WO2023117483A1

WO2023117483A1 - Method for efficient determination of deep coverage operation mode for plmn selection in cellular iot

Info

Publication number: WO2023117483A1
Application number: PCT/EP2022/085138
Authority: WO
Inventors: Mathias KURTH
Original assignee: Shenzhen GOODIX Technology Co., Ltd.
Priority date: 2021-12-23
Filing date: 2022-12-09
Publication date: 2023-06-29

Abstract

The invention relates to a method for efficient determination of deep coverage operation mode for PLMN selection in cellular loT access technology. The objective of the invention to present a method that is able to limit the deep coverage PLMN scans using NB- IoT access technology to carrier frequencies where noise-limited NB- IoT deployments can be found will be solved by - preparing a list of carrier frequencies by a UE according to a list of supported RE bands to search for PLMNs on; - selecting a carrier frequency out of the list of carrier frequencies and executing a PLMN scan at a normal coverage level by receiving and processing synchronization signals as well as reading and processing a SIB1 broadcasted by an eNodeB; if the PLMN scan fails, the normal coverage level is increased step-wise into a direction of deep coverage level, wherein the method further comprises the steps : - reusing time-contiguous RF samples received during the PLMN scan on the selected carrier frequency within a power detection procedure to detect whether the carrier frequency is occupied by another RE system, whereas - the power detection procedure comprises at least three threshold detection tasks each are configured to output a Boolean value indicating whether the carrier frequency is occupied by another RE system, and as soon as one of the at least three threshold detection tasks report that the selected carrier frequency is occupied by another RF system said selected carrier frequency is skipped from PLMN scan and is stored into an internal state memory of the UE.

Description

Method for efficient determination of deep coverage operation mode for PLMN selection in cellular loT

The invention discloses a method for ef ficient determination of deep coverage operation mode for PLMN selection in cellular loT access technology .

NarrowBand Internet-of-Things (NB- IoT ) is a 3GPP standard for the radio interface between a base station ( eNodeB ) and a user equipment (UE ) . The obj ectives of the standard are power ef ficient transport of infrequent small data with battery li fetimes of up to 10 years , extension of the cellular coverage , the support of a large number of devices within a cell as well as low complex devices with low costs [ 3GPP TR 23 . 720 ] .

Coverage Extension in NB- IoT

The deployment and mobility characteristics of ToT devices are much more diverse compared to mobile phones and, thus , the requirements in terms of cellular coverage are higher . When a smart-phone loses the network connectivity in a basement , tunnel or other signal-challenging location, it creates the inconvenience for the user to move out of this place to reacquire the signal . For an ToT device deployed in such a signal-challenging location, this results in a permanent service outage . NB- IoT addresses these requirements by allowing for the reuse of the existing cellular network infrastructure in order to benefit from the already existing density of cellular network deployments . In addition, NB- IoT enhances the link budget by an additional 20dB margin compared to LTE, WCDMA or GPRS to achieve a maximum coupling loss (MCL) of 164dB.

NB-IoT approaches the MCL of 164dB by reducing the system bandwidth to 200kHz in downlink and down to 3.75kHz in the uplink. In addition, the standard introduces repetitions of the physical layer signal in both uplink and downlink. As the number of repetitions increases, the code rate decreases and hence, the transmission becomes more reliable. Increasing the number of repetitions, however, increases also the time for reception and transmission. Thus, the operation of the UE within higher coverage levels is generally more energy demanding with adverse effects on the battery lifetime.

PLMN Selection across Coverage Levels

The public land mobile network (PLMN) selection is a procedure that is executed by the UE/IoT device to identify an RE carrier frequency for the UE/IoT device to camp on, so that the UE/IoT device can register with the PLMN and obtain network service. The PLMN selection procedure is executed after switch-on and after recovery from lack of coverage [cf. 3GPP TS 23.122] . For NB-IoT, in particular, the underlying PLMN selection involves a scan of all RE channels in the E-UTRA bands according to its capabilities [cf. 3GPP TS 36.304] .

Most NB-IoT modules inside the UEs available on the market have support for the 3GPP RE bands 1, 3, 5, 8, 20, 28, covering in total 270MHz. In addition, there are NB-IoT modules for the North-American market with additional support of 3GPP bands 2, 4, 12, 13, 17, 18, 19, 25, 66, 26, which adds another 344MHz.

The details of the PLMN selection procedure are not standardized within 3GPP, but left to UE implementation. The state-of-the-art approach for the PLMN selection within NB- loT is shown in figure 1.

According to the list of supported radio frequency (RE) bands (referred to as "rf_bands" in the flow chart of Fig.l) , the UE prepares a list of carrier frequencies (also called E-UTRA Absolute Radio Frequency Channel Numbers (EARFCNs) in NB-IoT) to search for PLMNs on. The UE executes a PLMN scan on the selected carrier frequency, which generally involves the operation of the RE module and the NB-IoT baseband processing to identify the synchronization signals of PLMNs as well as reading the PLMN identification that is broadcast by the eNodeB within System Information Block 1 (SIB1) .

Compared to Evolved Universal Terrestrial Radio Access Network (E-UTRAN) (LTE) , NB-IoT introduces the concept of coverage levels, which needs to be properly controlled within the PLMN selection procedure. As shown in figure 1, the state-of-the-art approach employs an outer loop around the PLMN selection: The PLMN selection is started with "normal coverage", i. e. the best coverage level. In case the PLMN selection fails as indicated by the registration result, the coverage level is increased step-wise into the direction of deep coverage; known steps are from normal coverage (Enhanced Coverage Level-ECLO) over extended coverage (ECL1) to deep or extreme coverage (ECL2) . The strategy to increase the coverage level from normal into deep coverage is beneficial for the overall run-time of the PLMN selection. Assuming that coverage levels are separated by lOdB each, it can also be assumed that the time and, thus, the energy, for the associated PLMN scan increases by a factor of 10.

In the following, figure 2 is used to illustrate the problem underlying this invention. The figure 2 shows the spectrum allocation of a complete RF band having four PLMNs . The PLMNs A, B, and D (solid line) are broadband technologies like 4G or 5G. PLMN C (dashed line) is a standalone NB-IoT deployment of 200kHz. PLMN B is special in the sense that it provides broadband service (solid line) and also operates a so-call in-band NB-IoT network (dashed line) within the broadband spectrum.

The coverage extension of NB-IoT targets noise-limited scenarios. In the example of figure 2, there are the noiselimited PLMNs B and C, i. e. the received signal from the PLMNs is below the UE noise floor. Thus, a NB-IoT UE would need to use coverage extensions to find these PLMNs.

PLMNs A and D in figure 2, on the other hand, are not noiselimited. However, they are also not NB-IoT deployments. The state-of-the-art approach will execute PLMN searches with coverage extensions in the spectrum of PLMNs A and D to find NB-IoT networks that are deeply interfered. However, the coverage extension of NB-IoT are not well suited to combat interference .

In other words, using deep coverage for PLMN searches on carrier frequencies where non NB-IoT deployments have been detected will waste time and energy of the UE . The NB- IoT coverage extension are designed for noise-limited scenarios , they are not applicable to interference limitations . Thus , the state-of-the-art approach for PLMN selection is wasting time and energy executing coverage-enhanced PLMN scans on spectrum of PLMN A and D as shown in figure 3 .

It is therefore an obj ective of the invention to present a method that is able to limit the deep coverage PLMN scans using NB- IoT access technology to carrier frequencies where noise-limited NB- IoT deployments can be found . In other words , a method should be presented that exclude all carrier frequencies from the deep coverage PLMN scans where a potential NB- IoT deployment would be limited due to interference .

The obj ective of the invention will be solved by a method according to independent claim 1 . The method for ef ficient determination of deep coverage operation mode for public land mobile network ( PLMN) selection in cellular Internet- of-Things ( loT ) access technology comprises the following steps :

- preparing a list of carrier frequencies by a user equipment (UE ) according to a list of supported radio frequency (RE) bands to search for PLMNs on;

- selecting a carrier frequency out of the list of carrier frequencies and executing a PLMN scan at a normal coverage level by receiving and processing synchroni zation signals as well as reading and processing a System Information Block 1 , S IB1 , broadcasted by an eNodeB ;

- i f the PLMN scan fails , the normal coverage level is increased step-wise into a direction of deep coverage level , wherein the method further comprises the steps : - reusing time-contiguous RF samples received during the PLMN scan on the selected carrier frequency within a power detection procedure to detect whether the carrier frequency is occupied by another RF system, whereas

- the power detection procedure comprises at least three threshold detection tasks each are configured to output a Boolean value indicating whether the carrier frequency is occupied by another RF system, and

- as soon as one of the at least three threshold detection tasks reports that the selected carrier frequency is occupied by another RF system said selected carrier frequency is skipped from PLMN scan and is stored into an internal state memory of the UE .

There are three main di f ferences to the known state-of-the- art approach :

1 . The power detection procedure is configured to detect the presence of another RF system within the carrier frequency and comprises at least three di f ferent tasks or functions . The time-contiguous RF samples received during the PLMN scan are reused within the power detection procedure : the advantage is that i . e . the power detection step does not increase the RF reception time . The detection is based on received power and will be explained in detail below .

2 . An internal state memory is configured to store whether another RF system has been detected for a particular/ selected carrier frequency out of the list of carrier frequencies .

3 . I f the method returns a decision that another RF system has been detected on the considered carrier frequency, said carrier frequency will be skipped from PLMN scan . Hence , the internal state in the internal state memory of the UE for said carrier frequency is updated to ' true ' , and will be not used for PLMN scan .

In a variant of the inventive method, a first threshold detection task of the at least three threshold detection tasks processes a received signal power of the time- contiguous RF samples within a channel clear assessment and outputs a first Boolean value according to a first configurable threshold .

The channel clear assessment based on energy detection . Similar to IEEE 802 . 11 ( see : IEEE 802 . 11 Speci fication) , the received signal power of the time-contiguous RF samples is observed within an observation window, which is , in the sense of this invention, a defined time duration in which the received time-contiguous RF samples on the selected carrier frequency on a RF channel are considered . I f a configured threshold is exceeded and the corresponding Boolean value is set to ' true ' , the channel is occupied . In other words , when the channel is occupied or "busy" for more than an energy detection threshold within the observation window, the carrier frequency is declared occupied .

In another variant of the inventive method, a second threshold detection task of the at least three threshold detection tasks checks a distribution of the received signal power of the time-contiguous RF samples for unimodality using statistical tests and thresholding and outputs a second Boolean value according to a second configurable threshold .

Thresholding on statistical test for unimodality on the distribution of the received signal power of the time- contiguous RF samples are used . The presence of a bursted RF communication system likely creates additional statistical modes in the distribution of the received signal power . This can be detected by the second threshold detection task . The phrases RF system and RF communication system are used synonymously .

In a variant of the inventive method, a Hartigans ' dip test is applied for testing the unimodality of the received signal power of the time-contiguous RF samples .

The Hartigans ' dip test is one possibility to detect another RF system on the selected carrier frequency as another RF communication system likely creates additional statistical modes in the distribution of the received signal power . The test checks i f statistical data has more than one mode in its distribution . The dip statistic provides information of the probability of the empirical distribution function being bimodal . By having a large value dip, the empirical data is more probable to have multiple modes .

In further variant of the inventive method, a third threshold detection task of the at least three threshold detection tasks detects OFDM-based waveforms in the time- contiguous RF samples , that have cyclic prefixes using a thresholding on an autocorrelation of the received time- contiguous RF samples and outputs a third Boolean value according to a third configurable threshold .

Thresholding on the autocorrelation of the received time- contiguous RF samples is used for detecting OFDM based waveforms that are having cyclic prefixes . Using autocorrelation on the cyclic prefix is a well-known technology for parameter estimation (see: Saputri, Desti Madya . "Autocorrelation Method for Cyclic Prefix OFDM Estimation." IJITEE (International Journal of Information Technology and Electrical Engineering) 3.3: 91-97.) , and for detection of OFDM systems in cognitive radio (see: Chaudhari, Sachin, Visa Koivunen, and H. Vincent Poor.

"Autocorrelation-based decentralized sequential detection of OFDM signals in cognitive radios." IEEE transactions on signal processing 57.7 (2009) : 2690-2700.) . However, it is also applicable for the presented algorithms/method. Both 4G and 5G cellular systems are OFDM based, but also other relevant RF systems like digital terrestrial television (DVB-T) use OFDM within the same frequency spectrum resources .

All three threshold detection tasks are active simultaneously within the inventive method. The output of each threshold detection task is a Boolean value indicating whether the carrier is occupied by another RF communication system. The output of all tasks is aggregated using logical OR, so that a final decision on the occupancy of the carrier is declared (Boolean value TRUE) as soon as one task reports an occupancy with another RF system (Boolean value TRUE) .

In another further variant of the inventive method, the received signal power of the time-contiguous RF samples is observed for a defined time duration, which is at least 20ms .

For NB-IoT, the best-case time to observe the received signal power of the time-contiguous RF samples is 20ms, but it is often larger or should be larger for real implementations .

In a variant of the inventive method, a state in the internal state memory for each selected carrier frequency out of the list of carrier frequencies is initiali zed to false when the method starts .

At the beginning of the inventive method, the states in the internal memory are initiali zed to false for each carrier frequency . Only i f another RF system is detected on the selected carrier frequency out of the list of carrier frequencies the state for said carrier frequency is turned to true . I f the state in the internal state memory for a carrier frequency out of the list of carrier frequencies is turned to true , this carrier frequency is skipped from the PLMN scan .

In a further variant of the inventive method, the at least first , second and third threshold is adaptable to characteristics of a used RF module of the UE and to regulate a receiver operating characteristic .

All three power or energy detection procedures above are having configuration parameters , also called thresholds , allowing to adapt the proposed method to the characteristics of the used RF module as well as to regulate the receiver operating characteristic, i . e . the tradeof f between a false positive and a false negative rate .

In a variant of the inventive method, the thresholds are set to a particular value to minimi ze a false positive rate of detected occupied carrier frequencies and to maximi ze the false negative rate of detected occupied carrier frequencies , whereas the threshold values are set during the UE development.

A false positive event corresponds to declaring the carrier frequency occupied even though there is no other RE communication system. Thus, false positives events introduce the risk of functional failure since deep coverage deployment might be missed when skipping the carrier frequency. A false negative event, on the other hand, corresponds to the failure to detect an existing RE communication system. Thus, this event does not affect the functional correctness, but it reduces the performance benefits of the proposed method compared to the state-of- the-art .

In another variant of the inventive method, the thresholds are set to values to achieve a small false positive rate, e. g. 1%, 0,1% or 0,01% false positive rate and are set during the UE development.

Without limiting the scope of the invented method, an example for setting the threshold that is having high practical relevance is setting the thresholds to achieve a fixed small false positive rate, e. g. 1%, 0.1%, or 0.01% false positive rate. The associated false negative rate is then determined by the receiver operating characteristic. The threshold selection is generally part of the device type characterization during the device development since the energy detection thresholds depend on RE type parameters like noise figure.

In a further variant of the inventive method, the power detection procedure reacts also on signals like TV broadcast and any type of cellular downlink as well as uplink signals, especially according to 2G to 5G .

As the proposed method uses a detection method based on received energy, it is able to react also on signals like TV broadcast and any type of cellular downlink as well as uplink signals ( 2G to 5G) .

The proposed method will be even more important when looking forward in time as it can be expected that a ) existing cellular deployments will become more crowed in the future and b ) more RF spectrum is reassigned to cellular technology in order to satis fy the ever-growing demand for cellular service .

The invention will be explained in more detail using exemplary embodiments .

The appended drawings show

Fig . 1 Flow Chart of a PLMN selection supporting coverage levels ( state-of-the-art ) ;

Fig . 2 Example : Spectrum Allocation of four PLMNs within one RF band;

Fig . 3 Time line for state-of-the-art PLMN selection;

Fig . 4 Time line for the proposed PLMN selection according to the inventive method;

Fig . 5 Flow Chart of the proposed PLMN selection supporting coverage levels according to the invention;

Fig . 6 Flow Chart for the energy or power detection procedure according to the invention . In Figure 4 , the resulting time line of PLMN search activities of the proposed method is shown for the previously discussed example of Figure 2 . Within the first stage of PLMN search at normal coverage , there is no di f ference to the state-of-the-art approach . When increasing the coverage level , however, the proposed approach will be able to exclude the spectrum used by PLMN A and D, which saves both time and energy, as PLMN A and D are broadband technologies and are neither noise-limited nor NB- IoT deployments . Applying the inventive method, the spectrum used by PLMN A and D is excluded, because the first threshold for energy detection returns ' true ' and hence the carrier frequencies are skipped from the PLMN scan .

Figure 5 shows the proposed PLMN selection that supports coverage levels according to the inventive method . The power detection step 5 is shown as a separate action in Figure 5 for the sake of clarity . In an optimi zed implementation, however, the PLMN scan and the power detection are combined within a single procedure . Regardless of the implementation details of the PLMN scan, this procedure generally requires time-contiguous receiving of RF samples on the considered carrier frequency for a defined time duration, also called "observation window" . For NB- IoT , the best-case time is 20ms , but it is often larger for real implementations .

The time-contiguous RF samples received during PLMN scan are reused within the power detection procedure 5 , i . e . the power detection step does not increase the RF reception time . At the beginning of the procedure , the coverage level is set 1 to normal coverage (ELO ) and a carrier frequency out of a list of supported radio frequency, RF, bands 2 to search for PLMNs on, is selected 3 . I f said selected carrier frequency was skipped 20 from PLMN scan, that is stored in the internal state memory, a next carrier frequency 3 from the list of carrier frequencies is selected 3 . I f the selected carrier frequency is not skipped 20 from the list of carrier frequencies a PLMN scan 4 and a power detection procedure 5 is performed on said carrier frequency . When the last carrier frequency of the list of carrier frequencies is processed, a PLMN selection 6 is performed . I f the registration 9 is success ful , the method will stop 10 , otherwise a further higher coverage level 7 is selected and the steps of the method starts again with the next higher coverage level 8 .

Figure 6 shows the overall procedure of the power/energy detection procedure 5 .

In a first step 13 , a received signal power of the time- contiguous RF samples is processed within a channel clear assessment 12 and a first Boolean value according to a first configurable threshold is outputted . I f 13 the busy duty cycle exceeds the configured threshold, which was selected during the device development , the RF carrier is declared occupied 18 and the internal state for said carrier frequency is updated to ' true ' and skipped from the PLMN scan and the power detection procedure stops 19 . Otherwise , the method continues with the next second threshold detection task 14 .

In a second step 15 , the received signal power of the time- contiguous RF samples is tested for unimodality 14 using statistical tests and thresholding and a second Boolean value according to a second configurable threshold is outputted . I f 15 the distribution of the received signal power of the time-contiguous RF samples is not unimodal and the second Boolean value returns ' true ' , then the tested carrier frequency of the RF carrier is declared occupied 18 and the internal state for said carrier frequency is updated to ' true ' and skipped from the PLMN scan and the power detection procedure stops 19 . Otherwise , the method continues with the next third threshold detection task 16 .

In the third step 17 , OFDM-based waveforms in the time- contiguous RF samples , that have cyclic prefixes are detected, using a thresholding on an autocorrelation of the received RF samples 16 and a third Boolean value according to a third configurable threshold is outputted . I f an OFDM signal 17 is detected, the tested carrier frequency of the RF carrier is declared occupied 18 and the internal state for said carrier frequency is updated to ' true ' and skipped from the PLMN scan and the power detection procedure stops 19 . Otherwise , the power detection procedure stops 19 .

The benefits of the proposed methods might be varying under real world conditions . Firstly, there might be the case that a UE-conf igured RF band is not available within the geographical region of the UE . This might be the case because the user did not de-configure the unavailable bands , or the UE might be operated in multiple regions . As RF spectrum is scares ( especially in the sub-GHz range ) , it can be expected that RF bands used within the considered region are exhibiting a high degree of utili zation . Thus , the proposed method will show the highest benefits when configuring only RF bands that are available in the considered geographical region . Even in the case the considered RE band is unavailable in the region of interest , the proposed method is still able to generate time and energy benefits as long as other RE systems are actively using the spectrum . As the proposed method uses a detection method based on received energy, it is able to react also on signals like TV broadcast and any type of cellular downlink as well as uplink signals ( 2G to 5G) .

List of Reference Signs define coverage level define list of supported radio frequency bands select carrier frequency out of the list of supported RF bands PLMN scan on selected carrier frequency Power detection procedure PLMN selection Test i f further coverage level is available Select next coverage level Test i f registration on PLMN is success ful Stop Start power detection procedure First step of a channel clear assessment Test i f a first threshold is exceeded Test of unimodality Test i f second threshold is true or false Autocorrelation test for OFDM based systems Test i f third threshold is true or false Update internal state of the internal state memory to skip RF carrier Stop power detection procedure Skip selected carrier frequency

Claims

Claims Method for efficient determination of deep coverage operation mode for public land mobile network, PLMN, selection in cellular Internet-of-Things , loT, access technology, whereas the method comprising the following steps :

- preparing a list of carrier frequencies (2) by a user equipment, UE, according to a list of supported radio frequency, RE, bands to search for PLMNs on;

- selecting (3) a carrier frequency out of the list of carrier frequencies and executing (4) a PLMN scan at a normal coverage level by receiving and processing synchronization signals as well as reading and processing a System Information Block 1, SIB1, broadcasted by an eNodeB;

- if the PLMN scan (4) fails, the normal coverage level is increased step-wise into a direction of deep coverage level, wherein the method further comprises the steps:

- reusing time-contiguous RE samples received during the PLMN scan (4) on the selected carrier frequency (3) within a power detection procedure (5) to detect whether the carrier frequency is occupied by another RE system, whereas

- the power detection procedure (5) comprises at least three threshold detection tasks (12, 14, 16) each are configured to output a Boolean value indicating whether the carrier frequency is occupied by another RE system, and - as soon as one of the at least three threshold detection tasks reports that the selected carrier frequency is occupied by another RF system said selected carrier frequency is skipped ( 18 ) from PLMN scan and is stored in an internal state memory of the UE . Method according to claim 1 , wherein a first threshold detection task ( 12 ) of the at least three threshold detection tasks processes ( 13 ) a received signal power of the time-contiguous RF samples within a channel clear assessment and outputs a first Boolean value according to a first configurable threshold . Method according to claim 1 , wherein a second threshold detection task ( 14 ) of the at least three threshold detection tasks checks ( 15 ) a distribution of the received signal power of the time-contiguous RF samples for unimodality using statistical tests and thresholding and outputs a second Boolean value according to a second configurable threshold . Method according to claim 1 , wherein a third threshold detection task ( 16 ) of the at least three threshold detection tasks detects ( 17 ) OFDM-based waveforms in the time-contiguous RF samples , that have cyclic prefixes using a thresholding on an autocorrelation of the received RF samples and outputs a third Boolean value according to a third configurable threshold . Method according to one of the former claims , wherein the received signal power of the time-contiguous RF samples is observed for a defined time duration, which is at least 20ms. Method according to one of the former claims, wherein a Hartigans' dip test is applied for testing the unimodality of the received signal power of the time- contiguous RF samples. Method according to one of the former claims, wherein a state in the internal state memory for each selected carrier frequency (3) out of the list of carrier frequencies is initialized to false when the method starts . Method according to one of the former claims, wherein the at least first, second and third threshold is adaptable to characteristics of a used RF module of the UE and to regulate a receiver operating characteristic. Method according to one of the former claims, wherein the thresholds are set to particular values to minimize a false positive rate of detected occupied carrier frequencies and to maximize the false negative rate of detected occupied carrier frequencies, whereas the threshold values are set during the UE development. Method according to one of the former claims, wherein the thresholds are set to values to achieve a small false positive rate, e. g. 1%, 0,1% or 0,01% false positive rate and are set during the UE development. Method according to one of the former claims, wherein the power detection procedure reacts also on signals like TV broadcast and any type of cellular downlink as well as uplink signals, especially according to 2G to 5G.