Numerical Investigation of Body-worn Ultra Wideband Antenna Localisation Techniques for Motion Capture Applications

This paper presents study of accurate body worn antennas localisation using Ultra Wideband time of arrival techniques. The localisation results obtained in presence of no obstacle are compared with those obtained in the presence of obstacles (e.g. wood and glass objects) near the base stations. Results show that accuracy in estimation of location of the sensor depends on position where it is placed on the human body as the presence of body would cause interference. It is observed that difierent base stations give difierent received signal output depending on line of sight (LOS) or non line of sight (NLOS). As expected there is more error obtained in estimation of sensor positions located on the human model due to presence of obstacles. The average accuracy in 3D localisation is around 2.5 to 4.5cm when no obstacle is present. Presence of obstacle leads to reduction in accuracy of results (2 to 3cm reduction) with glass showing higher (4 to 6cm) average error in comparison to wood (3 to 4.5cm). Localisation and motion tracking using body-worn sensors is emerging as an important research area based on Ultra Wideband (UWB) technology. Motion tracking itself is motivated by a va- riety of applications such as training of athletes, patient monitoring in medicine, localisation of people in home or o-ce environment. The choice of sensors such as compact, e-cient and low cost UWB antenna makes human localisation and activity monitoring a promising new application made possible by advances in UWB technology. Impulse Radio UBW technology ofiers various advantages such as high accuracy, robustness, low cost, high data rate, robustness towards multi- path, ease of implementation and low energy consumption making it suitable for short range human localisation (1). There are various wireless technologies available for human tracking and localisation such as infrared, ultrasound and Radio Frequency (RF) based technologies. Infrared signals are low power, and inexpensive but they cannot penetrate through obstructions, e.g. opaque objects require line of sight between transmitter and receiver and they are also susceptible against sunlight (2{4). Ultrasound signals provide high accuracies in the short range. The systems based on ultrasound technology are relatively cheap but the precision is lower in comparison with IR-based systems. Radio frequency is used most commonly for localization purpose because these signals can penetrate through obstacles and can propagate to long distances. The RF signals are classifled into narrow band (Radio Frequency Identiflcation (RFID) and Wireless Local Area Network (WLAN)) and Ultra Wideband technology (4). In the literature amongst these technologies, the ultra short pulse UWB based systems give high accuracy and due to its low power can carry signals through many obstacles that usually re∞ect signals at more limited bandwidths and at higher power (2{6). An accuracy of 10 to 15cm is achieved in commercial localisation systems with operating range of 50cm (7) and also sub-millimetre range accuracy is possible using carrier based UWB systems as proposed in literature (6). Thus UWB localisation has a widespread usage including localization of body worn sensors. In this paper UWB 3D human body localisation is studied using 14 difierent sensors on the body using CST simulations in presence of difierent obstacles and the accuracy is compared with situation when no obstacle is present. The objective of the work is to achieve accurate localisation of the human body using data fusion time of arrival and peak detection techniques.