Indoor navigation

In open, outdoor spaces satellite-based positioning provides accurate coordinates of a user’s receiver in a global reference frame. The accuracy in consumer applications typically reaches a couple of meters with respect to the true location. In urban areas, however, high-rise buildings obstruct the signal propagation paths, and in indoor areas acquiring useful satellite signal transmissions is often totally unattainable. Other navigation techniques are needed to support, augment, and replace satellite-based navigation so that, for example, pedestrian positioning will be seamless from outdoors to indoors and of sufficient accuracy. The potential augmenting navigation technologies include inertial navigation methods and wireless network positioning approaches. Wireless network positioning technologies include, among others, cellular network based positioning (for example based on GSM), positioning methods based on signal strength measurements of wireless local area networks (for example Wi-Fi), Bluetooth-based positioning, proximity sensing based on RFID tags, ultrasound- or infrared-based positioning, and time-delay based positioning derived from ultra wideband (UWB) signals.

Figure 1. Indoor navigation infrastructure with Wi-Fi and Bluetooth.

Inertial navigation

Inertial navigation is based on measuring accelerations and rotations/orientations. The term inertia indicates an object’s tendency to continue its motion if no external forces are applied to it. In inertial navigation, an accurate reference location is firstly required. Then, the acceleration and orientation of the inertial navigation system (INS) are measured with linear acceleration sensors (accelerometers) and angular velocity sensors (gyroscopes). By integrating this information about the motion, the INS can determine the user’s position, velocity, and orientation. Motion sensors suitable for inertial navigation are available in various accuracy, quality and price grades, depending on which kind of technology is utilized. Motion sensors implemented with a low-cost technology called Micro Electro Mechanical Systems (MEMS) are increasingly common in various consumer products. This enables a vast number of possibilities for inertial positioning. Acceleration and direction derived from MEMS sensors are, however, very noisy, but they can still augment GPS in signal-denied areas, such as tunnels.

MEMS-based inertial navigation is, however, a relative positioning method and cannot be used independently from other positioning systems. In practice, such inertial navigation systems must be calibrated regularly with an absolute position obtained from, for example, GPS. The longer the time that has elapsed since the last calibration, the worse the accuracy of the position obtained by the inertial navigation system. Temperature changes and magnetic disturbances, which are typical in indoor areas, degrade the sensor measurements even further.

Motion sensors can also be used for simple pedestrian step detection, enabling the derivation of odometer information and dead-reckoning to assist satellite-based navigation or some other localization result based on radio signals.

Cellular network based positioning

There are many technologies used for implementing cellular network based positioning. These include, for example, cell identification based localization, angle of arrival based positioning, and time difference of arrival based positioning. The positioning accuracy resulting from the use of cellular network signals depends greatly on the geographical distribution of the cellular network: the accuracy is typically between 100 metres and 10 kilometres, depending on the density of the cells and the implementation technology.

WLAN and Bluetooth

Wireless local area network based positioning, e.g. Wi-Fi based positioning, as well as Bluetooth-signal based positioning, are founded on signal strength measurements. Signal strength measurements from network access points enable localization based on previously prepared databases of these signal strengths, so-called “fingerprint databases”. The fingerprint databases are generated by measuring the signal strengths from multiple access points at certain reference points in the area of interest. Then, in the positioning phase location of the user’s receiver is determined by matching the real-time signal strengths to the fingerprint database generated in the training phase. The accuracy of wireless local area network based positioning depends essentially on the number of access points in the area of interest  and the topological distribution of the reference points represented in the fingerprint database. In good conditions, accuracy of a couple of meters can be achieved.

Bluetooth, on the other hand, is a short-range (around 10-100 metres) radio technology that can be applied for proximity sensing, data communication, as well as positioning. The fingerprinting approach can also be used for Bluetooth positioning and accuracy of around a couple of meters is typically attainable.

Other wireless systems

Other wireless systems that can be utilized for positioning include those based on Radio Frequency Identification (RFID) technology, infrared communications, UWB radio signals, and so-called GNSS pseudolites. RFID tags are activated with a specific radio frequency and can act as proximity sensors for positioning purposes. Similarly, infrared transmitters at fixed locations can transmit their location information over a short range and in narrow beams to a user terminal. Infrared communication, however, requires a good line of sight to the user terminal. UWB technology can be used for ranging between the transmitter and the receiver, which enables utilizing positioning computation similar to that used in GPS. UWB signals can also penetrate solid obstacles, such as walls, but the resulting attenuation and reflections do increase the error level. Because communications regulations restrict the allowable signal power, UWB is mostly suitable for short-range indoor positioning and provides an accuracy of approximately 5 metres if properly executed. GPS pseudo satellites, or “pseudolites”, can also assist GPS navigation in signal-denied areas, for example, in urban canyons and indoors, in order to extend the GPS signal coverage area. Pseudolites can be utilized for relaying differential signal error corrections or for ranging as in satellite–based positioning. In good signal conditions pseudolites can provide positioning with an accuracy of around 1 metre. The receiver must, however, be able to utilize the GPS channels that the pseudolites occupy. In addition, the GPS frequency band is strictly regulated and, thus, pseudolite usage has not gained momentum yet. Any terrestrial transmission disturbs the actual satellite transmissions.

Heidi Kuusniemi

D.Sc. (Tech.)
Adj. Prof.
Research Professor