Magnetic Field based Heading Estimation for Pedestrian Navigation Environments Muhammad Haris Afzal, Valérie Renaudin and Gérard Lachapelle Department of Geomatics Engineering University of Calgary Calgary, Canada {haris.afzal, valerie.renaudin, gerard.lachapelle}@ucalgary.ca Abstract—Heading estimation plays an important role in pedestrian navigation applications. With the advent of smart- phones equipped with MEMS sensors, it has become possible to utilize ones orientation information along with location. This combination has allowed researchers to investigate provisioning users with orientation aware location based services as well as seamless navigation in different environments using Pedestrian Dead Reckoning (PDR) techniques. Although gyroscopes are considered to be the primary sensors for orientation estimation, the errors associated with these sensors require periodic updates from other sources. In case of small hand held devices, these other sources are accelerometers for roll and pitch estimates and magnetic field sensors for the heading. In order to utilize the magnetic field sensors for heading estimation with respect to some known reference, it is desirable to measure only the Earth’s magnetic field components. Although this is achievable in the outdoors, presence of manmade infrastructure in all kinds of urban environments makes it impossible to sense only the Earth’s magnetic field at all times. These manmade magnetic anomalies caused by electronic devices, ferrous materials, mechanical and electrical infrastructures among others are the main culprits contaminating the magnetic field information. Therefore it is desirable to investigate how good one can estimate the heading using magnetic field alone in different pedestrian navigation environments by isolating the perturbed regions from the clean ones. In this paper, a detector is proposed that can identify the magnetic field measurements that can be used for estimating heading with adequate accuracy. The expected errors in the heading estimates are also output based on the test statistics, which allow the proposed detector to be utilized for sensor fusion and estimation of errors associated with gyroscopes. Real world data is acquired using a custom designed consumer grade sensor platform and a high accuracy reference system. Theoretical analysis and experimental results show that the proposed detector is capable of identifying the effects of perturbations on the Earth’s magnetic field, which provides users with a better estimate of magnetic heading in different pedestrian navigation environments. Keywords-Pedestrian Navigation; Magnetic Field; Indoor Orientation Estimation I. INTRODUCTION Magnetometers are sensors utilized for measuring the Earth’s magnetic field [1]. These sensors have now become small enough to be utilized in pedestrian navigation applications. The magnetic field is a three dimensional quantity and thus an orthogonal arrangement of magnetometers is required to observe it completely. By resolving these three axis measurements into horizontal and vertical fields, one can estimate the heading: the pointing direction of the sensor block with respect to the magnetic North. This heading can be easily referenced to true North with the help of magnetic field models like International Geomagnetic Reference Field (IGRF)[2]. With proper calibration of the magnetometers, it is possible to sense Earth’s magnetic field with significant accuracy, which can then be utilized for estimating heading within a few degrees of true North [3]. This is achievable if the environment is free from magnetic perturbations e.g. outdoor (country side) and airborne (aircraft). But in the context of pedestrian navigation, such perturbation free environments are seldom encountered and additional artificial magnetic fields are present causing magnetic anomalies or perturbations that change the magnetic field vector, which finally leads to errors in the heading estimates exceeding even 100° in some cases [4]. This brings one’s attention to alternate means of heading estimation. Gyroscopes measure the angular inertial forces that can be utilized for estimating the heading with respect to some reference. These sensors along with accelerometers have been used for almost half a century in Inertial Navigation Systems (INS). Such systems are also referred to as dead reckoning systems as the measurements of gyroscopes and accelerometers are effectively integrated forward in time with respect to reference navigation parameters [5]. Similar to the advancements in magnetic field sensors, the gyroscopes and accelerometers are also being miniaturized and are getting a lot of attention for pedestrian navigation applications [6]. With inertial navigation comes error growth due to its integration nature. Thus any errors associated with the sensors are accumulated causing an ever increasing error in the estimated The financial support of Research In Motion (RIM), the Natural Science and Engineering Research Council of Canada, Alberta Advanced Education and Technology and Western Economic Diversification Canada is acknowledged. 978-1-4577-1804-5/11/$26.00 ©2011 IEEE IPIN 2011, Session 7, Guimarães, Portugal, 21-23 September 2011