Evolutionary optimization of sensor deployment for an indoor positioning system with unknown number of anchors

We present an evolutionary multi-objective optimization method for sensor deployment applied to an indoor positioning system with range-difference measurements. Stationary sensors at known locations are used to obtain the position of a moving emitter. Coverage and accuracy of the positioning system depend on the number and location of the sensors for a given indoor space (floor plan) and on the properties of the system. The proposed method allows finding an optimum number and spatial distribution of the sensors automatically using constraints and different optimization criteria. We use the usual genetic operators for crossover and mutation and we have also introduced the possibility of deleting or adding a sensor in the sensor distribution corresponding to the offspring population. The method is applied to an infrared indoor positioning system showing two examples of sensor deployment with three and four objectives, respectively.