Pedestrian Indoor Localization and Tracking using a Particle Filter combined with a learning Accessibility Map

نویسندگان

  • Julian Straub
  • Samarjit Chakraborty
  • Martin Schäfer
چکیده

As mobile phones are starting to get equipped with inertial sensors, indoor navigation for pedestrians becomes an increasingly interesting topic in research. This work aims to develop and evaluate the use of a Particle Filter to deal with noisy sensor measurements of an Inertial Measurement Unit (IMU) providing localization and tracking of a pedestrian in indoor environments. Designed at the Institute for Real-Time Computer Systems (RCS), the so called PiNav-System was used, which can extract the motion of a person from inertial sensor measurements. On this basis a Particle Filter was implemented, which uses Dead Reckoning in combination with a geometric floor plan to localize and track a person wearing the PiNav-System in a building. In addition the concept of the Accessibility Map (AM) is proposed which reflects human walking preferences in the degree of accessibility of space in a floor and which makes it possible to exploit this information in the Particle Filter. Reinterpreting the AM as a Radial Basis Function Network, a special type of Neural Network, a method for learning accessibility of space in a floor is derived. Measurements show that the additional use of the AM in the Particle Filter yields an improvement in the localization accuracy of up to 32%, resulting in an average accuracy of 1.1m. Deploying the AM and the learning AM, also a more robust tracking is observed. Hence, besides the ability to monitor the walking patterns of a pedestrian in a building with a Particle Filter, the localization accuracy and the tracing robustness could be enhanced by the proposed AM. Nachdem inertiale Sensoren zunehmend in Handys eingebaut werden, wird Navigation in Gebäuden zu einem immer interessanteren Forschungsgebiet. Diese Arbeit befasst sich mit der Entwicklung eines Partikel Filters zur Fußgänger-Lokalisierung in Gebäuden. Das am RCS entwickelte PiNav-System extrahiert die Bewegungen seines Trägers aus den Messdaten der eingebauten inertialen Sensoren. Diese Daten über Schrittlänge und Schrittrichtung verwendet der Partikel Filter in Verbindung mit einer Stockwerkskarte um die Position der Person zu schätzen. Außerdem wird die so genannte Accessibility Map (AM) vorgestellt, welche das durchschnittliche Fußgängerverhalten durch den Grad der Zugänglichkeit in einem Raum darstellt und für den Partikel Filter nutzbar macht. Eine Lernregel für die Zugänglichkeit von Raumbereichen wird durch die Uminterpretation der AM als ein Radial Basis Function Network (RBFN) hergeleitet. In der Lokalisierungsgenauigkeit war eine Verbesserung um bis zu 32% auf 1.1m messbar, wenn die AM zusätzlich zu dem Stockwerksgrundriss verwendet wurde. Es zeigt sich außerdem, dass der Einsatz der AM oder der lernenden AM eine robustere Positionsschätzung zur Folge hat. Insgesamt konnte, neben der Möglichkeit die Laufmuster von Personen aufzuzeichnen, mit dem entwickelten System aus Partikel Filter und lernender AM die Lokalisierungsgenauigkeit verbessert und die Pfadschätzung robuster gemacht werden.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A Hybrid Indoor Localization and Navigation System with Map Matching for Pedestrians Using Smartphones

Pedestrian dead reckoning is a common technique applied in indoor inertial navigation systems that is able to provide accurate tracking performance within short distances. Sensor drift is the main bottleneck in extending the system to long-distance and long-term tracking. In this paper, a hybrid system integrating traditional pedestrian dead reckoning based on the use of inertial measurement un...

متن کامل

Automatic Generation of Topological 2D Indoor Maps for Real-Time Map-Based Filtering

Personal location information is regarded as the most important contextual information transmitted in ubiquitous systems. Many pedestrian indoor localization systems rely on mapmatching to constrain sensor errors. The maps required for computer aided localization and tracking need to incorporate a semantic structure. Such maps are not available on a larger scale and therefore most groups workin...

متن کامل

An Evaluation of Infrastructure-free and Infrastructure-based Indoor Positioning Methods with the Focus on Pedestrian Dead Reckoning

The expansion of location-based services (LBS) and their applications has led to a growing interest in localization, which can be done on the smartphone platform. Various positioning techniques can be used for indoor or outdoor positioning. Indoor positioning systems have been one of the most challenging technologies in location-based services over the past decade. Considering the increase of p...

متن کامل

A Novel Map-Based Dead-Reckoning Algorithm for Indoor Localization

Step counting-based dead-reckoning has been widely accepted as a cheap and effective solution for indoor pedestrian tracking using a hand-held device equipped with motion sensors. To compensate for the accumulating error in a dead-reckoning tracking system, extra techniques are always fused together to form a hybrid system. In this paper, we first propose a map matching (MM) enhanced particle f...

متن کامل

GridiLoc: A Backtracking Grid Filter for Fusing the Grid Model with PDR Using Smartphone Sensors

Although map filtering-aided Pedestrian Dead Reckoning (PDR) is capable of largely improving indoor localization accuracy, it becomes less efficient when coping with highly complex indoor spaces. For instance, indoor spaces with a few close corners or neighboring passages can lead to particles entering erroneous passages, which can further cause the failure of subsequent tracking. To address th...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2010