GNSS Technologies. GNSS integration with other positioning methods

Transcription

1 GNSS Technologies GNSS integration with other positioning methods

2 Content Location system alternatives RF types and classifications Locationing using RF signals Cellular positioning DTV-signal based locationing Summary Slides based on: GNSS Applications and Methods, by S. Gleason and D. Gebre-Egziabher (Eds.), Artech House Inc.,

3 Location system alternatives (1) Positioning techniques based on wireless communication can be divided roughly into proximity sensing, triangulation, fingerprinting, dead reckoning, and hybrid approaches When considering the alternatives for augmenting a GNSS receiver s capability, the usage of RF systems with GNSS is a very flexible approach RF systems have wider operation areas and lower vulnerability to environmental noise The most limiting factor is the transmission range of the RF signals RF propagation can be used for close-in applications as well as those applied over hundreds of kilometers Other aspects of a RF-based location systems that must be considered include transmission frequency and associated antenna size, multipath and signal penetration into buildings or other structures

4 Location system alternatives (2) NON-RF SYSTEMS Technology Pros Cons Inertial Other selfcontained sensors Database based Vision Ultrasound Precise with expensive sensors Works in a variety of environments No need for instrumentation Feasible for constricting INS errors High precision Works when instrumented Very inexpensive emitters and sensors High-precision Unusable without frequent corrections from external reference Unusable without frequent corrections from external reference Limited range Extensive instrumentation Limited range Very sensitive to surrounding noise RF SYSTEMS Technology Pros Cons Cellular Wide coverage Precision < 150 m typically Some instrumentation needed Local area networks Access points widespread Somewhat limited range Instrumentation needed RFID Low-cost Limited range Extensive instrumentation needed GNSS Worldwide coverage Free to use Accuracy 10 m unaided Limited indoor use

5 Location system alternatives (3) Characterization of RF location systems in various frequency bands: While at lower frequencies, efficient antennas are the key; at higher frequencies, the local environment becomes the driving factor in location determination Above 100 MHz, LOS signals become the dominant propagation type where buildings, vehicles and natural obstructions like trees begin to have greater effect

6 RF types and classifications (1) RF location technologies primarily break down into five major types: proximity direction-finding (DF) or angle of arrival (AOA) Doppler signal strength timing or phase

7 RF types and classifications (2) Proximity-based approaches include contact and nearcontact sensors such as in RF identification (RFID) these systems establish location based purely by presence operate in the near field within a few wavelengths If the tracked transmitter object is close enough for a receiver to get a signal, then the receiver and transmitter are clearly close to one another DF and AOA systems provide a means for determining the bearing of a RF transmitter from a receiver Two or more receivers can be used to triangulate on the two-dimensional horizontal location of the transmitter Additionally, altitude can be determined by a third, vertically oriented receiver

8 RF types and classifications (3) Location by Doppler uses the phenomenon that when a transmitter and receiver are traveling towards one another the received frequency is higher than transmitted, and when they are traveling away from each other the received frequency is lower By combining multiple measurements of frequency shifts it is possible to determine the location and velocity simultaneously Signal strength uses signal power or other signal-based metrics to form an estimate of range, or previously stored location information in a database, otherwise known as RF fingerprinting In the ranging case, multiple range estimates can be combined via lateration from multiple reference positions In the case of RF fingerprinting, previously collected reference values are compared to one or more current measurements to match the current location to the stored data

9 RF types and classifications (4) The RF fingerprinting principle Accuracy dependent on the numbers of base stations and the grid density of the reference points

10 RF types and classifications (5) Timing and phase location systems use measurements of the received phase of an RF signal or an additional timing modulation on the signal to estimate the range between a transmitter and receiver This is the case for GNSS Lateration is used to combine multiple range estimates to form a location estimate This timing information can be employed directly as in time-of-arrival (TOA) systems, or in round-trip-delay (RTD) measurement, or alternatively in time difference-ofarrival (TDOA) systems

11 Location by proximity (1) One of the simplest approaches to RF location is simply by sensing an RF transmitter that is at a known location The location precision of the receiver is the coverage area of the transmitter Depending on the propagation characteristic of the frequency, channel characteristics, directionality of the transmitter/receiver pair, and the effective aperture of the transmitting and receiving antenna, the range may be extended from a few meters to hundreds of kilometers The type of device most widely in use today representing proximity-based location is RFID

12 Location by proximity (2) - RFID An RFID system is comprised of at least one reader, which operates both as a power transmitter and receiver, and one or more passive tags, which are used as transmitters The RFID tags are not usually powered by batteries or other power sources but derive their power from the reader RF signal Depending on the power transmission of the reader the range to activate the tag can be as little as a centimeter or as large as 30 m or more RFID operates in different frequency bands; in lowfrequency operations it is employed at khz and is used mainly for short-range reads (less than 1 m), while at higher frequencies it can operate in the range of MHz and GHz; such operation is better suited for longer read ranges

13 Location by proximity (3) Bluetooth (IEEE ) is another wireless technology for short-range communications It uses a frequency hopping scheme that makes timing/phase measurements very difficult, but it is well-suited for proximity-based sensing since its range is dependent on its power class (optionally 1, 10, or 100 m) Similar to Bluetooth are WiFi/WLAN (802.11b/g/n) and ZigBee ( ) RF signals, which also operate in the ISM band of GHz and can also be used as proximity sensors and combined with GNSS In general, proximity-based location can be classified as an ultra-local or local area system where the range generally will not exceed 100 m

14 Location by proximity (4) Precision of the location estimate is inversely proportional to range Considerations of bandwidth, signal attenuation in the atmosphere, or through structures are generally not important the devices are often operating in the near-field of the RF bands, making far-field effects inconsequential These systems always put a constraint on the maximum number of simultaneous users for a fixed infrastructure, since they all depend on some external reader or base station with a limited number of channels

15 Case: AoA localization by Quuppa Uses bluetooth low energy (Bluetooth LE) Based on placing locators within the navigation environment and a transmitter on user One locator and single narrowband signal are enough to provide an accurate and reliable 2D location When the height of the locator and the approximate height of a Bluetooth equipment on the user are known, the positioning accuracy of the user is less than one meter when within 3-50 meters of the locator When two or more locators are used the location may be expanded to 3D. The angular positioning is based on basic trigonometry evaluating the elevation angle from the Bluetooth tag attached to the user to the locator GNSS Technologies

16 Location estimation using signal strength (1) Position-determination techniques based on signal strength and other signal-related properties such as signal-to-noise ratio (SNR), link quality indicator (LQI), response rate (RR), or carrier to noise ratio (C/N 0 ) have also been developed The two most common RF sources used in these studies are WiFi access points (IEEE ) and cellular towers (both GSM and CDMA), but also Bluetooth and recently Magnetic field based solutions have emerged There are basically two general classes that signal strength solutions break down into: distance estimation (leading to lateration) and pattern recognition (also known as RF fingerprinting)

17 Location estimation using signal strength (2) The decreasing and fluctuating signal level (due to both the environment and fading) as it travels from the transmitter is illustrated as Distance estimation (1): As the distance between the transmitter and the receiver increases, the received signal strength (RSS) decreases the RSS can be linked to a distance measurement In cellular and WiFi systems the RSS is used for handoff and traffic processing and is accessible without changing system architecture Propagation path loss with multipath

18 Location estimation using signal strength (3) Distance estimation (2): When discussing indoor location systems, it is necessary to model the transmission properties through walls and other obstructions The relative orientation of the receiving antenna and transmitter can greatly affect the received signal strength The drawback is the training phase (including retraining if the physical layout should change) and the variability of environmental factors such as the number of people in the building at a given time the overall accuracy compared to timing-based approaches is generally lower

19 Location estimation using signal strength (4) Pattern recognition (1): In contrast to approaches that seek to map signal strength into range, pattern recognition techniques have been applied to signal metrics to form location estimates This type of approach is often referred to as fingerprinting and uses previously stored measurements or calculations to map locations for later matching to observed measurements This approach takes advantage of the local-area fading characteristics by using features from multiple transmitters to uniquely identify a location Generally speaking, the finer the training grid the better the resultant accuracy Fingerprint generation phase is the training phase

20 Location estimation using signal strength (5) The basic concept of a pattern recognition locating solution using a fingerprint map from signal strengths

21 Location estimation using signal strength (6) Received signal strength indicator (RSSI) based probabilistic approach for pattern recognition: Data Training Phase Generating RSSI distribution statistical model Grid points Location Determination Phase User Locations x u, y u RSSI1, RSSI2,... x 1, y 1 RSSI1, RSSI2,... x 2, y 2 RSSI1, RSSI2, x n, y n RSSI1, RSSI2,... Training Data Model x i, y i Observations Location Estimates Grid point most close to the user

22 FGI: Positioning based on WLAN and Bluetooth Preparation of a fingerprint map based on signal strength (RSSI) Position computed based on the map and the received signal strength Accuracy dependent on the frequency of base stations Results at FGI WLAN open corridor Mean [m] Std Max Bluetooth BSE PKF IFSD Mean [m]: Std Max GNSS Technologies Different integration algortihms 13 base stations

23 Examples WiFi localization by Ekahau Ltd. Wi-Fi Planning, Deployment and Troubleshooting with Ekahau Site Survey Bluetooth localization by 9Solutions Ltd. Real time localization system using Bluetooth Low Energy and cloud technologies by 9Solutions Ltd: dded&v=zuwo7y5nh-s GNSS Technologies

24 Case: IndoorAtlas Fingerprinting using magnetic fields for indoor positioning Based on measuring the distortion of magnetic field caused by building s materials GNSS Technologies

25 Location using TOA, POA, and TDOA (1) GNSS systems use pseudoranges, timing measurements collected at a receiver that can be based on timemodulated spreading code (t) or the carrier phase (f) GNSS is the most obvious example of a time-of-arrival (TOA) and phase-of-arrival (POA) system Typically multiple TOA ranges are combined via multilateration The range measurements are represented for an individual transmitter by The same formulations as for TOA systems apply for POA systems; however, there is usually a carrier-cycle ambiguity between the transmitter and receiver n n ( t) r ( t) b( t) the advantages of POA systems is the precision of the carrier phase

26 Location using TOA, POA, and TDOA (2) In contrast to TOA, time-difference of arrival (TDOA) systems use arrival times from two transmitters at a receiver The time differences in a TDOA system are represented as line-of-position hyperbolic curves along which the receiver lies By using a second measurement the two-dimensional position can be determined The significant advantage of TDOA over TOA is that the receiver clock bias is not an important factor in location determination, since the receiver clock term b(t) will cancel upon pseudorange differencing

27 Emerged RF technologies for localization DASH7 (433 MHz ) an open source wireless sensor networking standard for wireless sensor networking, which operates in the 433 MHz unlicensed ISM band active RFID range of up to 2 km is being used for developing new location-based services NFC, near field communication (13.56 MHz) a set of short-range standards for smartphones and similar devices to establish radio communication with each other by touching them together or bringing them into close proximity, usually no more than a few centimetres allowing two-way communication between endpoints active RFID Bluetooth low-energy (2.45 GHz) a feature of Bluetooth 4.0 wireless radio technology, aimed at new, principally lowpower and low-latency, applications for wireless devices within a short range (up to 50 metres) active RFID ZigBee (868 MHz in Europe, 915 MHz in the USA and Australia, and 2.4 GHz worldwide) ZigBee is a specification for a suite of high level communication protocols using small, low-power digital radios based on an IEEE 802 standard for personal area networks

28 Cellular positioning techniques Network based: Cell ID (CI) Received Signal Strength Angle of Arrival (AOA) Timing Advance (TA) in GSM Round Trip Time (RTT) in 3G Time Of Arrival (TOA)/Time Difference of Arrival (TDOA) Mobile/Handset-based: GPS Hybrids: Assisted GPS (A-GPS) Enhanced Observed Time Difference (E-OTD) in GSM Observed Time Difference Of Arrival - Idle Period DownLink (OTDOA-IPDL) Advanced Forward Link Trilateration (A-FLT) in US CDMA

29 Few definitions Network-based: Base stations (BS) make measurements and a location centre/server calculates the position Mobile-based: Mobile station (MS) makes measurements and calculates position Mobile-assisted network-based: MS makes measurements and sends them to the location centre where the position is calculated Network-assisted mobile-based: BS makes the measurements and sends them to MS where position is calculated

30 Cellular positioning: Cell ID Cell ID (CI) Cell ID Accuracy depends on the size of the cell (e.g., 100 m 35 km) Network-based method Serving cell is used to locate the user: latitude/longitude of the cell site will give the location cell radius defines the accuracy/uncertainty of the position Most basic of all cellular location methods Low cost method (only minor software changes are required) Enhancements include: - CI +Timing Advance (TA) - GSM - CI + Round Trip Time (RTT) - UMTS - Enhanced Cell Global Identity (E-CGI): CI+RSS TA resolution is 550 m (in GSM) Maximum RTT resolution is 1/16 chip ~78 m/16 = 5 m (in UMTS) Included in 3GPP standard

31 Cellular positioning: RSS Received Signal Levels (Rx-level) or Rx Signal Strength (RSS) Estimated location -75dBm Received power: Site A: -57 dbm Site B: -75 dbm Site C: -93 dbm B A -57 dbm Typical accuracy: m C -93 dbm Network-based method Received signal level depends on the distance between MS and the BS Since signal levels are known, the location of the MS can be found in the intersection point of three circles (trilateration) Using propagation models, the distance can be estimated Changing propagation conditions cause problems limited accuracy; any obstruction in the signal path can cause large attenuation Low cost method

32 Cellular positioning: AOA Angle 1 BTS 1 Typical accuracy: m Angle 2 BTS 2 Angle Of Arrival (AOA) Network-based method Requires the use of adaptive antennas In order to estimate the position, the AOA measurements need to be done by at least 2 BSs AOAs are measured by the BS Non Line of Sight causes problems Location errors increase when distance between BS and MS increases Not a standardized method In 3G, smart antennas => no changes needed

33 Assisted GNSS/GPS MS calculates the position based on time of propagation measurements Determining the position based on time of propagation measurements requires satellite information, which is transmitted in information messages time information satellite clock correction data ephemeris (precise orbital parameters) almanac (coarse orbital parameters) coefficients for the ionospheric delay model coefficients to calculate UTC (Universal Coordinated Time) from GPS system time. The idea of A-GPS is to: Increase the sensitivity of GNSS/GPS receiver by providing information that cannot be heard directly from satellite Reduce time to first fix (TTFF) (from more than 30 sec to just a few seconds) acquisition assistance Increase the positioning accuracy if differential corrections are implemented

34 Accuracy of various cellular-based methods Openwave 2002

35 Wireless environment degradation sources The speed of wave propagation Fundamental source of difficulty for any of the RF location technique, which are based on time measurements Since radio waves propagates at the speed of light, in order to calculate position within 50 m accuracy window, the supporting time measurements must be accurate within 166 ns (approximately, 50 m = 166 ns * 3*10^8 m/s) Noise and interference One of the primary impediments to high capacity in CDMA-based systems is the near-far effect, which makes hard to receive signal from a distant user In most of the cases, noise and interference can be modeled as a random, zero mean process. Therefore, by averaging over the observation time, the impact of these errors can be reduced. Fading Un-beneficially affects the location measurements directly or indirectly Indirect corruption depends on the ratio between the desired and undesired signals. Similarly, averaging process can reduce the impact of fading. Non-line of sight propagation Multipath propagation is the dominant source of error in most of the cases. Only Line Of Sight (LOS) provides true information about the distance or the angle of the receiving signal. Accuracy in almost all RF location methods depends on the existence of LOS

36 Integration methods Combinations of location methods such as GNSS and signal strength as well as GNSS and proximity do not offer direct integration schemes The integration can be performed at a position/velocity level

37 Pseudolites (1) Probably the earliest combination of nonsatellite RF signals with GNSS was the use of pseudo-satellites, or pseudolites, employed during the original GPS constellation testing and verification stages The main disadvantages of the approach is that these pseudolite transmitters broadcast signals that may block the existing satellite transmissions, especially when the receiver is much closer to the pseudolite and the relative signal strength is overwhelming the input dynamic range of the receiver (the near-far problem) pulsing the signal address the near-far problem as well as ranging codes Pseudolite usage is subject to license from the authorities

38 Pseudolites (2) Synchronized pseudolites: synchronizing its signals using one or more GPS satellite signal for indoor applications the GPS signal can be rebroadcast indoors where signals may not be available directly from the satellites provide multiple broadcast sites with timing delays to allow TDOA measurements for precise indoor positioning

39 FGI example Pulsed pseudolites indoor and real-life GNSS integrated Accuracy of around 7 meters is achieved in a typical glass, concrete, and steel office building with pseudoliteproximity (6 pseudolites are installed in the 3-storey FGI building), floormap, and highsensitivity indoor GNSS signals fused together Pseudolites provide a convenient navigation aid indoors for a GNSS receiver without the need for using additional hardware 3. floor max mean GPS GPS+ PLs in PF GNSS Technologies

40 TV-based location integration Television signals are broadcast all over the world Television broadcasts are at a lower frequency than GNSS signals giving better structure penetration and transmit at a relatively high bandwidth of 6 MHz providing acceptable phase resolution A local reference receiver measures the signal timing of the television signals and GNSS signals and reports this to a location server The receiver to be located measures the TOA of the television and GPS signals and forwards this to the location server to compute a rangebased solution In a system implementation, combined TV+GPS location accuracies in a highly challenging urban environment have been reported to be even below 50 m

41 DTV Background Property of DTV signals for positioning High Tx power: 10 KW- 15 KW power Freq. diversity, 6-8 MHz signal bdw website) (source: Rosum corp. Fixed transmitter: Less Doppler effects Low frequency: UHF ( MHZ), VHF ( MHz) Horizontal signals, less attenuation from walls than roof Economy: Infrastructure savings DTV development : From DVB-T to DVB-T2 Multiple transmitter: Increase payload 30% Portable/mobile receiver + cell. comm. service provider Market adoption: UK ( ), Italy ( ), Sweden ( ), Finland ( ) by DNA corp. GNSS Technologies

42 TOA estimation with DVB-T signals Test platform Testing Scneario in Marseilles, Fr. GNSS Technologies

43 Evaluation criteria for location technologies Accuracy One of the most important performance measurements: commonly used criterion is the root mean of squared error (RMSE). Two-dimensional Circular Error probability (CERP) and three-dimensional Spherical Error Probability (SERP) are also used. Reliability Refers to the controllability of observations, that is ability to detect blunders/outliers and to determine the effect of blunders on the location accuracy Availability Refers mostly to coverage area within which positioning is available. Availability sometimes refers to both accuracy and availability. Latency Demand for short latency/delays is very important in case of emergency use of positioning Applicability Measure of all the physical limitations and requirements associated with the implementation and use of a certain technology.

44 Summary (1) Often stand-alone GNSSs are not sufficient enough in severely obstructed signal conditions for providing means to obtain a location Assistance is required in order for a GNSS receiver to be able to acquire and track the satellite signals AGPS/AGNSS Supportive systems for backing up GNSS or as additional support and verification are essential, especially in indoor conditions Pseudolite infrastructures in which a set of pseudo-satellites or pseudolites send a GNSS signal that is tracked by the receiver similarly to a normal GNSS satellite MEMS / INS sensors for propagating the initial position WLAN systems that are today occupying more and more public areas Cellular positioning Digital TV broadcasts RFID technology is based on radio signals that are exchanged between an RFID reader an RFID tags (Bluetooth low-energy/nfc/dash7) Bluetooth technology Also many other approaches; such as UWB, ZigBee, vision, ultrasound, infrared, etc

45 Summary (2)