Hsiu Wen Chang

An Artificial Neural Network Embedded Position and Orientation Determination Algorithm for Low Cost MEMS INS/GPS Integrated Sensors

Digital mobile mapping, which integrates digital imaging with direct geo-referencing, has developed rapidly over the past fifteen years. Direct geo-referencing is the determination of the time-variable position and orientation parameters for a mobile digital imager. The most common technologies used for this purpose today are satellite positioning using Global Positioning System (GPS) and Inertial Navigation System (INS) using an Inertial Measurement Unit (IMU). They are usually integrated in such a way that the GPS receiver is the main position sensor, while the IMU is the main orientation sensor. The Kalman Filter (KF) is considered as the optimal estimation tool for real-time INS/GPS integrated kinematic position and orientation determination. An intelligent hybrid scheme consisting of an Artificial Neural Network (ANN) and KF has been proposed to overcome the limitations of KF and to improve the performance of the INS/GPS integrated system in previous studies. However, the accuracy requirements of general mobile mapping applications can’t be achieved easily, even by the use of the ANN-KF scheme. Therefore, this study proposes an intelligent position and orientation determination scheme that embeds ANN with conventional Rauch-Tung-Striebel (RTS) smoother to improve the overall accuracy of a MEMS INS/GPS integrated system in post-mission mode. By combining the Micro Electro Mechanical Systems (MEMS) INS/GPS integrated system and the intelligent ANN-RTS smoother scheme proposed in this study, a cheaper but still reasonably accurate position and orientation determination scheme can be anticipated.

A MEMS Multi-Sensors System for Pedestrian Navigation

Micro-electro-mechanical system (MEMS) sensors are widely used in many applications due to their low cost, low power consumption, small size and light weight. Such MEMS sensors which are usually called multi-sensors include accelerometers, gyroscopes, magnetometers and barometers. In this research, Samsung Galaxy Note is used as the MEMS multi-sensors platform for pedestrian navigation. It contains a three-axis accelerometer, a three-axis gyroscope, a three-axis magnetometer and GPS receiver. Pedestrian Dead Reckoning (PDR) algorithms which include step detection, stride length estimation, heading estimation and PDR mechanization are carefully discussed in this paper. GPS solution is the major aiding source to reduce the MEMS IMU position, velocity and attitude errors when GPS signals are available. Magnetometers are also used to reduce the attitude errors of gyroscopes if there are no environment disturbances. A loosely-coupled extended Kalman Filter is implemented in the paper to fuse all the information to obtain the position result. Two typical scenarios are tested and analyzed in this paper: walking from outdoor to indoor and indoor walking. The MEMS multi-sensors system works well for both scenarios. To conclude, algorithms of MEMS multi-sensors system can provide an accurate, reliable and continuous result for pedestrian navigation on the platform of smart phone.

An ANN embedded RTS smoother for an INS/GPS integrated Positioning and Orientation System

Digital mobile mapping, which integrates digital imaging with direct geo-referencing, has developed rapidly over the past fifteen years. The direct geo-referencing is the determination of time variable position and orientation parameters for a mobile digital imager. The most common technologies used for this purpose today are satellite positioning by GPS and inertial navigation using an IMU. They are usually integrated in such a way that the GPS receiver is the main position sensor, while the IMU is the main orientation sensor. KF is considered the optimal estimation tool for teal-time INS/GPS integrated kinematic positioning and orientation determination. In post-mission processing, on the other hand, data from the whole mission can be used to estimate the trajectory. When filtering is used in the first step, an optimal smoothing algorithm can be applied to achieve higher accuracy for mobile mapping applications. An intelligent and hybrid scheme consisting of an ANN and KF is proposed to overcome the limitations of KF and to improve the performance of an INS/GPS integrated system from a previous study. However, the accuracy requirements of general mobile mapping applications cannot be achieved easily even by using an ANN-KF scheme. Therefore, this study proposes an ANN embedded RTS backward smoother to enhance the overall accuracy of POS parameters for a tactical grade INS/GPS integrated system in a post-mission mode. Combing the tactical grade INS/GPS integrated system and intelligent POS scheme proposed in this study, a cheap but reasonably accurate POS can be anticipated.

The performance analysis of the map-aided fuzzy decision tree based on the pedestrian dead reckoning algorithm in an indoor environment

Hardware sensors embedded in a smartphone allow the device to become an excellent mobile navigator. A smartphone is ideal for this task because its great international popularity has led to increased phone power and since most of the necessary infrastructure is already in place. However, using a smartphone for indoor pedestrian navigation can be problematic due to the low accuracy of sensors, imprecise predictability of pedestrian motion, and inaccessibility of the Global Navigation Satellite System (GNSS) in some indoor environments. Pedestrian Dead Reckoning (PDR) is one of the most common technologies used for pedestrian navigation, but in its present form, various errors tend to accumulate. This study introduces a fuzzy decision tree (FDT) aided by map information to improve the accuracy and stability of PDR with less dependency on infrastructure. First, the map is quickly surveyed by the Indoor Mobile Mapping System (IMMS). Next, Bluetooth beacons are implemented to enable the initializing of any position. Finally, map-aided FDT can estimate navigation solutions in real time. The experiments were conducted in different fields using a variety of smartphones and users in order to verify stability. The contrast PDR system demonstrates low stability for each case without pre-calibration and post-processing, but the proposed low-complexity FDT algorithm shows good stability and accuracy under the same conditions.

The Performance Analysis of Space Resection-Aided Pedestrian Dead Reckoning for Smartphone Navigation in a Mapped Indoor Environment

Smartphones have become indispensable in our daily lives. Their various embedded sensors have inspired innovations in mobile applications—especially for indoor navigation. However, the accuracy, reliability and generalizability of navigation all continue to struggle in environments lacking a Global Navigation Satellite System (GNSS). Pedestrian Dead Reckoning (PDR) is a popular method for indoor pedestrian navigation. Unfortunately, due to its fundamental principles, even a small navigation error will amplify itself, step by step, generally leading to the need for supplementary resources to maintain navigation accuracy. Virtually all mobile devices and most robots contain a basic camera sensor, which has led to the popularity of image-based localization, and vice versa. However, all of the image-based localization requires continuous images for uninterrupted positioning. Furthermore, the solutions provided by either image-based localization or a PDR are usually in a relative coordinate system. Therefore, this research proposes a system, which uses space resection-aided PDR with geo-referenced images of a previously mapped environment to enable seamless navigation and solve the shortcomings of PDR and image-based localization, and evaluates the performance of space resection with different assumptions using a smartphone. The indoor mobile mapping system (IMMS) is used for the effective production of geo-referenced images. The preliminary results indicate that the proposed algorithm is suitable for universal pedestrian indoor navigation, achieving the accuracy required for commercial applications.

A low complexity map-aided Fuzzy Decision Tree for pedestrian indoor/outdoor navigation using smartphone

With the great international popularity and various sensors embedded, smartphone becomes an excellent mobile and indoor navigator. Pedestrian Dead Reckoning (PDR) is one of the most common technologies for pedestrian and indoor navigation which is based upon pedometer and orientation sensor. But various errors tend to accumulated step by step in its present form. Therefore, this study proposes a novel map aided Fuzzy Decision Tree (FDT) without complex algorithm and individually tuning process to reduce the accumulated error, improve the generation ability and minimize the use of infrastructure. The rule-based FDT algorithm estimates the location based upon the map, sensors and expert knowledge after training once then for other new experiments. Various scenarios consisted of different test sites, smartphones and users are implemented in order to verify the performance of proposed algorithm. The results verify that once the proposed algorithm is well trained, it is able to maintain good position performance regardless of the users, fields and smartphones. In addition, the positioning solution can output the global coordinate because of the use of self-produced map for seamless navigation in both indoor and outdoor environments.

Cycling dead reckoning for enhanced portable device navigation on multi-gear bicycles

Abstract Based on the advancements in micro-electric mechanical systems (MEMS), lightweight and small-size inertial sensors are commonly used in various devices today. These advancements have dramatically increased the number of consumer devices that utilize MEMS-based sensors and the number of applications developed to promote interaction with those consumer devices. One application is navigation, which can be used to provide route guidance and other useful information to the user. MEMS inertial sensors can be used for navigation; however, they need aiding from other sources, such as updates from global navigation satellite system (GNSS) or other aiding sources. Nevertheless, using MEMS-based portable devices for cycling applications, in any orientation and without careful mounting or any external updates, is not enough for achieving acceptable navigation performance. We propose a new module to improve the use of low-cost motion sensors as a portable navigator for cycling without putting any constraints on the user and therefore working in different environments and in any device orientation. The proposed multi-gear cycling dead reckoning module involves an adaptive modeling algorithm. When trained, the models are used to enhance navigation. In cases where GNSS is not available or is degraded, this module can assist the integrated navigation solution. The proposed solution is general enough to apply to any type of bicycle and is also adaptive to various tire sizes and gears, as well as different cyclists. The experimental results demonstrate the performance and usefulness of the proposed solution.