Fusion of the magnetic and optical sensor information for motion capturing

Abstract

We propose a sensor fusion technique for motion capture system. In our system, two kinds of sensors are used for mutual assistance. Six magnetic sensors are attached on the arms and feet for assisting twelve optical markers and six optical markers, respectively, which are attached on the arms and feet of a performer. The optical sensor information is not always complete because the optical markers can be hidden due to obstacles. In this case, magnetic sensor information is used to link discontinuous optical sensor information. Sensor fusion seeks to overcome these drawbacks by integrating or combining information from independent two or more sensor readings. We can easily think that combining readings from several different kinds of sensors can reduce the uncertainties and give a system more accurate information than the reading data from a single sensor. The simplest case of fusion for a multi-sensor configuration that records the same property of the environment is to combine the data using averaging. Here it is assumed that nothing is known a priori about the sensors characteristics and thus all the readings have the same level of belief. In other words, when it is known that a particular sensor reading is more reliable than others, a weighted average of the sensor readings can be used instead. However these simple combining methods have some problems and are not suitable when two sensors characteristics are complex as the environmental conditions like our case. Thus we need formal and intelligent approaches to model the complex situation using two heterogeneous sensors. Here we use Neural Network, Fuzzy logic, and System Identification technique for modeling the relation between the sensors’ signals. In these modeling and testing, dynamic systems are constructed with sample input-output data. Finally, we determine the best model from the set of candidate models using the minimization of Least Squares Method (LSM) error. Our approach is very simple in the...