Accurate Indoor Positioning Using Multipath Components

Kurzfassung

Positioning is next to communication the most important field of applications for wireless radio transmissions. It can be achieved by ranging, i.e. the determination of the propagation distance of the radio wave from a transmitter at a known location to the receiver. Under pure line-of-sight (LoS) propagation conditions, the information of the wave’s traveled distance can be extracted from the amplitude, phase or the delay of the wideband signal. To obtain the three dimensional position of the receiver, ranges to at least four different transmitters need to be measured, assuming that the transmitters are synchronized and their positions are known. Signals of opportunity used for positioning are for example originated from mobile communication base-stations, Global Navigation Satellite Systems (GNSSs), dedicated ultra-wideband (UWB) transmitters or WLAN base-stations for indoor. Well known for positioning is the Global Positioning System (GPS) for example. Using the GPS, multipath reception degrades the accuracy of the positioning device as long as the receiver is based on standard methods for synchronization like a delay locked loop. Strategies to mitigate multipath effects on the ranging estimate are in general based on the estimation of the channel impulse response (CIR). To extract the range information the first arrived path is treated as the LoS path. Examples for this type of algorithms are maximum likelihood based or based on sequential Bayesian filters. All these methods have in common that they determine the multipath in order to remove the influence on the estimate of the LoS path. At least, three different signal transmitters are necessary for estimating the position and if the receiver’s clock is not synchronized an additional signal transmitter is necessary. In the field of indoor positioning with UWB signals a method using only a single anchor and known room geometry has been published. Within this approach, reflected signals are used for positioning knowing the location of walls and the physical transmitter. This paper focuses on multipath aided positioning. Hence, the paper uses additionally the multipath components of the signal for obtaining better positioning performances. Channel sounder measurements with a moving receive antenna showed, that some multipath components have a path life of several meters. The paper considers a quasi-stationary scenario, fixed transmitter positions and a moving receiver. To efficiently estimate and track the time-variant multipath components of the received signal, the paper uses a Space-Alternating Generalized Expectation-Maximization (SAGE) based Kalman filter method. The novel approach treats multipath signals as transmitters from virtual base-stations with a fixed location. Because of the multipath signal, the virtual base stations are time synchronized to the real transmitter, so emitting a signal at the same time. Therefore, using multipath signals, positioning with only one physical anchor is possible. To use the information of the multipath components of a signal, the positioning algorithm has to estimate at the same time the position of itself and the position of the origin of the reflected signal, called virtual base-station. Furthermore, by estimating the virtual base-stations also accurate position estimation in non-line-of-sight (NLoS) scenarios is possible. Especially in these scenarios, accurate position estimation with common and advanced position estimation algorithm is impossible. Additionally, the new approach does not rely on any prior information such as the room layout or information collected in a database for finger printing. To verify the proposed algorithm, simulation results consider a scenario with a moving receiver and 2 base-stations. The signal is transmitted at a frequency of 2 GHz, a bandwidth of 100 MHz and the signal is reflected and diffracted by terrestrial objects. Thus, the signals reaching the receiving antenna consist of multiple paths, which include constructive and destructive interference and phase shifting of the signal. By using the mentioned SAGE Kalman approach the algorithm is able to estimate the CIR of the received signal. At the beginning of the track both base-stations are LoS, however after the half of the track, the LoS components of the received signals disappear and we obtain a typical NLoS scenario. With a common receiver, accurate position estimation for the second half of the track is not possible. These receivers track the first path as a LoS path and obtain a wrong position estimate. With advanced algorithms, such as tracking filters, which try to estimate the LoS path, better position estimations are possible. However, by using the proposed novel algorithm, accurate position estimation is still possible. The algorithm is able to use the multipath components and the estimated virtual base-stations as an input source. Therefore, also in the NLoS scenario accurate position estimation is possible. To summarize, this paper shows a positioning approach using the multipath components of the received signal as an additional position information source. Therefore, the novel algorithm derived in this paper estimate the position of the virtual base-stations, where the multipath components are treated as transmitters. Additionally, a room layout or information about the surrounding area as a prerequisite is not needed. Furthermore, the algorithm enables accurate position estimation in serve multipath environments and especially in NLoS scenarios. In the final paper, we will present the derivation of the proposed algorithm and explain the advantage of the algorithm in detail. Furthermore, we will provide additional simulation results which show the advantage of using the novel algorithm and compare it to common and advanced algorithms.