In the past decade mobile phones have undergone a tremendous paradigm change. At first only intended for human-to- human communication purposes, the mobile phones evolved towards “smartphones” offering services and techniques for different applications.
The technical break-through became possible by new human-machine interfaces (e.g., touchscreens, speech technologies), high and affordable data rates of mobile networks, and by new powerful and simultaneously energy efficient microprocessors.
Modern smartphones support a multitude of wireless communication protocols, be it for cellular communications (e.g., UMTS, LTE), for short range, high data rate connections (BT, WLAN) or for very short range contactless data exchange, e.g., for payment services (NFC).
One of the first add-ons not related to communication was the integration of GPS (Global positioning system) receivers, thus enabling positioning and navigation services for the user. Smartphone-based positioning has found widespread use both for indoor positioning and as an alternative to built-in navigation system for cars.
Their main advantages compared to dedicated devices for car navigation are reduced costs, as no extra hardware needs to be purchased, and increased flexibility, because the device can be used both for on-board and off-board navigation.
However, a well-known deficiency of GPS-only navigation is that no positioning information is available during GPS dropouts, which may occur, e.g., in narrow street canyons, tunnels or parking garages.
Higher availability of precise positioning information is therefore obtained by built-in car navigation systems which utilize inertial measurement units (IMUs) and other information (e.g., car speed information, vector maps, map matching) to predict the position of the car based on the last position information before the GPS dropout.
This is particularly important for applications where the po- sitioning information is used for electronic toll collection, such as in the German truck toll collecting system.
In the following we propose an approach which aims at achieving the same high precision positioning as built-in car navigation systems, however without additional hardware and installation costs.
To this end we realized a sensor fusion algorithm on a smartphone which combines the GPS data with angular velocity information from the gyroscope sensors of the smartphone and further, with the car speed information obtained from the vehicle’s CAN-Bus via a wireless CAN-Bus-to-Bluetooth (CAN-BT) adapter.
On the smartphone a strapdown algorithm and an error-state Kalman filter are used to fuse the different sensor data streams. Sensor fusion is carried out by an error-state Kalman filter, whose complexity has been reduced such that it operates well below real-time.
The experimental results show that the system is able to maintain higher positioning accuracy during GPS dropouts, thus improving the availability and reliability, compared to GPS- only solutions.
To read more of this external content, download the article from the author archives on line at the University of Paderborn.