Optimization of Sensor Network Coverage Utilizing the Dependency between Sensor Relationships and Observed Phenomena

Jan Gorecki^* and Andrzej Kucharski

Coverage is an important issue when it is necessary to obtain a certain level of quality deploying any kind of telecommunication network. Placing network nodes in particular area does not guarantee that the network will be operating according to defined performance indicators. The network nodes must be planned accordingly in such way that coverage of this area meets expectations. \newline This paper deals with information coverage of sensing service identifying quality of service parameters which is coverage probability and error range. It is shown that different phenomena are changing significantly sensor relationships what has the influence on the coverage and k-coverage of the sensor area (i.e. number of sensors that cover each point of the sensing area). Sensor relationship between two sensors exists when particular sensor is able to cover another one with desired coverage probability and acceptable error range.\newline This paper defines the orthogonal phenomena and contrary phenomena which are crucial to be taken into account while designing and further optimizing the sensor grid. Two phenomena are orthogonal when they do not create common relationships between the sensors. Two phenomena are contrary when they create opposite relationships between the sensor. It is defined how to determine the Orthogonality Factor (OF) and Contrary Factor (CF) between 2 phenomena. \newline Having identified the phenomena, sensor relationship, desired coverage probability and acceptable error range, it is possible to run optimization algorithm to determine which sensors can be removed from the field in order to have lower cost of deployment of sensor network. The more complete set of simulated phenomena is, the more reliable optimization process is.\newline Described optimization method can be utilized in the following cases valid for sensor networks:\newline - Reduction of cost related to sensor network deployment and maintenance. \newline - Increase sensor network reliability by defining for which phenomena the sensor network is optimized.\newline - Optimizing the grid to get desired k-coverage, i.e. determining how many sensors cover a certain point of the grid.\newline - Determining which neighbouring sensor can take over the measurement activities upon sensor failure.

Mathematics Subject Classification: 94A13

Keywords: Sensor Networks, coverage, information coverage, optimization, dimensioning, neighbor relationships, optimization algorithms, OPEX/CAPEX savings, time series similarities

Minisymposion: Optimal Design of Spatio-Temporal Networks and Systems