|My News|| |
Time To First Fix Will IRNSS solve the problem?
|My News|| |
Amongst GNSS receiver design criteria such as accuracy, sensitivity, channels etc., the Time To First Fix (TTFF) is an important criterion, which defines how fast a navigation solution is available to the user since receiver power on. TTFF is defined as the time that a receiver takes to acquire and track a minimum of four satellites and extract the necessary information (ephemeris – primary parameter) from the demodulated navigation data bits. In the past decade, there has been a constant demand from the user community to optimize TTFF specifications. For example, for E911 use, Global Navigation Satellite System (GNSS) receivers integrated as a part of mobile handsets obtain assistance from a base station (terrestrial link) to enhance the TTFF.
This article provides a top level description of methods to achieve fast TTFF and characterizes it using GPS and GLONASS L1 receiver, then generalizes the result for emerging systems such as the Indian Regional Navigation Satellite System (IRNSS). Using an example, the need for Fast TTFF in single frequency mode of operation is illustrated. A brief overview of receiver operations is first presented from a TTFF perspective.
Receiver Operations leading to TTFF
At a top level, the receiver needs to solve three unknowns in position – latitude, longitude and altitude and receiver time, based on the measurements performed on the signal Eq. (1). Typically, the measurements within a receiver are generated from a relatively inferior grade oscillator in comparison with the onboard atomic clock . To facilitate position estimation, measurements need to be synchronous to the satellite clocks, which necessitate the estimation of the local clock misalignment . With this as the fourth unknown, position and time solution mandates measurements to four satellites for navigation estimation. Based on measurements to four satellites, a user position solution is obtained.
The Radio Frequency (RF) signals from the satellites incident on the user antenna are processed in the RF down conversion of the receiver, which generates suitable digital Intermediate Frequencies (IF) for further baseband processing. Signal detection and measurements are performed on the baseband signals. Detection involves estimating the signal in code and Doppler dimension and measurements, in the estimatio of pseudorange and range rates. In addition, the Navigation (NAV) data obtained from the detection process enables the receiver to estimate the satellite position. This along with measurements enables navigation solution estimation. The TTFF consists of operations from processing the signal incident on the user antenna leading to position computation post power-on. A GNSS user whose primary objective is to obtain the navigation solution would require it immediately from poweron. The need for fast TTFF is a major specification of handheld and critical military applications. The following section explains in detail the various methods practised in industry to enhance TTFF.
Fast TTFF methods – Without external assistance
Based on the inputs (ephemeris, almanac user position and time) available at power-on, a receiver is categorized into four different modes – Cold, Warm, Hot and Snap. The following paragraph explains each mode of receiver operation and its achievable TTFF.
In this mode of operation, a receiver has no prior inputs available within the receiver. With power-on, the receiver needs to search for satellites available from the entire constellation. Subsequently, the system time estimation, measurements to four satellites and NAV data collection are performed. For GPS L1, this process takes 32 to 36 s assuming the initial search is based on fast signal acquisition algorithms .
These receivers have access to satellites’ almanac, approximate user position and time as shown in Figure 2. They are either supplied externally or maintained internally in the receiver. In the internal configuration case, the time at the last power cycle is maintained and estimated typically in Real Time Clock (RTC). Similarly, user position and almanac are maintained in non-volatile memory of baseband processing section. The advantages of these parameters are twofold: First, the receiver will be able to compute the list of visible satellites and restrict the signal detection to those fewer satellites (than complete constellation) that are actually above the horizon. Second, with the almanac, the receiver will be able to compute approximate satellite position; with such an approximate user position, the geometric range can be established. This results in a reduced search range in the code and Doppler domain, which enhances acquisition time. Subsequently, the remaining operations leading to TTFF is similar to cold start and takes 32 s.
In addition to the warm start estimates, with a valid ephemeris at power-on, TTFF reduces to 6-8 s and this mode of operation is termed as Hot start . With ephemeris, the only parameter that needs to be established is the system time. Typically, this mode is adopted in most automotive grade receivers.
The last mode of receiver operation is the Snap start where, in addition to Hot start parameters, the receiver clock is also estimated internally. The advantage with this approach is further reduction of system time estimation. With bit synchronization established on four satellites, TTFF is achieved instantaneously. This mode assumes that the receiver was powered on recently .
Fast TTFF – Assisted GPS (AGPS)
The various start modes assumed receivers have the necessary information stored as a part of its internal memory or assisted from external network, the former being receiver specific, the latter being addressed by AGPS. AGPS works on the principle of client server architecture, with the receiver (handset) operated as client. Based on the levels of service offered, AGPS can be categorized as: Ephemeris Assistance and Absolute .
Ephemeris assistance is similar to the Hot start mode of operation. With a link to the server, the receiver will be able to obtain approximate position, time and ephemeris. Subsequently, the receiver operations are similar to Hot start with an achieved TTFF of the order of 6-8 s. This mode assumes the receiver to have a constant link with the server, synchronize to the network and thus be able to predict the code and the Doppler search ranges precisely. The absolute mode of AGPS is similar to the Snap start mode of receiver, which provides the TTFF in 2-3 s. However, the load on the server is higher in comparison with the ephemeris assistance mode .
With the details of receiver start modes, the next step is to experimentally understand the TTFF from power-on and quantify various parameters underlying it. Towards this, the experimental setup shown in Figure 3 is employed .
The apparatus consisted of a GPSGLONASS receiver from Accord Software & Systems Pvt Ltd, which was used as a platform to profile various receiver parameters. The receiver supports 32 GPS and 14 GLONASS channels. The receiver does not have the provision of storing or estimating any parameters from previous power-on or hardware to support external data.
The receiver was connected to a GPS GLONASS antenna, placed in a surveyed location under open sky. A Digital Storage Oscilloscope (DSO) is connected to the boot-pin of the Digital Signal Processor (DSP) to profile the application boot time (time required to load the software to DSP post power-on). Th e boot-time with this receiver is around 250 ms, as shown in Figure 3. Generalizing this component, let,
Tb be the time taken to boot the receiver application (2)
Subsequent to booting, the receiver is programmed with 32 GPS and 14 GLONASS satellites to respective channels for further processing. The satellites visible at the receiver antenna take between 2 to 8 s for signal acquisition. This is due to the large Doppler search range supported by the receiver. Generalizing this component, let
Ta be the time taken to acquire the visible satellites (3)
Subsequent to acquisition, each channel takes a finite time for bitsynchronization. This is around 800ms. Generalizing this component, let
Tbs be the time required for bitsynchronization (4)
The next activity on each of the bitsynched channel is the collection of NAV data, specifically the ephemeris. This takes anywhere between 18 and 30 s for GPS and 8 and 30 s for GLONASS channels, respectively. Let
Teph be the time required to collect ephemeris. (5)
Finally, from the NAV data extracted and the measurements formulated, the user position is computed. Using Eq.(2) through Eq .(5), the TTFF for any Lone of Sight (LOS) receiver is given by
Subsequently, for each tracked channel, the almanac is collected, which takes 12.5 minutes from the instant of the first tracked channel in case of GPS and 2.5 minutes in the case of GLONASS. Generalizing this component, let
Talm be the time taken to collect almanac (7)
Following the almanac collection, the receiver applies the ionosphere correction to the pseudoranges to improve accuracy. Accounting for this correction, the overall timing equation of the receiver is given by
Based on 50 trials, the above timing components are profiled and tabulated i n Table 1. Figure 4 summarizes the various timing components of a generic receiver.
Assuming the user receiver has a massive correlator architecture (((1023 correlator/ C/A code chip)/channel), acquisition of the signal and bit synchronization occurs approximately in 2 s for all the channels. It is evident from Table 1 that the major component underlying the TTFF for an open sky user is which needs to be reduced to achieve fast TTFF . Generalizing, in order to minimize TTFF, achievable from satellite should be minimized.
Need for fast TTFF
Subsequent to TTFF analysis, studies were carried out on the ICD and the data sheets of selected civilian and military receivers to establish their TTFF for single and dual frequency operation. In addition, there is no documentary evidence describing the sequence of operation or assumptions with which a military signal receiver operates from power-on leading to TTFF. The datasheets of receiver manufacturers supporting military signals are the sole means to access related TTFF performance.
Assuming this number represents the best achievable TTFF for All In View (AIV) architecture, military signal cold starts are at most comparable to civilian ones.
The following describes an example highlighting the limitations of the LOS TTFF achievable in single frequency. A user with a handheld unit would expect it to output position immediately from power-on. In the scenario shown in Figure 1, it is not practical to have a terrestrial link. This requires that the TTFF achievable from the satellite be minimal.
H owever, from a single frequency perspective the existing (GPS and GLONASS L1) or the emerging (GPS L2C, GPS L5 and GALILEO) systems are not optimal w.r.t TTFF. With some improvements in FNAV method of NAV data transmission in GALILEO, the TTFF is reduced in dual frequency .
The above shows how the TTFF of the IRNSS might be enhanced in order to lead to better performance. IRNSS a regional system, the benefits should be derived in single and dual mode of receiver operation . In addition, if an additional frequency to the existing proposed L5 and S1 of IRNSS is added, it can assist the TTFF of GNSS in AGPS/HOT start mode of operation . The parameters to be established are given in Table 2.
This article described the top level operations of a GNSS receiver leading to position estimation. The methods employed within a receiver and using AGPS external assistance to enhance signal acquisition and reduce the TTFF were introduced. The analysis of TTFF w.r.t various receiver parameters were presented. It was shown that the TTFF is largely dictated by Teph. Using a real world scenario, the need for an effective LOS TTFF from the satellites was illustrated. TTFF being a design parameter for any GNSS system, the article presented requirements under different mode of operation for civilian and military cases that will assist in establishing optimal IRNSS TTFF parameters.
The first author would like to thank the management of Accord Software & Systems Pvt Ltd India for partially supporting his doctoral studies.
Ka plan, E.D. and C. Hegarty (2006) Understanding GPS Priciples and Applications, Artech House
Van Diggelen, F. (2009) A-GPS: Assisted GPS, GNSS and SBAS, Artech House
Bao, J. and Y. Tsui (2004) Fundamentals of Global Positioning System Receivers: A Software Approach, Wiley Series
Rao, V. G., G. Lachapelle, and M. Sashidharan (2011) “Proposed LOS Signal Design for IRNSS to Reduce TTFF in a Single Frequency Receiver,” in Proceedings of the GNSS Signals 2011 Workshop, Toulouse, 12 pages
Galileo (2008) Galileo Open Service Signal In Space Interface Control Document, OS SIS ICD, Draft I, European Space Agency / European GNSS Supervisory Authority, http://www. gsa.europa.eu/go/galileo/os-sis-icd
Rao, V. G., G. Lachapelle, and S.B. Vijaykumar (2011) “Analysis of IRNSS over Indian Sub-continent,” in Proceedings of the International Technical Meeting, , San Diego, The Institute of Navigation, 13 pages