Base station vendor classifier methods and devices

Abstract

A base station information-based vendor classification process includes extracting, decoding, and grouping the base station information. The grouped base station information is a list associated with types of cellular network technology that is filtered based on the types used by the transmitters. The filtered list is converted to vectors of feature / value pairs, encoded, and provided to a pre-trained classifier. The classifier provides a classification confidence for base station vendors, and an action can be taken based on a comparison between the classification confidence and a list of allowed vendors.

Description

BASE STATION VENDOR CLASSIFIER METHODS AND DEVICESCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 557,019, filed February 23, 2024, and entitled, “Base Station Vendor Classifier Methods and Devices.”STATEMENT OF GOVERNMENT INTEREST

[0002] This invention was made with government support under grants ITE-2226460 and OAC -2326928 awarded by the National Science Foundation. The government has certain rights in the invention.FIELD OF THE DISCLOSURE

[0003] Field of the disclosure is cellular wireless communications, specifically, wireless device security.BACKGROUND

[0004] A base station interfaces with wireless phones and other cellular enabled portable devices to provide access to various services, e.g., communications with traditional phone networks, other wireless networks and internet data access. Identification of base station vendor equipment when a portable device connects to a base station can provide benefits, for example, detecting eavesdropping. . Device security against infiltration from adverse base stations can provide security to industry, governments and in conflict situations, including on or near battlefields.

[0005] Present long term evolution (LTE) and fifth generation (5G) networks are logically split into the radio access network (RAN) and mobile core. The RAN includes distributed base stations that wirelessly connect to user equipment (UEs), for example, but not limited to, cellular phones. In commercial networks, the base stations are typically sold by one of five vendors: Ericsson, Nokia, Samsung, Huawei, and ZTE. In the United States, Europe and other places, Ericsson, Nokia, and Samsung are typically considered benign. The United States and other governments are concerned with the possibility that Huawei and ZTE could have been required to place backdoors in their base station software at the behest of Chinese intelligence. Such backdoors could allow United States adversaries to surveil and eavesdrop oncommunications in cellular networks around the world. This ability to eavesdrop presents a security threat to sensitive communications, including immediate storage and later decryption of cyber-attacks, tracking of device and user movements, and the identification of users based upon communication metadata.

[0006] A System Information Block (SIB) is a component of the LTE and 5G network stacks. The SIB carries information that enables a UE to access a cell (i.e. a geographic area covered by a base station), perform cell re-selection, and provide information related to intrafrequency, inter-frequency, and inter-radio access technology (RAT) cell selections. SIB messages are numbered from 1 to 13 in LTE and from 1 to 21 in 5G, belong to the radio resource control (RRC) layer, and are broadcasted by the base station in the Physical Data Shared Channel (PDSCH). In 5G, some SIBs are sent when the UE requests them. Fourth generation (4G) and 5G SIBs are defined by the 3GPP in TS 36.331 (see, 3GPP. Evolved universal terrestrial radio access (E-UTRA), RRC, protocol specification (available online)), and TS 38.331 (see, 3GPP. Nr, radio resource control (RRC), protocol specification (available online)), respectively.

[0007] SIB messages include SIB1, SIB2, SIB3, SIB4, and SiB5. SIB1 messages can include cell-access-related information such as a public land mobile network (PLMN) identity list, a PLMN identity, a tracking area code, a cell identity, and a cell status. SIB1 messages can also include cell selection information such as a minimum receiver level, and scheduling information about other SIBs such as SI message type and periodicity, SIB mapping info and SI window length. SIB2 messages provide information regarding access barring such as an access probability factor, an access class barring list, and an access class barring time. SIB2 messages can also include semi-static common channel configuration (CCH) information such as a random access channel (RACH) parameters and physical random access channel (PRACH) configuration, uplink frequency information such as uplink E-UTRA absolute radio frequency channel number (EARFCN), uplink bandwidth, and additional emission information, and UE timers and constants. SIB3 messages include information and parameters associated with intrafrequency cell reselections, SIB4 messages include information related to intra-frequency neighboring cells, and SIB5 messages include information related to inter-frequency neighboring cells.

[0008] Parameters transmitted in the SIBs result from configurations of the layers, such as the physical layer (PHY), the media access control layer (MAC), and the RRC layer, of the base station stack. These configurations and SIBs change depending on the deployment type and the base station policies. Example factors that influence the configurations associated withSIBs are re-selection policies, neighbor cells / frequencies, operator preferences, and base station capabilities. A typical base station can send parameters through the SIBs that are used by the UE to access the cell.SUMMARY

[0009] Systems and methods in accordance with embodiments of the present disclosure include classifying a base station by a device. A system of one or more computers can be configured to perform particular operations or actions by virtue of having software, firmware, hardware, or a combination of them installed on the system that in operation causes or cause the system to perform the actions. One or more computer programs can be configured to perform particular operations or actions by virtue of including instructions that, when executed by data processing apparatus, cause the apparatus to perform the actions. One general aspect includes a method for classifying a base station by a device. The method includes extracting base station information about the base station from data received by electronic user equipment from a transmitter, decoding the base station information into a text-based format, grouping the base station information based on the transmitter, creating a list of the base station information for each transmitter, filtering the list based on cellular network technology used by the transmitter and the electronic user equipment, converting the filtered list into vectors of feature / value pairs, encoding the vectors, and providing the encoded vectors to a classifier. The method also includes receiving from the classifier a classification confidence for a pre-selected number of base station vendors, and taking an action based on a comparison between the classification confidence and allowed vendors. Other embodiments of this aspect include corresponding computer systems, apparatus, and computer programs recorded on one or more computer storage devices, each configured to perform the actions of the methods.

[0010] Implementations may include one or more of the following features. The base station information may include a system information block (SIB). The device may include a modem and / or a mobile hot spot. The filtering may include removing configuration parameters that include carrier specific information, frequency information, or are time-dependent. The configuration parameters may include a mobile network code, a mobile country code, a base station identification, subframe numbers, and the frequency information. Converting the filtered list into vectors may include organizing the filtered list in a tree structure, iterating, using a preorder traversal convention, over the organized filtered list including (a) generating a feature / value pair when a leaf node is encountered, where the feature is a path from the leaf node to a root of the tree structure, and the value is the leaf node, (b) repeating (a) over the leafnodes, and (c) repeating (a) and (b) over the organized filtered list. The classifier may include a multi-layer perceptron (MLP) neural network. The MLP may include an input layer, a first internal layer, a second internal layer, and an output layer. The action may include alerting a user, changing vendors automatically, or instructing the user to change to a specific vendor. Implementations of the described techniques may include hardware, a method or process, or computer software on a computer-accessible medium.

[0011] One general aspect includes a computer system for classifying a base station. The computer system includes a hardware processor, and a non-volatile storage medium storing instructions that when executed by the hardware processor perform operations. The operations may include extracting a SIB from an air interface of the base station, decoding the SI,; converting the extracted SIB into a vector of elements, each element representing a parameter from the extracted SIB, and classifying, using of a pre-trained model, a likely vendor of the base station based at least upon the vector, the likely vendor having a vendor type, the pretrained model being trained from data obtained from base station classifiers of a plurality of vendors. Other embodiments of this aspect include corresponding computer systems, apparatus, and computer programs recorded on one or more computer storage devices, each configured to perform the actions of the methods.

[0012] Implementations may include one or more of the following features. The computer system may include a user radio extracting and decoding the SIBs, and a GPU-powered device, remote from the user radio, converting the extracted SIBs into the vector and classifying the likely vendor of the base station. The operations may include removing duplicate SIBs, allowing one SIB for each of the vendor types, the removing leaving remaining SIBs, and deleting, from the remaining SIBs, the SIBs that include operator-specific information, leaving a set of the SIBs. The operations may include filtering, from the set of the SIBs, the parameters that are operator-specific or vendor-independent. The operations further may include taking an action based on a comparison between the classified likely vendor and a list of allowed vendors. The action may include alerting a user, changing vendors automatically, or instructing the user to change to a specific vendor. The operations further may include: encoding the vectors. Implementations of the described techniques may include hardware, a method or process, or computer software on a computer-accessible medium.

[0013] One general aspect includes a computer program product for classifying base stations by a device. The computer program product comprises a computer readable storage medium having program instructions embodied therewith. The program instructions are executable by a computing device to cause the computing device to perform operations. Theoperations include extracting base station information about the base station from data received by electronic user equipment from a transmitter, decoding the base station information into a text-based format, grouping the base station information based on the transmitter, creating a list of the base station information for each transmitter, filtering the list based on cellular network technology used by the transmitter and the electronic user equipment, converting the filtered list into vectors of feature / value pairs, encoding the vectors, and providing the encoded vectors to a classifier. The operations also include receiving from the classifier a classification confidence for a pre-selected number of base station vendors, and taking an action based on a comparison between the classification confidence and a list of allowed vendors. Other embodiments of this aspect include corresponding computer systems, apparatus, and computer programs recorded on one or more computer storage devices, each configured to perform the actions of the methods.

[0014] Implementations may include one or more of the following features. The device may include a modem and / or a mobile hot spot. Implementations of the described techniques may include hardware, a method or process, or computer software on a computer-accessible medium.

[0015] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present teachings, as claimed.BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate aspects of the present teachings and together with the description, serve to explain the principles of the present teachings.

[0017] FIG. l is a schematic block diagram of a system in accordance with embodiments of the present disclosure;

[0018] FIG. 2 is a schematic block diagram of a SIB decoding module in accordance with embodiments of the present disclosure;

[0019] FIG. 3 is a schematic block diagram of a processing module in accordance with embodiments of the present disclosure;

[0020] FIG. 4 is a schematic block diagram of a prediction module in accordance with embodiments of the present disclosure;

[0021] FIG. 5 is a schematic block diagram of a SIB-based vendor classification process in accordance with embodiments of the present disclosure;

[0022] FIG. 6 is a schematic block diagram of an automatic protection architecture in accordance with embodiments of the present disclosure;

[0023] FIG. 7 is a schematic block diagram of an automatic multi-device protection architecture in accordance with embodiments of the present disclosure; and

[0024] FIG. 8 is a method in accordance with embodiments of the present disclosure.

[0025] It should be noted that some details of the figures have been simplified and are drawn to facilitate understanding rather than to maintain strict structural accuracy, detail, and scale.DETAILED DESCRIPTION

[0026] Reference will now be made in detail to the present teachings, examples of which are illustrated in the accompanying drawings. In the drawings, like reference numerals have been used throughout to designate identical elements. In the following description, reference is made to the accompanying drawings that form a part thereof, and in which is shown by way of illustration specific examples of practicing the present teachings. The following description is, therefore, merely exemplary.

[0027] Methods and devices in accordance with embodiments of the present disclosure classify base station vendors based on information extracted from System Information Block (SIB) messages. A deployed base station retains parameters transmitted through the SIBs using vendor default values. The SIB parameters broadcast by a deployed base station result primarily from hardware and software choices made by the vendor. Vendor default values are not specified by the Third Generation Partnership Project (3 GPP) and are set by vendors according to their implementation and capabilities. Methods and devices in accordance with embodiments of the present disclosure leverage vendor default parameters transmitted via base station configuration messages such as, for example, but not limited to, SIB messages to fingerprint and classify the base station. The SIB message format is defined by the 3GPP. Throughout this description, without loss of generality, SIB messages are used as examples of base station configuration messages.

[0028] Referring now to FIG. 1, a configuration of a base station SIB classifier includes an SIB decoding module, a processing module, and a prediction module. Distinct modules can be deployed in various locations, for example, the SIB decoding module can execute in specialized hardware using software defined radios, whereas the processing and prediction modules can be deployed in GPU-powered machines. Other configurations are possible, for example, a single module can decode the SIB, process the decoded information, and predictthe classification of the base station. In some configurations, the single module executes on a cellular phone. In some configurations, the cellular phone configured to classify base stations can allow other cellular phones to connect to a network through the configured cellular phone.

[0029] In some configurations, the single module executes as a mobile hot spot. In some configurations, the single module executes as a Wi-Fi router. In some configurations, when the mobile hot spot is used to classify the base station vendor, a cellular phone can be placed in airplane mode with Wi-Fi, and calls can be placed by the cellular phone through the mobile hot spot. Such a configuration protects the cellular phone from unwanted surveillance, and the cellular phone internal execution paths are not modified to include base station classification. In some configurations, more than one cellular phone can place calls through the mobile hot spot.

[0030] Referring now to FIG. 2, a SIB decoding module 101 is shown. The SIB decoding module 101 extracts SIBs in a digital form from the air interface of a given base station. The SIB decoding module begins execution when the downlink frequency used by the base station from which the SIBs are extracted is provided to a frequency selector task 201. The frequency selector task 201 tunes a radio receiver that is part of, for example, a software defined radio or a modem, to listen to the downlink frequency. Automatic gain control (AGC) 203 can execute in parallel with the frequency selector task and in the background. AGC 203 is a closed-loop feedback regulating software task that adapts the amplitude of the input signal to the expected amplitude of the receiver. When the AGC 203 is executing, the SIB decoding module 101 attempts to acquire the primary and secondary synchronization signals. In some configurations, the primary synchronization signal is located in the last orthogonal frequency-division multiplexing (OFDM) symbol of the first time slot of the first subframe of the radio frame. The secondary synchronization signal is located in the same subframe as the primary synchronization signal, but one symbol before. These two signals are used by the primary / secondary synchronization signal decoding module 205 to obtain the physical layer cell identity (PCI), which implies that the primary / secondary synchronization signal decoding module 205 is synchronized with the cell because the primary / secondary synchronization signal decoding module 205 knows the location of the reference signals and can proceed with the decoding of cell information. The first message that is decoded is referred to as the master information block (MIB). The MIB is a radio resource control (RRC) channel message used for synchronizing with the network and accessing the cell. The MIB format is defined in the 3GPP. In some configurations, the MIB is transmitted over a broadcast control channel (BCCH) logical channel every 40 ms using six physical resource blocks (PRBs) withquadrature phase-shift keying (QPSK) modulation which occupies 24 bits. In some configurations, the MIB is extracted and decoded by a MIB decoding task 207. The MIB carries information for the usage of the cell, such as the cell bandwidth, sub-frame number (SFN), and physical channel (PHICH) configuration. The information carried in the MIB is used as the input for the SIB1 decoding task 209. The SIB 1, whose format is defined by 3 GPP, carries cell-access-related information, cell selection information, and scheduling information about other SIBs. In some configurations, this information is provided to the SIB decoding task 211 which extracts the SIBs transmitted by the cell, specified in the scheduling information field of SIB1. Decoded SIBs are the output of the SIB decoding task. In some configurations, SIB decoding includes converting the SIB to Abstract Syntax Notation One (ASN. l). ASN.l is an interface description language (TDL) for defining data structures that can be serialized.

[0031] Referring now to FIG. 3, a SIB processing module 103 converts decoded SIBs (for example, a bit array) of a base station into a vector. The elements of the vector represent parameters of the SIBs. The SIB processing module 103 includes a SIB parsing task 303 that parses decoded SIBs 301 that were encoded using ASN.l and converts them into a human- readable format (e.g., XML). The ASN. l description of the SIB structures is specified by the 3GPP. The SIBs as XMLs are interpreted as a tree structure and flattened using, for example, a depth-first search, by, for example, but not limited to, a tree flattening task 305. The result is a hash map where the leaves of the tree are values in the hash map. The key to retrieve the values in the hash map is the path to that leaf. The path is a list of nodes from the root. The SIB’s parameters, the tree leaves, are captured as well as the structure of the SIB, the path. This process is applied to the decoded SIBs, represented by an XML structure, and then grouped together by, for example, but not limited to, a SIB grouping task 307. The SIB grouping task 307 removes duplicate SIBs, leaving one SIB per type, and filters those SIBs that contain operator-specific information, for example, SIB4, SIB5, etc.. The resulting list of SIBs, SIB1, SIB2, and SIB3, is filtered 309 to delete the parameters that are operator-specific or vendorindependent such as, for example, but not limited to, PLMN ID, for example, a mobile network code (MNC) and mobile country code (MCC), IDs, for example, base station ID, cell ID, tracking area, or frequencies, etc., and time-dependent values, for example, sub-frame number (SFN).. In general, parameters that are tied to frequency, time, and carrier are deleted. A vector is generated 311 with the keys (root-to-leaf path) as columns and the values as attributes of the vector. This vector is the output of the SIB processing module.

[0032] Referring now to FIG. 4, a prediction module 105 in accordance with embodiments of the present disclosure is shown. The prediction module 105 predicts the vendor of the basestation based at least on the vector 401 generated by the SIB processing module. Some of the param eters / attributes in the SIBs are categorical, and can be converted 403 to binary format using, for example, but not limited to, One Hot Encoding. The resulting vector is used as the input, along with a pretrained model 409, trained through vectors obtained from base station classifiers of a plurality of vendors, to, for example, but not limited to, a prediction task 405 that generates a vendor suggestion. The prediction task can include a classifier that can take the form of a neural network that is configured as follows. In some configurations, the input layer includes 1500 neurons, the first hidden layer includes 1000 neurons with RELU activation, the second hidden layer includes 25 neurons with RELU activation, the output layer includes 10 neurons with softmax activation. Tensorflow modules include tf.keras. Sequential, tf.keras.layers.Dense for the TF modules. Data from the SIBs are used, except what is excluded in the filtering process. Input is the input vector as described herein, One Hot encoded to Is and 0s. Output is a vector of probabilities, where each vector element corresponds to a vendor

[0033] The prediction is converted 407 to a vendor name 411 (e.g. Huawei, Nokia, Ericsson, Samsung, etc.) that is accompanied by a confidence level that the prediction is accurate. In some configurations, the user is informed of the vendor. In some configurations, the system can automatically switch the user to a pre-selected or dynamically-determined vendor. The classification can be uploaded to a database in which cells, location, and classification are stored. The provider can proactively decide which carrier to use in a given location, and the user can be instructed to manually change the SIM card, or a soft request could be done to change the carrier.

[0034] Referring now to FIG. 5, a SIB-based vendor classification process includes, but is not limited to including, actions such as SIB extraction 503, SIB decoding 505, SIB grouping 507, SIB filtering 509, SIB vectorization 513, feature filtering 515, feature encoding 517, and prediction 519. With respect to SIB extraction 503, the SIB messages are used internally by the modem to establish a radio connection with the network, and so they are not exposed to the operating system through interfaces such as, for example, the Radio Interface Layer (RIL) for Android, or the mobile broadband interface model (MBIM) / Qualcomm internet protocol camera (IPC) Router (QRTR) / Qualcomm mobile station modem (MSM) interface (QMI) protocols for external modem processors. The SIBs along with other internal cellular and radio specific information are exposed to vendor diagnostic tools for debugging and verification purposes through the modem diagnostic channel. The modem diagnostic channel can be used to establish a connection with the modem using the modem diagnostic protocol (specific for each modem vendor) to specify to the modem what information is to be extracted. The requestcan be limited to RRC messages. When the SIB message is extracted from incoming data at entry point 501, the SIB message is encoded using, for example, ASN.l encoding. Other encoding can be used, or the SIB message can be provided in a unencoded form. The ASN.1 format is defined by 3GPP for transmission over the radio air interface. The SIB decoder 505 decodes the SIB into a text-based format for structured data representation such as JSON or XML. SIBs are extracted from the modem and decoded, possibly serially (one after the other). The modem might try to connect with several base stations in a short period of time, and so the SIBs obtained in that short period of time might belong to different base stations. To account for this, the SIBs are grouped 507 based on the base stations that transmitted them, creating a list of SIBs for each base station. A group includes the SIBs used for classification. In LTE, SIB1, SIB2, and SIB3 are used for classification, In 5G SIB1 and SIB2 are used for classification. Other SIBs received from the base station can be ignored. In some configurations, individual SIBs are not classified, groups of SIBs are classified. The SIBs collected from the target base station are converted into a vector where each position of the vector corresponds with a generated feature- value pair. To generate the feature-value pair, the SIB is serialized as follows. The SIB structure (for example, XML or JSON) is iterated over as if it were a tree, using, for example, but not limited to, a preorder traversal convention. When a leaf node is encountered, a feature-value pair is generated, where the feature is the path from the leaf to the root of the tree and the value is the leaf itself, thus capturing the values and their locations and ordering. The process is repeated over the leaves in the tree (SIB). This process is repeated for the SIBs, and the feature-value pairs are concatenated. SIB vectorization 513 concatenates the result of encoding into a vector in which the entries are numeric. When the SIBs have been serialized into feature / value pairs and concatenated into a vector 513, the features are filtered 515. To filter 515 the features, configuration parameters that, for example, but not limited to, contain carrier specific information, frequency information, or are time-dependent are removed. Pairs that are removed can include, but are not limited to including, MCC / MNC, base station IDs, subframe numbers (SFN), and frequency information. The remaining configuration parameters are encoded 517 using, for example, One Hot encoding. The result of the feature encoding 517 is used as the input vector of a classifier, for example, but not limited to, a multilayer perceptron (MLP) neural network. In some configurations, the neural network has four layers: an input layer (for example, 1500 neurons), a first internal layer (for example, 1000 neurons and rectified linear (RELU) activation), a second internal layer (for example, 25 neurons and RELU activation), and an output layer (for example, 10 neurons and softmax activation). The prediction 519 returns, for example, a 10-element vector 521 where theelements include a classification confidence for the ten base station vendors that are supported, for example.

[0035] Referring now to FIG. 6, the SIB -based vendor classification process described above can ensure that cellular equipment data and control network traffic traverse benign infrastructure, for example, but not limited to, a pre-defined list of cellular equipment vendors. This protection process executes in the cellular equipment and starts with a SIB-based vendor classification process 605 that interfaces with a modem 601 in the cellular equipment through the modem’s diagnostic interface 603. The SIB-based vendor classification 605 requests modem internal information using SIB extraction 503 (FIG. 5) to retrieve encoded SIBs. The SIB-based vendor classification 605 outputs a vendor prediction 607 for a serving cell. The serving cell is the base station that the modem 601 is currently using to connect with the network. The vendor prediction is provided to a vendor selection module 609. The vendor selection module 609 compares the received vendor prediction with a list of allowed vendors. The list of allowed vendors is predefined and, in some configurations, can be modified by, for example, a user or an automated monitor. If the predicted vendor is included in the list of allowed vendors, the vendor selection module 609 awaits the modem’ s connecting with another base station to check a vendor prediction. If the predicted vendor is not included in the list of allowed vendors, the vendor selection module 609 provides an alert, and instructs the modem 601 to switch carriers using, for example, but not limited to, come to attention (AT) commands 611 through a serial interface. To monitor the available carriers, the vendor selection module 609 periodically instructs the modem 601 to scan for carriers using the same serial interface and the corresponding AT commands 611 as used to instruct the modem 601 to switch carriers. When selecting a next carrier from the list of available carriers, the vendor selection module 609 uses, for example, a round-robin strategy until it finds a carrier whose base station vendor predictions are included in the list of allowed vendors. If none of the available carriers are included in the list, the vendor selection module 609 can iterate over the available carriers until a vendor is included in the list of allowed vendors, or vendor selection module 609 can connect to the carrier with less confident vendor predictions based on, for example, but not limited to, a priority list of carriers. Automatic protection can be provided for devices with cellular modems, for example, but not limited to, cellular phones, laptops with USB cellular modems, and Wi-Fi+cellular routers.

[0036] Referring now to FIG. 7, multiple devices 711 / 713 can be protected at the same time through the use of a hotspot 707. A device 707, referred to herein as a gateway, that implements automatic protection, serves as a network gateway (hotspot) to devices 711 / 713that route their traffic through the gateway. The gateway device 707 ensures connection to a carrier 705 that uses allowed base station vendors. The gateway device 707 forwards network traffic from the WiFi / Ethernet interface 709 used by client devices 711 / 713 to the cellular interface that is exposed by the cellular modem.

[0037] Referring now to FIG. 8, method 800 for classifying a base station using a device can include, but is not limited to including, but not limited to, extracting 802 base station information about the base station from data received by electronic user equipment from a transmitter at the base station. In some configurations, the device that is classifying the base station is a modem, for example a modem in a cellular phone. In some configurations, the device is a mobile hot spot. In some configurations, the base station information includes SIBs. Other configurations are contemplated by the present disclosure. The method 800 can further include decoding 804 the base station information into a text-based format. In some configurations, the base station information, in, for example, the SIB, is decoded into a textbased format such as, but not limited to, JSON or XML. In some configurations, SIBs are extracted from a modem and decoded. The method 800 can include grouping 806 the base station information based on the transmitter at the base station. The reason for grouping the base station information is because the modem or mobile hot spot, for example, might try to connect with several base stations in a short period of time. In some configurations, a group includes SIBs used for classification. The method 800 includes creating 808 a list of base station information for each transmitter. For example, in LTE technology, SIB1, SIB2, and SIB3 are used for classification, whereas in 5G technology, SIB4 and SIB5 are used for classification. Other possible technologies and classification information are contemplated by the present disclosure. The method 800 includes filtering 810 the list based on the technology (for example, but not limited to, LTE and 5G) used by the transmitter and the electronic user equipment in order to enable the modem or mobile hot spot to ignore other SIBs. The method 800 includes converting 812 the filtered list into vectors of feature / value pairs. In some configurations, the feature / value pairs are created by iterating over the received and filtered SIB as if it were a tree structure until a leaf node is encountered. The feature is the path from the leaf to the root of the tree, and the value is the leaf. In some configurations, the process is repeated over the leaves in the tree, then repeated for the SIBs, and the feature / value pairs are concatenated. In some configurations, the features are filtered by removing configuration parameters such as, but not limited to, carrier specific information, frequency information, and time-dependent information. In some configurations, MCC / MNC, base station IDs, subframe numbers, and frequency information are removed.

[0038] The method 800 includes encoding 814 the remaining configuration parameters (vectors). In some configurations, One Hot encoding is used. The method 800 includes providing 816 the encoded vectors to a classifier. In some configurations, the classifier is a multi-layer perceptron neural network. In some configurations, the neural network has four layers. The method includes receiving 818 a classification confidence for a pre-selected number of base station vendors. The method includes taking 820 an action based on a comparison between the classification confidence and allowed vendors. For example, if there is a high confidence that the vendor associated with the SIB is not an allowed vendor, the modem or mobile hot spot could inform the user. The user could physically replace the equipment’s SIM card with a SIM card of an allowed vendor. In some configurations, the modem or mobile hot spot can automatically switch the user equipment to an allowed vendor. In some configurations, the modem or mobile hot spot can evaluate the possible communication options and choose a best option for communications for the user equipment.

[0039] While the present teachings have been illustrated with respect to one or more implementations, alterations and / or modifications can be made to the illustrated examples without departing from the spirit and scope of the appended claims. In addition, while a particular feature of the present teachings may have been disclosed with respect to only one of several implementations, such feature may be combined with one or more other features of the other implementations as may be desired and advantageous for any given or particular function. As used herein, the terms “a”, “an”, and “the” may refer to one or more elements or parts of elements. As used herein, the terms “first” and “second” may refer to two different elements or parts of elements. As used herein, the term “at least one of A and B” with respect to a listing of items such as, for example, A and B, means A alone, B alone, or A and B. Those skilled in the art will recognize that these and other variations are possible. Furthermore, to the extent that the terms “including,” “includes,” “having,” “has,” “with,” or variants thereof are used in either the detailed description and the claims, such terms are intended to be inclusive in a manner similar to the term “comprising.” Further, in the discussion and claims herein, the term “about” indicates that the value listed may be somewhat altered, as long as the alteration does not result in nonconformance of the process or structure to the intended purpose described herein. Finally, “exemplary” indicates the description is used as an example, rather than implying that it is an ideal.

[0040] It will be appreciated that variants of the above-disclosed and other features and functions, or alternatives thereof, may be combined into many other different systems or applications. Various presently unforeseen or unanticipated alternatives, modifications,variations, or improvements therein may be subsequently made by those skilled in the art which are also intended to be encompasses by the following claims.

Claims

CLAIMS1. A method for classifying a base station by a device, the method comprising: extracting base station information about the base station from data received by electronic user equipment from a transmitter; decoding the base station information into a text-based format; grouping the base station information based on the transmitter; creating a list of the base station information for each transmitter; filtering the list based on cellular network technology used by the transmitter and the electronic user equipment; converting the filtered list into vectors of feature / value pairs; encoding the vectors; providing the encoded vectors to a classifier; receiving from the classifier a classification confidence for a pre-selected number of base station vendors; and taking an action based on a comparison between the classification confidence and allowed vendors.

2. The method of claim 1, wherein the base station information comprises: a system information block (SIB).

3. The method of claim 1, wherein the device comprises: a modem.

4. The method of claim 1, wherein the device comprises: a mobile hot spot.

5. The method of claim 1, wherein the filtering comprises: removing configuration parameters that include carrier specific information, frequency information, or are time-dependent.

6. The method of claim 5, wherein the configuration parameters comprise: a mobile network code, a mobile country code, a base station identification, subframe numbers, and the frequency information.

7. The method of claim 1, wherein converting the filtered list into vectors comprises: organizing the filtered list in a tree structure; iterating, using a preorder traversal convention, over the organized filtered list including:(a) generating a feature / value pair when a leaf node is encountered, where the feature is a path from the leaf node to a root of the tree structure, and the value is the leaf node;(b) repeating (a) over the leaf nodes; and(c) repeating (a) and (b) over the organized filtered list.

8. The method of claim 1, wherein the classifier comprises: a multi-layer perceptron (MLP) neural network.

9. The method of claim 8, wherein the MLP comprises: an input layer, a first internal layer, a second internal layer, and an output layer.

10. The method of claim 1, wherein the action comprises: alerting a user, changing vendors automatically, or instructing the user to change to a specific vendor.I L A computer system for classifying a base station comprising: a hardware processor; a non-volatile storage medium storing instructions that when executed by the hardware processor perform operations comprising: extracting a system information block (SIB) from an air interface of the base station; decoding the SIB; converting the extracted SIB into a vector of elements, each element representing a parameter from the extracted SIB; and classifying, using of a pre-trained model, a likely vendor of the base station based at least upon the vector, the likely vendor having a vendor type, the pre-trained model being trained from data obtained from base station classifiers of a plurality of vendors.

12. The computer system of claim 11, further comprising: a user radio extracting and decoding the SIBs; and a GPU-powered device, remote from the user radio, converting the extracted SIBs into the vector and classifying the likely vendor of the base station.

13. The computer system of claim 11, wherein the operations comprise: removing duplicate SIBs, allowing one SIB for each of the vendor types, the removing leaving remaining SIBs; and deleting, from the remaining SIBs, the SIBs that include operator-specific information, leaving a set of the SIBs.

14. The computer system of claim 13, wherein the operations comprises: filtering, from the set of the SIBs, the parameters that are operator-specific or vendorindependent.

15. The computer system of claim 11, wherein the operations further comprise: encoding the vectors.

16. The computer system of claim 14, wherein the operations further comprise: taking an action based on a comparison between the classified likely vendor and a list of allowed vendors.

17. The computer system of claim 16, wherein the action comprises: alerting a user, changing vendors automatically, or instructing the user to change to a specific vendor.

18. A computer program product for classifying base stations by a device, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to perform operations comprising: extracting base station information about the base station from data received by electronic user equipment from a transmitter; decoding the base station information into a text-based format; grouping the base station information based on the transmitter;creating a list of the base station information for each transmitter; filtering the list based on cellular network technology used by the transmitter and the electronic user equipment; converting the filtered list into vectors of feature / value pairs; encoding the vectors; providing the encoded vectors to a classifier; receiving from the classifier a classification confidence for a pre-selected number of base station vendors; and taking an action based on a comparison between the classification confidence and a list of allowed vendors.

19. The computer program product of claim 18, wherein the device comprises: a modem.

20. The computer program product of claim 18, wherein the device comprises: a mobile hot spot.

Images