Many emerging applications based on robotics or augmented reality (e.g. for museums, BIM.. .) require accurate indoor positioning, that obviously can not be achieved by global positioning systems or traditional radio systems using received signal strength indicators. Ultra-wideband radio is a promising technology to enhance this accuracy, but its robustness still suffers in noisy environment and most of localization algorithms keep a prohibitive complexity to be embedded on target nodes. The Newton-Gauss algorithm represents a good trade-off between positioning performance and processing needs at the node level. As some recent wireless body area networks embed inertial measurement unit, the target direction can be used to further enhance the accuracy while keeping an acceptable complexity. In fact, the direction allows to narrow the research area of the mobile position. Thanks to this additional information, our direction-aided Newton-Gauss algorithm allows a gain of more than 14% in terms of accuracy over classical Newton-Gauss algorithm.