User equipment and method for mobility prediction using artificial intelligence

Abstract

The invention provides user equipment and a method for performing mobility prediction by using artificial intelligence. In one novel aspect, UE mobility prediction is performed through machine learning techniques based on UE measurement data. In an embodiment, the UE obtains a set of mobility-related data, inputs the set of mobility-related data into a mobility AI model for UE mobility prediction, and obtains UE mobility prediction based on the mobility AI model. In an embodiment, UE mobility prediction is range based. In another embodiment, two independent AI models are applied to predict UE mobility in different cases, such as within and not within the service area. In an embodiment, the UE obtains mobility feedback from one or more UE applications and fine-tunes the mobility AI model based on the mobility feedback.

Description

Technical Field

[0001] This invention generally relates to a wireless communication, and more particularly, to an apparatus and method for predicting user equipment (UE) mobility using artificial intelligence (AI). Background Technology

[0002] Artificial intelligence (AI) and machine learning (ML) have permeated a wide range of industries, bringing significant productivity gains. In the field of mobile communication systems, these technologies are leading the transformation. Mobile devices are increasingly using AI-ML models to replace traditional algorithms to improve performance, user experience, and reduce complexity / overhead.

[0003] In legacy networks using the 3rd Generation Partnership Project (3GPP) 5G New Radio (NR), AI-ML technologies are being leveraged to address the challenges arising from the increased complexity of anticipated deployments. UE mobility measurement and prediction have traditionally operated based on algorithms such as the Doppler effect. In some cases, estimating UE mobility becomes difficult. Furthermore, in densely populated urban environments, defining multipaths to predict UE mobility is challenging.

[0004] Therefore, UE mobility prediction needs to be improved and enhanced. Summary of the Invention

[0005] This invention provides an apparatus and method for using AI to perform user equipment (UE) mobility prediction or detection. In a novel aspect, UE mobility prediction is performed based on UE measurement data using machine learning techniques.

[0006] In a first aspect, the present invention provides a method for a user equipment (UE) to use an artificial intelligence (AI) model in a wireless network. The method includes: the UE acquiring a set of mobility-related data; inputting the set of mobility-related data into a mobility AI model for UE mobility prediction; and obtaining a UE mobility prediction result based on the mobility AI model. In one embodiment, the mobility AI model predicts the UE's mobility using modem measurement data. In another embodiment, the UE mobility prediction is range-based. In another embodiment, two independent AI models are applied to predict UE mobility under different situations. When the UE is within the service area of ​​the wireless network, a neural network-based in-service model is used. When the UE is outside the service area of ​​the wireless network (OOS), a neural network-based out-of-service model is used. In one embodiment, the set of mobility-related data includes one or more UE data, which includes: one or more UE signal measurements from serving cells with different RX antennas, one or more UE signal measurements from neighboring cells with different RX antennas, the number of times the UE serving cell changes within a period of time, UE full-band power scan results, frequency received signal strength indication (RSSI) sniffer results, and timing advance, wherein the one or more UE signal measurements from the serving cell or the neighboring cell include reference signal received power (RSRP) measurement, reference signal received quality (RSRQ) measurement, signal interference plus noise ratio (SINR) measurement, or RSSI measurement.

[0007] In one embodiment, the UE determines a mobility AI model for UE mobility prediction based on one or more selection factors. In one embodiment, the one or more selection factors include whether the UE is within the service area of ​​the wireless network or not within the service area of ​​the wireless network (OOS). In one embodiment, the set of mobility-related data is configured based on the one or more selection factors. In one embodiment, UE mobility prediction is range prediction and generates mobility labels. In one embodiment, the range prediction is labeled as one of a set of mobility feature labels or as a speed range label. In one embodiment, the set of mobility feature labels includes stationary, walking, running, driving, traffic jam, highway, and high speed. In another embodiment, the mobility label is applied to the mobility AI model. In one embodiment, the UE obtains mobility feedback from one or more UE applications and fine-tunes the mobility AI model based on the mobility feedback. In one embodiment, the fine-tuning is performed by the UE on the device. In one embodiment, the mobility AI model is trained by the UE on the device or obtained from the wireless network.

[0008] On the other hand, the present invention provides a user equipment (UE), including a transceiver, a collection module, a mobility module, and a prediction module. The transceiver is used to transmit and receive radio frequency (RF) signals in a wireless network; the collection module is used to acquire a set of mobility-related data; the mobility module is used to perform UE mobility prediction using an artificial intelligence (AI) mobility model based on the set of mobility-related data; and the prediction module is used to obtain the result of the UE mobility prediction. Optionally, the collection module is further used to obtain mobility feedback from applications of one or more UEs; and the mobility module is further used to fine-tune the mobility AI model based on the mobility feedback.

[0009] These and other objects of the invention will be readily understood by those skilled in the art upon reading the following detailed description of the preferred embodiments illustrated in the accompanying drawings. A detailed description will be given in the following embodiments with reference to the accompanying drawings. Attached Figure Description

[0010] A more complete understanding of the invention can be obtained by reading the following detailed description and referring to the examples given in the accompanying drawings.

[0011] Figure 1 This is a schematic system diagram illustrating an exemplary wireless network that supports UE mobility prediction using AI, according to embodiments of the present invention.

[0012] Figure 2 An exemplary top-level diagram for UE mobility detection using AI is shown in an embodiment of the present invention.

[0013] Figure 3 The present invention illustrates different UE mobility AI models and a schematic diagram for selecting the appropriate model for using AI to predict UE mobility.

[0014] Figure 4 An exemplary program diagram for range prediction of UE mobility prediction using AI is shown in an embodiment of the present invention.

[0015] Figure 5 An exemplary flowchart for predicting UE mobility using AI is shown in an embodiment of the present invention.

[0016] In the following detailed description, numerous specific details are set forth for illustrative purposes so that those skilled in the art can more thoroughly understand the embodiments of the invention. However, it will be apparent that one or more embodiments may be practiced without these specific details, and different embodiments may be combined as needed, and should not be limited to the embodiments illustrated in the accompanying drawings. Detailed Implementation

[0017] The following description illustrates preferred embodiments of the present invention and is intended only to exemplify the technical features of the invention, not to limit the scope of the invention. Throughout this specification and claims, certain terms are used to refer to specific elements. Those skilled in the art should understand that manufacturers may use different names for the same element. Therefore, this specification and claims do not distinguish elements by differences in name, but rather by differences in function. The terms "element," "system," and "device" used in this invention can refer to computer-related entities, where the computer can be hardware, software, or a combination of hardware and software. The terms "comprising" and "including" as used in the following description and claims are open-ended terms and should be interpreted as "comprising, but not limited to...". Furthermore, the term "coupled" refers to an indirect or direct electrical connection. Therefore, if a device is described as coupled to another device, it means that the device can be directly electrically connected to the other device, or indirectly electrically connected to the other device through other means or connections.

[0018] Unless otherwise indicated, the corresponding numbers and symbols in the various figures generally refer to the corresponding parts. The figures are drawn to clearly illustrate the relevant parts of the embodiments and are not necessarily drawn to scale.

[0019] The terms "basically" or "roughly" as used in this document mean that, within an acceptable range, a person skilled in the art can solve the technical problem to be solved and basically achieve the desired technical effect. For example, "roughly equal to" means a method that a person skilled in the art can accept with a certain margin of error from "exactly equal to" without affecting the correctness of the result.

[0020] Figure 1This is a schematic system diagram illustrating an exemplary wireless network supporting user equipment (UE) mobility prediction according to embodiments of the present invention. Wireless network 100 includes one or more fixed base infrastructure units (also simply referred to as "base station units") forming a network distributed across a geographical area. Base station units may also be referred to as access points, access terminals, base stations, Node-B, eNode-B (eNB), gNB, or other terms used in the art. For example, a base station provides services to multiple mobile stations located within a service area (such as a cell or cell sector). In some systems, one or more base stations are coupled to a controller to form an access network coupled to one or more core networks. gNBs 106, 107, and 108 are base stations in the wireless network whose service areas may overlap or not. For example, user equipment (UE) 101 or mobile station 101 is located within the service area covered by gNBs 106 and 107. For example, UE 101 or mobile station 101 is located within the service area of ​​gNB 106 and is connected to gNB 106. UE 102 or mobile station 102 is located outside the service area of ​​wireless network 100 (out of service, OOS) and is not connected to any base station in wireless network 100. gNB 106 is connected to gNB 107 via Xn interface 121. gNB 106 is connected to gNB 108 via Xn interface 122. A 5G network entity (such as a controller, referred to as 5GC in the diagram) 109 is connected to gNBs 106, 107, and 108 via NG connections 131, 132, and 133, respectively.

[0021] For example, UE 101 moves from 101a to 101b. Both 101a and 101b are within the service area / range of gNB 106. In another case, UE 101 moves from 101a to 101c, where UE 101c is within the service area of ​​gNB 107. As another example, UE 102, which is out of service (OOS), moves from 102a to 102b, where 102b is also outside the service area of ​​wireless network 100. In another case, UE 102 moves from 102a to 102c, where 102c is within the service area of ​​gNB 107. In conventional UE mobility prediction 170, it is difficult to detect UE mobility using only modem measurement information (such as serving cell measurements and neighboring cell measurements). In one case 171, measurement data for UEs out of service (OOS), such as UE 102, is limited. For example, when UE102 moves from 102a to 102b, the available measurement data for UE mobility detection is more limited. In case 172, the Doppler effect formula has limitations. Although the Doppler effect formula can estimate UE mobility through the correlation between the serving cell and the UE, it struggles to estimate mobility when the UE moves along a tangent direction (e.g., UE 101 moves from 101a to 101b). In another case 173, multi-path formation is difficult in densely populated urban areas. Machine learning techniques are well-suited for predicting UE mobility using measurement information.

[0022] In a novel aspect 110, AI-based UE mobility prediction is performed in the wireless network 100. In step 111, the UE acquires its mobility-related data (i.e., data relating to the UE's mobility). In step 112, the UE mobility-related data is input / sent to a mobility AI model (also interchangeably referred to as an AI model) 115. In one embodiment, the AI ​​mobility model 115 is selected by the UE based on one or more selection rules. In step 113, the UE obtains the result of the UE mobility prediction based on the AI ​​mobility model.

[0023] Figure 1 It also illustrates a simplified block diagram of a base station and mobile device / UE that supports AI-based UE mobility detection. Figure 1A simplified block diagram of a UE (such as UE 101) is provided. For example, the UE has an antenna 166 for transmitting and receiving radio frequency (RF) signals. A transceiver 163, coupled to the antenna, receives RF signals from the antenna 166, converts them into baseband signals, and sends them to a processor 162. The transceiver 163 also converts baseband signals received from the processor 162 into RF signals and sends them to the antenna 166. The processor 162 processes the received baseband signals and invokes different functional modules to perform functions within the UE 101. A memory 161 stores program instructions and data 165 to control the operation of the UE 101.

[0024] The UE also includes a set of control modules that perform functional tasks. These control modules can be implemented through circuitry, software, firmware, or a combination thereof. For example, a collection module 191 is used to acquire a set of mobility-related data. A mobility module 192 is used to perform UE mobility prediction using an artificial intelligence (AI) mobility model based on this set of mobility-related data. A prediction module 193 is used to acquire the results of the UE mobility prediction.

[0025] Figure 1 A simplified block diagram of a base station (such as gNB 106) is further illustrated. The gNB has an antenna 156 for transmitting and receiving radio frequency (RF) signals. A transceiver 153, coupled to the antenna 156, receives RF signals from the antenna 156, converts them into baseband signals, and sends them to a processor 152. The RF transceiver 153 also converts baseband signals received from the processor 152 into RF signals and sends them to the antenna 156. The processor 152 processes the received baseband signals and invokes different functional modules to perform functions within the gNB 106. A memory 151 stores program instructions and data 154 to control the operation of the gNB 106. The gNB 106 also includes a set of control modules 158 for performing functional tasks related to communication with mobile stations. These control modules can be implemented through circuitry, software, firmware, or a combination thereof.

[0026] Figure 2An exemplary top-level diagram for UE mobility detection using AI is illustrated according to an embodiment of the present invention. In step 210, the UE acquires a set of mobility-related data (also interchangeably referred to as a "mobility-related dataset"). The UE performs measurements with a modem measurement data provider or other data provider; that is, the UE performs measurements and obtains measurement data through a modem module or other modules within the UE that can provide measurement data, to obtain the mobility-related data. The UE collects one or more UE mobility-related datasets, for example, UE mobility-related dataset #1211 and UE mobility-related dataset #2212. In step 220, the UE inputs one or more mobility-related datasets into a mobility AI model for UE mobility prediction. In one embodiment, the UE selects a mobility AI model for UE mobility prediction based on one or more selection factors 225. In one embodiment, the selection factors include whether the UE is in a service area (labeled "IN SERVICE" in the diagram) or not in a service area (i.e., outside the service area, labeled "OOS" in the diagram). If the UE is in-service (i.e., "on-network state"), in step 221, the UE applies the serving mobility AI model #1 (i.e., the first mobility AI model). If the UE is out-of-service (i.e., "off-network state"), in step 222, the UE applies the unserving mobility AI model #2 (i.e., the second mobility AI model, different from the first mobility AI model). In step 230, the UE performs mobility prediction based on the mobility AI model to obtain the UE mobility prediction result. In one embodiment 231, the mobility prediction is a range prediction and outputs a velocity range label. In one embodiment 235, the range label is dynamically configured and can be dynamically updated. In one embodiment 236, the range label is one of a set of mobility characteristics. For example, some examples of this set of mobility characteristics may include static, walking, running, driving, traffic jam, freeway, and high speed. In another embodiment 237, the range label can be a speed range label, for example, 10 to 15 km / hr. In one embodiment, the range label is predefined or dynamically configured and used for training UE mobility prediction. In another embodiment, the prediction described above is velocity prediction.

[0027] In one embodiment, the UE performs feedback on the UE mobility prediction and fine-tunes based on the feedback. In step 250, the generated UE mobility prediction result is sent to the modem application user, such as the UE's mobility services. In step 251, the modem application user provides feedback on the UE mobility prediction. In step 231, the result of the UE mobility prediction can also be provided as feedback input. In step 260, the mobility AI model is fine-tuned based on the application user's feedback and the output / result of the UE mobility prediction. In one embodiment, fine-tuning is performed on the device. In one embodiment, fine-tuning is performed on a selected mobility AI model based on one or more selection rules. In one embodiment, when the UE is within a service area, the fine-tuning is performed on a serviced mobility AI model (i.e., the first mobility AI model); when the UE is not within a service area, the fine-tuning is performed on a non-serviced mobility AI model (i.e., the second mobility AI model).

[0028] Figure 3 This invention illustrates different mobility AI models and a schematic diagram of selecting a UE mobility prediction model using AI. In one embodiment, two independent AI models are applied to predict UE mobility under different conditions. When the UE is in a service-enabled (under / in service) scenario, the service-enabled model based on a neural network (NN) is applied. When the UE is in an out-of-service (out-of-service) scenario, the out-of-service model based on a neural network (NN) is applied. gNB 303 serves a geographic area. Figure 3In the example, UE 302 is in the service area of ​​the wireless network. UE 301 is not in the service area of ​​the wireless network. When the UE has no service, the measurement data available for UE mobility detection is more limited. In one embodiment, UE mobility-related data includes one or more elements / information (UE data), including: UE signal measurements from serving cells with different RX antennas, one or more UE signal measurements from neighboring cells with different RX antennas, the number of times the UE's serving cell changes over a period of time, the UE full band power scan result, the Received Signal Strength Indicator (RSSI) sniffer result, and time advance. The one or more UE signal measurements from the serving cell or neighboring cells include Reference Signal Received Power (RSRP) measurement, Reference Signal Received Quality (RSRQ) measurement, Signal-to-Interference-plus-Noise Ratio (SINR) measurement, or RSSI measurement. In one embodiment, the out-of-service dataset 320 includes band / frequency level scan results. The in-service dataset 310 includes measurement information of the UE's serving cell and / or neighboring cells. In one embodiment, the UE determines / selects its mobility AI model based on one or more factors / conditions 330. In one embodiment, one or more factors / conditions 330 refer to whether the UE is in or not within the service area of ​​the wireless network. In one embodiment, factors / conditions 330 are applicable to different applications, including AI model application 350, feedback procedure 360, and AI model training 370. For example, when the UE is determined to be in the service area of ​​the wireless network, the in-service mobility AI model is applied to obtain UE mobility prediction, and / or, the UE collects the in-service dataset and feeds it to the in-service mobility AI model for training, and / or, user feedback is collected for fine-tuning the in-service mobility AI model.When a UE is determined to be outside the service area of ​​a wireless network, a serviceless mobility AI model is applied to obtain UE mobility predictions, and / or, the UE collects a serviceless dataset and uses it to train the serviceless mobility AI model, and / or, feedback from the application user is collected to fine-tune the serviceless mobility AI model.

[0029] Figure 4 An exemplary process for range prediction using AI for UE mobility prediction is illustrated according to embodiments of the present invention. In one embodiment, UE mobility prediction is range prediction. In step 411, the UE acquires a mobility-related dataset. In step 412, optionally, the UE acquires feedback information for UE mobility prediction. In step 421, the UE determines which mobility AI model to apply based on one or more factors / conditions. In one embodiment, the factors / conditions include whether the UE is within or outside the service area of ​​the wireless network. In step 430, the mobility AI model is trained using the mobility-related dataset and optional feedback information / feedback dataset. The mobility-related dataset and feedback dataset are selected based on one or more factors / conditions. For example, when the UE is within the service area of ​​the wireless network, the in-service dataset and optional in-service feedback dataset are used to train the in-service AI model. When the UE is outside the service area of ​​the wireless network, the out-of-service dataset and optional out-of-service feedback dataset are used to train the out-of-service AI model. In one embodiment 431, the mobility AI model is trained by the UE. The UE acquires mobility-related data and optionally acquires additional mobility-related data from the network side. Optionally, the UE acquires feedback datasets from application users and, optionally, additional feedback data from the network side. The UE uses the mobility-related datasets and the optional feedback datasets to train a mobility AI model on the device. In another embodiment 432, the mobility AI model is trained by the network side. The UE collects mobility-related datasets and sends the collected mobility-related datasets to the network side. Optionally, the UE acquires feedback datasets from the UE's application users and sends the feedback datasets to the network side. The network side uses the mobility-related datasets collected by the UE, the optional network-side mobility-related datasets, and the optional feedback datasets acquired from the UE, the network, or both to train the mobility AI model.

[0030] In one embodiment 401, a mobility AI model is used to obtain a velocity range-based output, which has a mobility characteristic label or a speed range label. In one embodiment 461, the mobility AI model is range-based by applying dynamic mobility characteristic labels or speed range labels, and the UE uses the mobility AI model to obtain range-based predictions (491). In another embodiment 462, the mobility AI model is not range-based and generates velocity predictions. In step 481, the velocity predictions are generated using the mobility AI model. In step 482, a range label is appended to the generated velocity predictions.

[0031] Figure 5 An exemplary flowchart for using AI to predict UE mobility is shown according to an embodiment of the present invention. In step 501, the UE acquires a set of mobility-related data. In step 502, the UE inputs the set of mobility-related data into a mobility AI model for UE mobility prediction. In step 503, the UE obtains the result of the UE mobility prediction based on the mobility AI model.

[0032] While the invention has been described by way of example and according to preferred embodiments, it should be understood that the invention is not limited to the disclosed embodiments. Rather, it is intended to cover various variations and similar structures (as will be apparent to those skilled in the art), such as combinations or substitutions of different features in different embodiments. Therefore, the scope of the appended claims should be given the broadest interpretation to cover all such variations and similar structures.

Claims

1. A method for user equipment (UE) to perform mobility prediction using artificial intelligence (AI) in a wireless network, comprising: The UE acquires a set of mobility-related data; The mobility-related data is input into the mobility AI model used for UE mobility prediction; as well as The mobility AI model is used to obtain the UE mobility prediction results.

2. The method as described in claim 1, wherein, The method also includes: The mobility AI model used for mobility prediction of the UE is determined based on one or more selection factors.

3. The method as described in claim 2, wherein, The selection factors include: The UE is either within the service area of ​​the wireless network or outside the service area of ​​the wireless network (OOS).

4. The method of claim 2, wherein, This set of mobility-related data is configured based on one or more of these selection factors.

5. The method of claim 1, wherein, This set of mobility-related data includes one or more UE data, which includes: One or more UE signal measurements from serving cells with different RX antennas, one or more UE signal measurements from neighboring cells with different RX antennas, the number of times the UE serving cell changes within a period of time, UE full-band power scan results, frequency received signal strength indication (RSSI) sniffer results, and timing advance, wherein the one or more UE signal measurements from the serving cell or the neighboring cell include reference signal received power (RSRP) measurement, reference signal received quality (RSRQ) measurement, signal interference plus noise ratio (SINR) measurement, or RSSI measurement.

6. The method of claim 1, wherein, The UE mobility prediction is a range prediction and generates mobility labels.

7. The method of claim 6, wherein, The mobility label is one of a set of feature labels or a speed range.

8. The method of claim 6, wherein, The mobility label was generated by the mobility AI model.

9. The method of claim 1, wherein, The method also includes: Obtain mobility feedback from one or more UE applications; and The mobility AI model is then fine-tuned based on this mobility feedback.

10. The method of claim 9, wherein, This fine-tuning is performed by the UE on the device.

11. The method of claim 1, wherein, The mobility AI model is either trained on the UE on the device or obtained from the wireless network.

12. A user equipment (UE), comprising: A transceiver is used to send and receive radio frequency (RF) signals in a wireless network. The collection module is used to acquire a set of mobility-related data; The mobility module is used to perform UE mobility prediction using an artificial intelligence (AI) mobility model based on this set of mobility-related data. as well as The prediction module is used to obtain the results of UE mobility prediction.

13. The UE as described in claim 12, wherein, The mobility module is also used to: determine the mobility AI model for predicting the mobility of the UE based on one or more selection factors, wherein the one or more selection factors include whether the UE is within the service area of ​​the wireless network or outside the service area of ​​the wireless network (OOS).

14. The UE as described in claim 13, wherein, This set of mobility-related data is configured based on one or more of these selection factors.

15. The UE as claimed in claim 12, wherein, This set of mobility-related data includes one or more UE data, which includes: One or more UE signal measurements from serving cells with different RX antennas, one or more UE signal measurements from neighboring cells with different RX antennas, the number of times the UE serving cell changes within a period of time, UE full-band power scan results, frequency received signal strength indication (RSSI) sniffer results, and timing advance, wherein the one or more UE signal measurements from the serving cell or the neighboring cell include reference signal received power (RSRP) measurement, reference signal received quality (RSRQ) measurement, signal interference plus noise ratio (SINR) measurement, or RSSI measurement.

16. The UE as claimed in claim 12, wherein, The UE mobility prediction is a range prediction and generates mobility labels.

17. The UE as claimed in claim 16, wherein, The mobility label was generated by the mobility AI model.

18. The UE as claimed in claim 12, wherein, The collection module is also used to obtain mobility feedback from one or more UE applications; and the mobility module is also used to fine-tune the mobility AI model based on the mobility feedback.

19. The UE as described in claim 18, wherein, This fine-tuning is performed by the UE on the device.

20. The UE as claimed in claim 12, wherein, The mobility AI model is either trained on the UE on the device or obtained from the wireless network.

Images