Portable Navigation Utilizing Sensor Technologies in Wearable and Portable Devices
MetadataShow full item record
Portable navigation may seem similar to other navigation problems where system availability, reliability, and accuracy are the objectives. However, more challenges arise due to the portability side of the problem such as having no constraints on the usage of the portable devices and its orientation or position relative to the user or platform. Sensors’ cost, computation capabilities, and battery life are additional challenges. An example of some of these challenges is when the designer tries to use more computation power to enhance the accuracy, even if the device’s computation capability tolerate this increase, it may shorten the battery life of the device. This research targets the development of methods to enhance portable navigation for portable and wearable devices in all environments, even in the absence of wireless positioning technologies (like GNSS and WiFi). The enhancement is through benefiting from the presence of multiple connected devices, each including at least one inertial sensor assembly. This research proposes new methods involving the integration of multiple devices and investigates how it can enhance portable navigation. The proposed methods have a computational demand sufficient to meet battery life constraints for portable devices. In addition, the integrated solution can scale to any number of devices at any point of time. Moreover, the method does not assume any predefined configurations of the devices or relative orientations among them and with respect to each other. The contributions of this research are achieved through three methods. The first enhances the accuracy of step length estimation. The second method was developed to accurately estimate the misalignment angle between the portable devices and the hosting platform, for example a person while walking or a vehicle while driving. Finally, the overall accuracy of the estimated position and velocity is enhanced while maintaining an efficient computational cost by developing a new method utilizing a new source of Kalman filtering update. The proposed methods were tested on a large number of real world trajectories collected using several portable/wearable devices during walking, running, and driving. The proposed methods resulted in significantly improving the accuracy of the portable navigation solution.