obstacles
  • Optimization formulations
  • Probability modeling
  • Wireless visual sensor networks

Fingerprint

Dive into the research topics of 'Novel visual sensor deployment algorithm in occluded wireless visual sensor networks'. Together they form a unique fingerprint.

Cite this