Valerie Renaudin

Adaptative pedestrian displacement estimation with a smartphone

Pedestrian dead reckoning is one of the most promising processing strategies of inertial signals collected with a smartphone for autonomous indoor personal navigation. When the sensors are held in hand, step length models are usually used to estimate the walking distance. They combine stride frequency with a finite number of physiological and descriptive parameters that are calibrated with training data for each person. But even under steady conditions, several physiological conditions are impacting the walking gait and consequently these parameters. Frequent calibration is needed to tune these models prior to relying on free inertial navigation solutions in indoor locations. Two hybridization filters are proposed for calibrating the step length model and estimating the navigation solution. They integrate either GNSS standalone positions or GNSS Doppler depending on the coupling level. A data collection performed with four test subjects show the variations of these parameters for the same person during his journey and effectiveness of the calibration for improving the estimation of walking distances. Thanks to the new filters, the error on the travelled distance gets reduced to 7% with the loosely coupled filter and 2% with the tightly coupled filter.

Smartphone based gait analysis using STFT and wavelet transform for indoor navigation

In this paper, we propose a frequency domain analysis for characterizing the walking gait in the context of indoor navigation, without assuming that the sensors are rigidly attached to the body. Firstly, frequency analysis is performed using Short Time Fourier Transform (STFT) since the statistical properties of the signal are changing over time but are assumed contant over a hort window. Globally STFT can extract step/stride frequency, but STFT is found non optimal for fast motion transitions. Wavelet Transform (WT) analysis is then introduced. Contrary to STFT, WT uses a size-adjustable window, which offers more advantages for human gait features extraction. When the time period of interest comprises a high frequency, the window is short, while when the local area comprises a low frequency, the window size is enlarged. This WT propriety is found to be critical our smartphone based gait analysis. Experimental assessment is performed with a smartphone Nokia Lumia 920 and a foot mounted MEMS grade inertial used as reference. These results are encouraging for designing a robust and adaptable real-time motion detection solution for smartphone in the context of indoor navigation.

Engineering, Human, and Legal Challenges of Navigation Systems for Personal Mobility

Walking is now promoted as an alternative transport mode to polluting cars and as a successful means to improve health and longevity. Intelligent transport systems navigation services are now directly targeting travelers due to smartphones and their embedded sensors. However, after a decade of research, no universal personal navigation system has been successfully introduced and adopted to improve personal mobility. An analysis of the underlying reasons is conducted, looking at the engineering, human, ethical, and legal challenges. First, contrary to adopting classical mechanization equations linked to solid state physics, location technologies must address complex personal dynamics using connected objects. Second, human factors are often not sufficiently considered while designing new technologies. The needs and abilities of travelers are not systematically addressed from a user-centered perspective. Finally, people want to benefit from location-based services without sharing personal location data to uncontrolled third bodies. Europe is a pioneer in the protection of individuals from personal identification through data processing since location data has been recognized as personal data, but the challenges to enforce the regulation are numerous. The recommendation of “privacy by design and default” is an interesting key to conceive the universal personal navigation solution. Alternative solutions are highlighted, but they definitively require a more interdisciplinary conception.

A multi-hypothesis particle filtering approach for pedestrian dead reckoning

A Map aided Pedestrian Dead Reckoning (PDR) algorithm is proposed to mitigate the drift errors and step detection limitations of pedestrian dead reckoning algorithm with handheld sensors in indoor and outdoor spaces. Specific to this context is the changing lever-arm between the handheld device and the pedestrian center of mass that introduces a misalignment between the inertial sensors and the walking directions. To address these challenges, an adaptive routing graph is built based on possible pedestrians motions, which depend on personal mobility profile and surroundings. An adaptive decision process is also developed to fuse map data with GNSS positions and PDR outputs in a particle filter. The performance is assessed with 1km walk experiments. Main contributions are (1) the calibration of the PDR step length model using both GNSS and map data during straight line travels with miss/over-detected steps modeled by the particle filter; (2) the estimation of angular misalignment between the walking and the handheld unit pointing directions in geometrically constrained areas; (3) a dynamic choice of opportune periods and measurements to calibrate the PDR outputs and improve the positioning process.

Toward a free inertial pedestrian navigation reference system

As free inertial pedestrian navigation systems with foot mounted inertial mobile unit mature, it becomes feasible to conceive a sufficiently accurate derived solution for assessing other indoor navigation systems and assisting research activities. The design of this reference solution and the remaining challenges are at the heart of this paper. A new filter with a quaternion based state vector exploits signals (acceleration and magnetic field) and motions of opportunity for estimating the navigation solution. Quasi Static Field (QSF) updates, Magnetic Angular Rate Updates (MARU) and Angular Gradient Update (AGU) are frequently applied for mitigating the low-cost inertial sensors errors. Experimental tests performed using a post-processed GPS differential solution as reference show an average accuracy of 1.2 m over 500 m walking path: 0.5 %. Discussions about the data acquisition protocol for meeting a 1 meter horizontal accuracy within two standard deviations of the mean (95.45%) is conducted.

Hybrid visual and inertial position and orientation estimation based on known urban 3D models

More and more pedestrians own devices (as a smartphone) that integrate a wide array of low-cost sensors (camera, IMU, magnetometer and GNSS receiver). GNSS is usually used for pedestrian localization in urban environment, but signal suffers of an inaccuracy of several meters. In order to have a more accurate localization and improve pedestrian navigation and urban mobility, we present a method for city-scale localization with a handheld device. Our central idea is to estimate the 3D location and 3D orientation of the phone camera based on the knowledge of the street furnitures, which have a high repeatability and a large coverage area in the city. Firstly, the use of inertial measurements acquired with an IMU in the vision based method allows to accelerate the calculation of the position and orientation. Secondly, the weighted fusion between the rotation matrices calculated with the vision and the inertial processes allows to give the more importance in the calculation with the highest confidence. With a good points selection, this provides a localization that is in the GNSS post-processed measurement precision use for determining the position and the orientation of the street furnitures. Performances are presented in terms of accuracy of positionning. The final aim is to have with our method a precision good enough to be able to propose in future works a on site display in augmented reality.

Pedestrian Dead Reckoning Navigation with the Help of A⁎-Based Routing Graphs in Large Unconstrained Spaces

An -based routing graph is proposed to assist PDR indoor and outdoor navigation with handheld devices. Measurements are provided by inertial and magnetic sensors together with a GNSS receiver. The novelty of this work lies in providing a realistic motion support that mitigates the absence of obstacles and enables the calibration of the PDR model even in large spaces where GNSS signal is unavailable. This motion support is exploited for both predicting positions and updating them using a particle filter. The navigation network is used to correct for the gyro drift, to adjust the step length model and to assess heading misalignment between the pedestrian’s walking direction and the pointing direction of the handheld device. Several datasets have been tested and results show that the proposed model ensures a seamless transition between outdoor and indoor environments and improves the positioning accuracy. The drift is almost cancelled thanks to heading correction in contrast with a drift of 8% for the nonaided PDR approach. The mean error of filtered positions ranges from 3 to 5 m.

[POSTER] An Inertial, Magnetic and Vision Based Trusted Pose Estimation for AR and 3D Data Qualification on Long Urban Pedestrian Displacements

In the context of pedestrian navigation, urban environment constitutes a challenging area for both localization and Augmented Reality (AR). In order to display 3D Geographic Information System (GIS) content in AR and to qualify them, we propose to fuse the pose estimated using vision thanks to a precisely known 3D urban furniture model with rotation estimated from inertial and magnetic measurements. An acquisition conducted in urban environment on a long pedestrian path permits to validate the contribution of sensors fusion and allows to qualify the pose estimation needed for AR 3D GIS content characterization.

Pedestrian track estimation with handheld monocular camera and inertial-magnetic sensor for urban augmented reality

Urban environment constitutes a challenging area for pedestrian navigation. However, with the recent increase of pedestrians owning devices (e.g. smartphones), complementary data provided by integrated low cost sensors (camera, Inertial and Magnetic Measurement Unit and GNSS receiver) may be used in a coupling process to accurately estimate the pose (i.e. 3D position and 3D orientation) of a handheld device. Additionally, the actual development and availability of 3D GIS content constitutes a mine of data usable for camera pose estimation. In the context of pedestrian navigation in urban environment, to update a Pedestrian Dead-Reckoning process and to improve the positioning accuracy, we propose to fuse the pose estimated through a vision process thanks to a precisely known 3D model with inertial and magnetic measurements. Experimental data collected in an urban environment, on a long pedestrian path with sparse known models permit to validate the benefit of sensors fusion process. This results in an improved positioning accuracy that enhances the Pedestrian Dead-Reckoning process and enables to display 3D information in Augmented Reality. Performance are presented in terms of positioning accuracy and compared to commonly used solutions.

Foot-mounted pedestrian navigation reference with tightly coupled GNSS carrier phases, inertial and magnetic data

Many indoor navigation systems have been developed for pedestrians and assessing their performances is a real challenge. Benefiting from a reference solution that is accurate enough to evaluate other indoor navigation systems and assist novel research is of prime interest. The design and algorithms of a foot-mounted reference navigation system titled PERSY (PEdestrian Reference SYstem) are presented in this paper. Quasi static phases of the acceleration and the magnetic field are used to mitigate inertial sensor errors in indoor spaces. Differential indoor/outdoor GNSS phase measurements are added to the strapdown EKF to improve the positioning accuracy with a correlation between low and high frequency velocity estimates. Experiments conducted with four persons over a 1.4 km walking distance show a 0.22% positioning mean error.