Shape Estimation Using Location-Unknown Distance Sensors: Iterative-Closest- Point-Based Approach

We discuss the shape estimation of moving target objects using distributed ranging sensors. Due to the inability to carefully design sensor locations or assign global positioning systems (GPS) to inexpensive sensors, it is often necessary to assume that the sensor locations and target object locations are unknown. Although methods have been developed that can be applied in such situations, the sensing results are assumed to include no noise and cannot be applied to practical situations in which sensing noise exists. We propose a method of estimating the whole shape of a moving target object in the presence of sensing noise. Our method analyzes continuous reports on the measured distance of a target object from distributed sensors and determines the sensing directions of sensors using a novel algorithm, which we developed inspired by the iterative closest point (ICP) algorithm. On the basis of the obtained sensing directions, the whole shape of the object can be estimated. We conducted extensive numerical simulations and an experiment using actual laser sensors to demonstrate the effectiveness and feasibility of our method.