Decoding path generation

Abstract

According to an example embodiment, a client device is configured to: receive at least one sidelink signal in a time slot on a plurality of subchannels; identify, using at least one detection model, a plurality of potentially active demodulation reference signal configurations in a time slot across the plurality of subchannels based on the at least one sidelink signal, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot; identify a payload transmission in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot; and create at least one decoding path for the payload transmission in the time slot based on the identified payload transmission in the time slot and the identified plurality of potentially active demodulation reference signal configurations. Apparatus, methods, and computer programs are disclosed.

Description

Technical Field

[0001] The present application generally relates to the field of wireless communications. Specifically, the present application relates to a client for wireless communications and related methods and computer programs. Background Art

[0002] Sidelink (SL) communication allows direct communication between two client devices without going through a base station. In at least some operating modes, SL resource allocation is based on autonomous allocation or selection of resources by the transmitting client device from a preconfigured pool of transmission resources. Resource selection can be based on simple random selection or sensing-based selection. Even after the sensing procedure occurs, collisions between multiple transmissions may occur. Therefore, it may be desirable for a client device to be able to decode multiple overlapping SL transmissions. Summary of the Invention

[0003] The scope of protection sought by various exemplary embodiments of the present invention is set out in the independent claims. Exemplary embodiments and features described in this specification that do not fall within the scope of the independent claims (if any) are to be construed as examples useful for understanding various exemplary embodiments of the present invention.

[0004] An example embodiment of a client device includes at least one processor and at least one memory including computer program code. The at least one memory and the computer program code are configured to, with the at least one processor, cause the client device to: receive at least one sidelink signal in a time slot across multiple subchannels; identify, using at least one detection model, a plurality of potentially active demodulation reference signal configurations from a plurality of possible demodulation reference signal configurations in the time slot across the multiple subchannels based on the at least one sidelink signal, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot; identify a plurality of payload transmissions in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot; and create at least one decoding path for the plurality of payload transmissions in the time slot based on the identified plurality of potentially active demodulation reference signal configurations. For example, the client device can effectively generate the at least one decoding path.

[0005] An example embodiment of a client device includes components for performing the following operations: receiving at least one sidelink signal in a time slot on multiple subchannels; using at least one detection model, identifying a plurality of potentially active demodulation reference signal configurations among a plurality of possible demodulation reference signal configurations in the time slot across the multiple subchannels based on the at least one sidelink signal, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot; identifying a plurality of payload transmissions in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot; and creating at least one decoding path for the plurality of payload transmissions in the time slot based on the identified plurality of payload transmissions in the time slot and the identified plurality of potentially active demodulation reference signal configurations.

[0006] In an example embodiment, alternatively or in addition to the above example embodiments, the demodulation reference signal transmissions in a time slot include at least a demodulation reference signal transmission for a control channel and a demodulation reference signal transmission for a data channel, and wherein the plurality of payload transmissions include at least the control channel and the data channel. For example, the client device may generate at least one decoding path for the control channel transmission and the data channel transmission based on the demodulation reference signal transmissions.

[0007] In an example embodiment, alternatively or in addition to the above example embodiments, the plurality of payload transmissions include physical sidelink control channel (PSCCH) transmissions and / or physical sidelink shared channel (PSSCH) transmissions, and the demodulation reference signal transmissions include PSCCH demodulation reference signal (DMRS) transmissions and / or PSSCH DMRS transmissions. For example, the client device may generate at least one decoding path for the PSCCH and PSSCH transmissions based on the DMRS.

[0008] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one detection model includes at least one machine learning model.For example, the client device can use the at least one machine learning model to effectively identify multiple potentially active demodulation reference signal configurations.

[0009] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one machine learning model includes at least one of: a multi-label multi-class classifier, a neural network, a deep neural network, or a decision forest. For example, the client device can effectively identify multiple potential active demodulation reference signal configurations.

[0010] In an example embodiment, alternatively or in addition to the above example embodiments, at least one machine learning model is configured to take as input modulated symbols or in-phase and quadrature-phase IQ samples of a subchannel of the plurality of subchannels and output a probability value for each of a plurality of possible demodulation reference signal configurations active in the subchannel, wherein each of the plurality of possible demodulation reference signal configurations defines an active demodulation reference signal transmission in the subchannel. For example, a client device can effectively identify a plurality of potentially active demodulation reference signal configurations based on the modulated symbols or IQ samples.

[0011] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to: identify a payload transmission in a time slot based on a plurality of identified potentially active demodulation reference signal configurations in the time slot by performing correlation between symbols across subchannels in the plurality of subchannels, so as to identify payload transmissions occurring across subchannels in the plurality of subchannels. For example, the client device can effectively identify payload transmissions occurring across subchannels.

[0012] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one machine learning model includes a first machine learning model trained for low mobility and a second machine learning model trained for high mobility. For example, the client device can effectively identify multiple potentially active demodulation reference signal configurations in low mobility and high mobility scenarios.

[0013] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to create at least one decoding path for the payload transmission in such a manner that the at least one decoding path maximizes the number of decoded transmissions. For example, the client device can effectively generate the at least one decoding path in such a manner that the number of decoded transmissions is maximized.

[0014] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to create at least one decoding path for the payload transmission by: calculating a reference signal received power (RSRP) for each of a plurality of potentially active demodulation reference signal configurations; ranking the plurality of potentially active demodulation reference signal configurations according to the calculated RSRP for each potentially active demodulation reference signal configuration; and creating the at least one decoding path based on the ranking of the plurality of potentially active demodulation reference signal configurations. For example, the client device can effectively generate the at least one decoding path such that the strongest transmission is decoded first.

[0015] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to: create at least one decoding path based on the permutation by excluding from the at least one decoding path potentially active demodulation reference signal configurations having an RSRP below a preconfigured threshold RSRP. For example, the client device can effectively generate the at least one decoding path in such a manner that the strongest transmission is decoded first based on the threshold RSRP.

[0016] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to decode the payload transmission according to the created at least one decoding path. For example, the client device can effectively decode the transmission.

[0017] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to: store inputs and outputs of the at least one machine learning model in response to the at least one decoding path successfully decoding multiple payload transmissions in a time slot; and retrain the at least one machine learning model using the stored inputs and outputs. For example, the client device can effectively retrain the at least one machine learning model for a new scenario.

[0018] An example embodiment of a method includes receiving at least one sidelink signal in a time slot across a plurality of subchannels; identifying, using at least one detection model, a plurality of potentially active demodulation reference signal configurations among a plurality of possible demodulation reference signal configurations in the time slot across the plurality of subchannels based on the at least one sidelink signal, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot; identifying a plurality of payload transmissions in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot; and creating at least one decoding path for the plurality of payload transmissions in the time slot based on the identified plurality of payload transmissions in the time slot and the identified plurality of potentially active demodulation reference signal configurations. For example, the method can effectively generate the at least one decoding path.

[0019] In an example embodiment, alternatively or in addition to the above example embodiments, the demodulation reference signal transmission in the time slot includes at least a demodulation reference signal transmission for a control channel and a demodulation reference signal transmission for a data channel, and wherein the plurality of payload transmissions includes at least the control channel and the data channel. For example, the method may generate at least one decoding path for the control channel transmission and the data channel transmission based on the demodulation reference signal transmission.

[0020] In an example embodiment, alternatively or in addition to the above example embodiments, the plurality of payload transmissions include physical sidelink control channel (PSCCH) transmissions and / or physical sidelink shared channel (PSSCH) transmissions, and the demodulation reference signal transmissions include PSCCH demodulation reference signal (DMRS) transmissions and / or PSSCH DMRS transmissions. For example, the method may generate at least one decoding path for PSCCH and PSSCH transmissions based on the DMRS.

[0021] In an example embodiment, alternatively or in addition to the above example embodiments, at least one detection model includes at least one machine learning model.For example, the method can use at least one machine learning model to effectively identify multiple potential active demodulation reference signal configurations.

[0022] In an example embodiment, alternatively or in addition to the above example embodiments, at least one machine learning model includes at least one of the following: a multi-label multi-class classifier, a neural network, a deep neural network, or a decision forest. For example, the method can effectively identify multiple potential active demodulation reference signal configurations.

[0023] In an example embodiment, alternatively or in addition to the above example embodiments, at least one machine learning model is configured to take as input modulated symbols or in-phase or quadrature-phase IQ samples of a subchannel of a plurality of subchannels and output a probability value for each of a plurality of possible demodulation reference signal configurations active in the subchannel, wherein each of the plurality of possible demodulation reference signal configurations defines an active demodulation reference signal transmission in the subchannel. For example, the method can effectively identify a plurality of potentially active demodulation reference signal configurations based on the modulated symbols or IQ samples.

[0024] In an example embodiment, alternatively or in addition to the above example embodiments, identifying a payload transmission in a time slot based on a plurality of potentially active demodulation reference signal configurations identified in the time slot includes performing correlation between symbols across subchannels in a plurality of subchannels to identify payload transmissions occurring across subchannels in the plurality of subchannels. For example, this method can effectively identify payload transmissions occurring across subchannels.

[0025] In an example embodiment, alternatively or in addition to the above example embodiments, the at least one machine learning model includes a first machine learning model trained for low mobility and a second machine learning model trained for high mobility. For example, the method can effectively identify multiple potentially active demodulation reference signal configurations in both low mobility and high mobility scenarios.

[0026] In an example embodiment, alternatively or in addition to the above example embodiments, at least one decoding path for the payload transmission is created in such a manner that the at least one decoding path maximizes the number of decoded transmissions. For example, the method can effectively generate at least one decoding path in such a manner that the number of decoded transmissions is maximized.

[0027] In an example embodiment, alternatively or in addition to the above example embodiments, creating at least one decoding path for the payload transmission includes: calculating a reference signal received power (RSRP) for each of a plurality of potentially active demodulation reference signal configurations; rank-ordering the plurality of potentially active demodulation reference signal configurations according to the calculated RSRP for each potentially active demodulation reference signal configuration; and creating at least one decoding path based on the rank-ordering of the plurality of potentially active demodulation reference signal configurations. For example, the method can effectively generate the at least one decoding path such that the strongest transmission is decoded first.

[0028] In an example embodiment, alternatively or in addition to the above example embodiments, at least one decoding path based on the permutation is created by excluding from the at least one decoding path potentially active demodulation reference signal configurations having an RSRP below a preconfigured threshold RSRP. For example, the method can effectively generate at least one decoding path in such a way that the strongest transmission is decoded first based on the threshold RSRP.

[0029] In an example embodiment, alternatively or in addition to the above example embodiments, the method further comprises: decoding the payload transmission according to the created at least one decoding path.For example, the method may efficiently decode the transmission.

[0030] In an example embodiment, alternatively or in addition to the above example embodiments, the method further includes: storing inputs and outputs of the at least one machine learning model in response to the at least one decoding path successfully decoding multiple payload transmissions in a time slot; and retraining the at least one machine learning model using the stored inputs and outputs. For example, the method can effectively retrain the at least one machine learning model for a new scenario.

[0031] An exemplary embodiment of a computer program product comprises a program code configured to perform a method according to any one of the above exemplary embodiments when the computer program product is executed on a computer. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings, which are included to provide a further understanding of example embodiments and constitute a part of this specification, illustrate example embodiments and together with the description help to explain the principles of example embodiments. In the drawings:

[0033] Figure 1 shows example embodiments of the subject matter described herein, illustrating example systems in which various example embodiments of the present disclosure may be implemented;

[0034] Figure 2 An example embodiment of the subject matter described herein is shown, illustrating a client device;

[0035] Figure 3 An example of the subject matter described herein is shown, illustrating a sidelink transmission collision;

[0036] Figure 4 Another example of the subject matter described herein is shown, illustrating a sidelink transmission collision;

[0037] Figure 5 An example of the subject matter described herein is shown, illustrating a sidelink control channel and a sidelink data channel during a transmission collision;

[0038] Figure 6 An example embodiment of the subject matter described herein is shown, illustrating a flow chart of a sidelink transmission decoding process;

[0039] Figure 7 An example embodiment of the subject matter described herein is shown, illustrating a physical sidelink control channel demodulation reference signal configuration;

[0040] Figure 8 An example embodiment of the subject matter described herein is shown, illustrating a physical sidelink shared channel demodulation reference signal configuration;

[0041] Figure 9 An example embodiment of the subject matter described herein is shown, illustrating the flow of data in a machine learning model;

[0042] Figure 10 Another example embodiment of the subject matter described herein is shown, illustrating the flow of data in a machine learning model;

[0043] Figure 11 An example embodiment of the subject matter described herein is shown, illustrating a machine learning model including a component trained for low mobility and a component trained for high mobility;

[0044] Figure 12 An example embodiment of the subject matter described herein is shown, illustrating a flow chart of a decode generation process; and

[0045] Figure 13 An example embodiment of the subject matter described herein is shown, illustrating a method.

[0046] In the drawings, the same reference numerals are used to designate the same parts. DETAILED DESCRIPTION

[0047] Reference will now be made in detail to example embodiments, examples of which are illustrated in the accompanying drawings. The detailed description provided below in conjunction with the accompanying drawings is intended as a description of this example and is not intended to represent the only form in which the present disclosure may be constructed or utilized. This description sets forth the functionality of the example and the sequence of steps for constructing and operating the example. However, the same or equivalent functions and sequences may be implemented by different example embodiments.

[0048] Figure 1 Example embodiments of the subject matter described herein are shown, illustrating example systems in which various example embodiments of the disclosure may be implemented.

[0049] Sidelink (SL) communication between client devices 200 via the PC5 interface is based on the principle of one-to-many broadcast to the sender. This means that, in principle, no connection needs to be established at the radio access level for SL communication between client devices 200, regardless of whether the SL communication is for unicast, multicast, or broadcast services.

[0050] On the one hand, a transmitting (Tx) client device 200_1 can use resources from a (pre-)configured resource pool to transmit a SL to a receiving (Rx) client device 200_2, or a group of Rx client devices 200_2, or all Rx client devices 200_2 in the vicinity of the Tx client device 200_1, at least for transmitting SL control information (SCI) used as a scheduling assignment for SL data transmission. On the other hand, the Rx client device 200_2 may need to continuously monitor the (pre-)configured resource pool to receive SL, at least receiving all SCI instances and determining whether the received SCI and corresponding SL data transmission indicate that the Rx client device 200_2 receives or does not receive the SRC corresponding to the Tx side and the DST corresponding to the Rx side based on the (multiple) source (SRC) and / or destination (DST) IDs indicated in the received SCI instances. This can be applied to all types of SL casting: unicast, multicast, or broadcast.

[0051] There are two resource allocation modes specified for SL transmission, referred to as Mode 1 and Mode 2. Mode 1 is based on the use of scheduled resources or grants from the serving base station (BS). This means that the Tx client device 200_1 may need to be in the RRC connected state of the serving BS in order to obtain the allocated Mode 1 resources. Mode 2 is based on the autonomous allocation or selection of resources by the Tx client device 200_1 from a preconfigured Tx resource pool. Resource selection in Mode 2 can be based on a simple random selection or a sensing-based selection. The latter may be preferred and used for normal operation, while the former may be used for special operations or situations with a specific preconfigured resource pool. Mode 2 can be used for the Tx client device 200_1 in coverage (IC) or out of coverage (OoC), in RRC idle, RRC inactive, or RRC connected state.

[0052] Figure 2 is a block diagram of a client system 200 according to an example embodiment.

[0053] According to an example embodiment, the client device 200 includes one or more processors 202 and one or more memories 204 including computer program code. The client device 200 may also include a transceiver 205 and other elements such as an input / output module ( Figure 2 Not shown) and / or communication interface ( Figure 2 not shown).

[0054] According to an example embodiment, the at least one memory 204 and the computer program code are configured to, with the at least one processor 202, cause the client device 200 to receive at least one sidelink signal in time slots on a plurality of subchannels.

[0055] Client device 200 may receive at least one sidelink signal by, for example, receiving at least one sidelink signal using transceiver 205. Alternatively, some other device / module / component may receive at least one sidelink signal and provide at least one sidelink signal to client device 200, and client device 200 may obtain at least one sidelink signal. In this document, "receive" may mean receiving a signal or obtaining data corresponding to a signal.

[0056] For example, client device 200 may sample, demodulate, and decode at least one sidelink signal before using the at least one sidelink signal for further processing as disclosed herein.

[0057] In this document, a time slot may also be referred to as a time slot for short.

[0058] The client device 200 may also be configured to identify, using at least one detection model, a plurality of potentially active demodulation reference signal configurations among a plurality of possible demodulation reference signal configurations in a time slot across a plurality of subchannels based on the at least one sidelink signal, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot.

[0059] Demodulation reference signal (DMRS) transmissions may include, for example, physical sidelink control channel (PSCCH) DMRS transmissions and / or physical sidelink shared channel (PSSCH) DMRS transmissions.

[0060] The client device 200 may also be configured to identify a plurality of payload transmissions in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot.

[0061] In this document, a payload transmission may refer to any transmission that includes payload data to be received by a client device. The payload data may include, for example, control channel data or data channel data.

[0062] Payload transmissions may include, for example, control channel (such as PSCCH) transmissions and / or data channel (such as PSSCH) transmissions.

[0063] The client device 200 may also be configured to create at least one decoding path for the plurality of payload transmissions in the time slot based on the identified plurality of payload transmissions and the identified plurality of potentially active demodulation reference signal configurations in the time slot.

[0064] In this document, a decoding path may refer to the order in which payload transmissions are decoded. A decoding path may also be referred to as a decoding order, etc.

[0065] The client device 200 may create at least one decoding path in such a way that the at least one decoding path maximizes the number of decoded transmissions.For example, the at least one decoding path may indicate that a PSCCH transmission should be decoded before an associated PSSCH transmission.

[0066] According to an example embodiment, the demodulation reference signal transmission in the time slot includes at least a demodulation reference signal transmission for a control channel and a demodulation reference signal transmission for a data channel, and wherein the plurality of payload transmissions includes at least the control channel and the data channel.

[0067] According to an example embodiment, the multiple payload transmissions include physical sidelink control channel (PSCCH) transmissions and / or physical sidelink shared channel (PSSCH) transmissions, and the demodulation reference signal transmissions include PSCCH demodulation reference signal (DMRS) transmissions and / or PSSCH DMRS transmissions.

[0068] The client device 200 may establish a relationship between the PSCCH DMRS and the PSSCH DMRS based on, for example, correlation of their channels or based on correlation between associated transmitter imperfections, such as imperfections in an oscillator at the transmitter, non-idealities in an amplifier, and the effects of these on the transmitted signal.

[0069] Although the client device 200 may be depicted as including only one processor 202, the client device 200 may include more processors. In an example embodiment, the memory 204 can store instructions, such as an operating system and / or various applications.

[0070] In addition, the processor 202 is capable of executing stored instructions. In an example embodiment, the processor 202 can be implemented as a multi-core processor, a single-core processor, or a combination of one or more multi-core processors and one or more single-core processors. For example, the processor 202 can be implemented as one or more of various processing devices (such as a coprocessor, a microprocessor, a controller, a digital signal processor (DSP), a processing circuit system with or without an attached DSP) or various other processing devices (such as, for example, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a microcontroller unit (MCU), a hardware accelerator, a dedicated computer chip, etc.). In an example embodiment, the processor 202 can be configured to execute hard-coded functionality. In an example embodiment, the processor 202 is implemented as an executor of software instructions, wherein the instructions can specifically configure the processor 202 to perform the algorithms and / or operations described herein when the instructions are executed.

[0071] The memory 204 may be implemented as one or more volatile memory devices, one or more non-volatile memory devices, and / or a combination of one or more volatile and non-volatile memory devices. For example, the memory 204 may be implemented as a semiconductor memory (such as a mask ROM, a PROM (programmable ROM), an EPROM (erasable PROM), a flash ROM, a RAM (random access memory), etc.).

[0072] The client device 200 can be any of various types of devices used by an end-user entity and capable of communicating in a wireless network. Such devices include, but are not limited to, smartphones, tablet computers, smart watches, laptop computers, Internet of Things (IoT) devices, and the like. The client device 200 may include, for example, a mobile phone, smartphone, tablet computer, smart watch, or any handheld or portable device or any other apparatus, such as a vehicle, robot, or repeater. The client device 200 may also be referred to as a user equipment (UE). The client device 200 may communicate with a network node device via, for example, an air / space-borne vehicle communication connection (such as a service link).

[0073] Some of the terms used herein may follow the naming scheme of current forms of 4G or 5G technology. However, this terminology should not be considered limiting, and the terminology may change over time. Therefore, the following discussion of any example embodiment may also apply to other technologies, such as 6G.

[0074] At least some example embodiments disclosed herein may enable identification of decoding paths, which in turn enables the client device 200 to optimize decoding.

[0075] Figure 3 An example of the subject matter described herein is shown, illustrating a sidelink transmission collision.

[0076] In SL resource allocation mode 2, collisions between multiple transmissions may occur even after the sensing procedure has occurred. In SL resource allocation mode 1, this may also occur at the cell edge for sidelink client devices.

[0077] A conflict may be the result of sensing and selecting the same resource at the same time. An example of this is Figure 3 , where a semi-persistently scheduled (SPS) transmission 301 from client device B collides with a transmission 302 from client device C to client device D, and an SPS transmission 303 from client device C collides with a transmission 304 from client device B to client device A.

[0078] Alternatively or additionally, a collision may also be the result of two nearby gNBs allocating the same SL resources to a SL client device in the cell edge.

[0079] Alternatively or additionally, collisions may also be due to mobility. For example, SPS transmissions from different sending client devices may overlap in time. An example of this is shown in Figure 4. Initially, Client Device B can perform an SPS transmission 401 to Client Device A, and Client Device C can perform an SPS transmission 402 to Client Device D without conflict. However, when Client Device C and Client Device D move closer to Client Device A and Client Device B, an SPS transmission 403 from Client Device B to Client Device A may conflict with an SPS transmission 404 from Client Device C to Client Device D, and vice versa.

[0080] It may be desirable for an Rx client device to detect the presence of a collision so as to be able to indicate to its desired Tx client device to avoid it, e.g., by selecting another Tx timeslot and / or to be able to detect and decode as many colliding transmissions as possible, e.g., based on continuous interference cancellation.

[0081] The resource configuration in the sidelink resource pool defines the minimum information required for the Rx UE to be able to decode the transmission, including the number of subchannels, the number of PRBs per subchannel, the number of symbols in the PSCCH, where the time slot has a physical sidelink feedback channel (PSFCH), and other configuration aspects not relevant to the present invention.

[0082] However, details of the actual sidelink transmission (ie, payload) are provided in the PSCCH (SCI Phase 1) for each individual transmission, including: time and frequency resources, DMRS configuration for PSSCH, MCS, PSFCH, etc.

[0083] Figure 5 An example of the subject matter described herein is shown, illustrating a sidelink control channel and a sidelink data channel during a transmission collision.

[0084] Since different transmissions within a time slot (such as PSCCH and PSSCH) can partially or completely overlap each other, in order for the receiver to be able to "straighten out" these transmissions, the receiver should deduce the decoding order, also known as decoding path, on a time slot basis. Otherwise, transmissions may be lost.

[0085] Before a PSSCH transmission can be decoded, the associated PSCCH should be detected and decoded first. However, if the PSCCH overlaps with another stronger transmission (such as PSCCH or even PSSCH), the stronger transmission may need to be decoded first and subtracted from the composite signal.

[0086] It should be noted that even if the PSCCH is decoded, the associated PSSCH may not be successfully decoded.

[0087] Figure 5The example illustrates a collision of various sidelink transmissions sent over multiple subchannels 510 in a time slot 511. A PSSCH transmission 501 by UE D is interrupted by a PSCCH transmission 502 by UE E and a PSSCH transmission 503 by UE B. A PSSCH transmission 504 by UE C in time slot 511 also does not occur to interrupt the PSCCH transmission 501 by UE D.

[0088] Figure 6

[0014] An example embodiment of the subject matter described herein is shown, illustrating a flow diagram of a sidelink transmission decoding process.

[0089] In operation 601, the client device 200 may transition to the next time slot. Figure 6 The process presented in can be repeated for each time slot.

[0090] In operation 602, the client device 200 may detect all potential transmissions within the time slot 511. This may include detecting PSCCH DMRS in all subchannels 510 and detecting all possible PSSCH DMRS configurations that exist across all subchannels 510. This may be accomplished using, for example, a multi-label, multi-class classifier customized for a specific PSCCH DMRS and PSSCH DMRS configuration. Specifically, as disclosed herein, each classifier may output a probability value for each configuration of each channel active in the time slot 511.

[0091] In operation 603, the client device 200 may identify characteristics of the transmission, such as PSSCH details, based on the detected PSCCH (when these are decodable), for example by decoding the first stage SCI. This operation may be performed based on the soft output of the classifier, and the client device 200 may decide which configurations to test based on, for example, the output probability being greater than a threshold (e.g., p=0.5).

[0092] In operation 604 , the client device 200 may use the output of operation 603 to create PSCCH and PSSCH decoding paths.

[0093] Operations 602 to 604 may construct a decoding path.

[0094] In operation 605, client device 200 may subtract the decoded transmission from all subchannels 510 in time slot 511. In the case of an iterative decoder, client device 200 may return to operation 602 to identify the remaining transmissions.

[0095] In operation 606, the client device 200 may identify transmissions in time slots across all sub-channels 510. For example, the client device 200 may detect all DRMS configurations that are present.

[0096] The client device 200 may check whether additional transmissions are detected in operation 607. If additional transmissions are detected, the client device 200 may return to operation 603. Otherwise, the client device 200 may return to operation 601 to repeat the process for other time slots.

[0097] Figure 7 An example embodiment of the subject matter described herein is shown, illustrating a physical sidelink control channel demodulation reference signal configuration.

[0098] An example of a possible PSCCH DMRS configuration 700 is shown in Figure 7 The PSCCH duration is part of the PSCCH DMRS configuration resource pool. Figure 7 The configurations depicted in may correspond to configurations that the SL Tx client device 200 may select from. The PSCCH DMRS configurations may be randomized uniformly to increase robustness to collisions.

[0099] Figure 8 An example embodiment of the subject matter described herein is shown, illustrating a physical sidelink shared channel demodulation reference signal configuration.

[0100] An example of a possible PSSCH DMRS configuration 800 is shown in Figure 8 The PSSCH DMRS configuration may be done based on, for example, the mobility conditions of the client (affecting the number of DMRSs) and the number of data payloads (affecting the number of selected time slots).

[0101] Figure 9 An example embodiment of the subject matter described herein is shown, illustrating the flow of data in a machine learning model.

[0102] According to an example embodiment, the at least one detection model comprises at least one machine learning (ML) model.

[0103] For example, Figure 9 The example embodiment includes a first ML model 901 for PSSCH DMRS configuration classification and a second ML model 902 for PSCCH DMRS configuration classification.

[0104] According to an example embodiment, the at least one machine learning model includes at least one of: a multi-label multi-class classifier, a neural network, a deep neural network, or a decision forest.

[0105] According to an example embodiment, at least one machine learning model is configured to take as input 903 modulated symbols or in-phase or quadrature-phase IQ samples of a subchannel in the plurality of subchannels 510 and output a probability value 904 for each of a plurality of possible demodulation reference signal configurations active in the subchannel, wherein each of the plurality of possible demodulation reference signal configurations defines an active demodulation reference signal transmission in the subchannel.

[0106] At least one detection model may take as input the modulation symbols or IQ samples of a single subchannel. Thus, at least one detection model may detect the presence of DMRS at each symbol within the subchannel. Client device 200 may use at least one detection model for each subchannel.

[0107] For example, at least one detection model may output a soft value for each DMRS configuration detected as active within time slot 511 .

[0108] The ML classifiers 901, 902 may include multi-label, multi-class classifiers, which may be implemented using various algorithms such as deep neural networks (DNNs), decision forests (DFs), etc. with cross entropy or KL divergence loss and, for example, sigmoid activation functions.

[0109] Figure 10 Another example embodiment of the subject matter described herein is shown, illustrating the flow of data in a machine learning model.

[0110] The received signal samples 1001 may be split into real part 1002 and imaginary part 1003 and fed to at least one ML model into the input layer 1004. Figure 10 In an example embodiment, at least one ML model includes a DNN having N hidden layers 1005. The output layer 1006 of the DNN can produce a K-dimensional output. For example, for a PSSCH DMRS classifier, K=7, otherwise K=3. The k-th output value p k It can represent the probability that at least one signal of configuration k exists in the received sidelink signal.

[0111] One benefit of using a detection model (such as an ML model) instead of, for example, sequential detection is that the detection model can provide a one-shot result, i.e., all configurations can be detected simultaneously. With sufficient training, the classification can be robust to noise—whereas other options may be sensitive to noise.

[0112] At least one ML model or any component thereof (such as classifiers 901, 902) may be pre-trained in a simulated (or replay) environment to establish a baseline. For example, the received signal used for training may be obtained by:

[0113] - Change the number Z of SL clients sending simultaneously (including the case of overlapping resource selection). For example, in the case of S={1, 2, ..., 100}, change Z in the set S.

[0114] - Changing the noise regime. For example, simulating the SNR in the range R, for example with R = {-10, -5, 0, 5} dB.

[0115] - Changing the mobility regime For example, a Doppler dispersion channel with maximum Doppler shift is simulated in the range D, for example with D = {0, 5, 50} Hz.

[0116] Additionally, some example embodiments may include an online training component, which may include at least some of the following:

[0117] - Identification of valid inputs and outputs.

[0118] - Each time a PSCCH DMRS or PSSCH DMRS configuration is correctly detected (ie, PSCCH is successfully detected and decoded, and then the indicated PSSCH DMRS configuration matches the detected PSSCH DMRS configuration), the corresponding input may be stored for training.

[0119] - The samples used for training can only be refreshed once in a while, which corresponds to the frequency with which the characteristics of the environment (congestion, propagation, receiver speed) change.

[0120] - For example, training can follow a reinforcement learning approach or be based on parallel training every x time periods.

[0121] - For example, training can be triggered based on elapsed time or when the number of collected samples exceeds a certain threshold.

[0122] According to an example embodiment, the at least one memory and the computer program code are further configured to, together with the at least one processor, cause the client device 200 to: store inputs and outputs of the at least one machine learning model in response to the at least one decoding path successfully decoding multiple payload transmissions in time slot 511; and retrain the at least one machine learning model using the stored inputs and outputs.

[0123] Figure 11 An example embodiment of the subject matter described herein is shown, illustrating a machine learning model that includes a component trained for low mobility and a component trained for high mobility.

[0124] According to an example embodiment, the at least one machine learning model includes a first machine learning model trained for low mobility and a second machine learning model trained for high mobility.

[0125] Low mobility (lowM) may also be referred to as low Doppler, and high mobility (highM) may also be referred to as high Doppler.

[0126] Client device 200 may include inertial sensors 1101. Based on inertial sensor signals and / or PDSCH detection 1102, client device 200 may perform Doppler estimation 1103. In other example embodiments, client device 200 may perform Doppler estimation based on other available information. If client device 200 detects low mobility, the client device may use a low mobility classifier 1104. Otherwise, client device 200 may use a high mobility classifier 1105.

[0127] One reason to differentiate between low mobility and high mobility is the size of the observation window fed to the ML model. For high-mobility clients, a short observation window is necessary, while for low-mobility clients the opposite is true. Therefore, a first machine learning model trained for low mobility can be trained with a long observation window, and a second machine learning model trained for high mobility can be trained with a short observation window.

[0128] Although at least one detection model is implemented using an ML model in some example embodiments disclosed herein, the at least one detection model may also be implemented in other ways. For example, for each of the possible DMRS sequences in the PSCCH and / or PSSCH, the at least one detection model may include a separate detector, such as a matched filter. When the output of the separate detector is above a preconfigured threshold, the sequence may be considered detected. For example, when the output is assumed to be a soft value between 0 and 1, as an illustrative example, the sequence may be considered present when the corresponding value is above 0.7.

[0129] Figure 12 An example embodiment of the subject matter described herein is shown, illustrating a flow chart of a decode generation process.

[0130] use Figure 12 The client device 200 may detect PSCCH and PSSCH transmissions on consecutive subchannels based on correlation of the PSSCH DMRS configurations detected in each adjacent subchannel.

[0131] After obtaining information on which PSCCH DMRS configurations and PSSCH DMRS configurations are active in the subchannel 510 of the time slot 511, the client device 200 may apply Figure 12 The process described in

[15] is used to identify individual transmissions. This process can be applied to a subset of the most likely detected configurations, such as p i Configurations above a certain threshold, where p irepresents the probability that at least one signal of configuration k exists in the received signal.

[0132] According to an example embodiment, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device 200 to create at least one decoding path for payload transmission in such a manner that the at least one decoding path maximizes the number of decoded transmissions.

[0133] In operation 1201 , the client device 200 may transition to a first sub-channel.

[0134] In operation 1202 , the client device 200 may detect PSCCH DMRS(s) in a current subchannel.

[0135] In operation 1203 , the client device 200 may store the PSCCH DMRS configuration detected for each subchannel.

[0136] In operation 1204 , the client device 200 may detect a PSCCH DMRS in a current subchannel.

[0137] In operation 1205 , the client device 200 may store the detected PSSCH DMRS configuration for each subchannel.

[0138] In operation 1206, the client device 200 may check whether the current subchannel is the first subchannel. If the current subchannel is the first subchannel, the client device 200 may perform operation 1207 and transition to the next subchannel and transition back to operation 1202. Otherwise, the client device 200 may transition to operation 1208.

[0139] In operation 1208, the client device 200 may correlate the detected PSSCH DMRS configuration with a detected PSSCH DMRS configuration from a previous subchannel using the stored PSSCH DMRS.

[0140] According to an example embodiment, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device 200 to: identify a payload transmission in a time slot 511 based on a plurality of potentially active demodulation reference signal configurations identified in the time slot 511 by performing correlation between symbols across subchannels in the plurality of subchannels 510 so as to identify payload transmission occurring across subchannels in the plurality of subchannels 510.

[0141] In operation 1209 , the client device 200 may store the identified transmission.

[0142] In operation 1210, client device 200 may check whether the current subchannel is the last subchannel in time slot 511. If the current subchannel is the last subchannel in time slot 511, client device 200 may transition to the next time slot and transition to operation 1201. Otherwise, the client device may transition to operation 1207.

[0143] The client device 200 may be configured to create at least one decoding path for the plurality of payload transmissions in the time slot 511 further based on the amplitude / power of the DMRS.

[0144] According to an example embodiment, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device 200 to create at least one decoding path for the payload transmission by: calculating a reference signal received power, RSRP, for each of a plurality of potentially active demodulation reference signal configurations; arranging the plurality of potentially active demodulation reference signal configurations according to the calculated RSRP of each potentially active demodulation reference signal configuration; and creating the at least one decoding path based on the arrangement of the plurality of potentially active demodulation reference signal configurations.

[0145] According to an example embodiment, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device 200 to: create at least one decoding path based on the permutation by excluding from the at least one decoding path potentially active demodulation reference signal configurations having an RSRP below a preconfigured threshold RSRP.

[0146] According to an example embodiment, the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device 200 to decode the payload transmission according to the created at least one decoding path.

[0147] The client device 200 may create at least one decoding path using, for example, the following operations:

[0148] The RSRP associated with each potentially active DMRS configuration detected for both PSSCH and PSCCH is calculated.

[0149] The different (potential) transmissions are ranked based on PSCCH and PSSCH RSRP.

[0150] In order to successfully decode the transmission, both PSCCH and PSSCH RSRP are required to be strong, ie above a predefined threshold RSRP_high.

[0151] All transmissions with an associated RSRP below a configured threshold RSRP_low (eg associated with receiver sensitivity) are excluded from the decoding path.

[0152] Figure 13 An example embodiment of the subject matter described herein is shown, illustrating a method.

[0153] According to an example embodiment, method 1300 includes receiving 1301 at least one sidelink signal in a time slot on a plurality of subchannels.

[0154] Method 1300 may also include identifying 1302 a plurality of potentially active demodulation reference signal configurations among a plurality of possible demodulation reference signal configurations in a time slot 511 across a plurality of subchannels based on at least one side link signal using at least one detection model, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot 511.

[0155] The method 1300 may further include identifying 1300 a plurality of payload transmissions in the time slot 511 based on the identified plurality of potentially active demodulation reference signal configurations in the time slot 511 .

[0156] The method 1300 may further include creating 1304 at least one decoding path for the plurality of payload transmissions in the time slot 511 based on the identified plurality of payload transmissions in the time slot 511 and the identified plurality of potentially active demodulation reference signal configurations.

[0157] It is to be understood that the order in which operations 1301 to 1304 are performed may vary. Figure 13 The example embodiments depicted in FIG.

[0158] Method 1300 can be performed by Figure 2 The method 600 is performed by the client device 200. Other features of the method 600 result directly from the functionality and parameters of the client device 200. The method 600 may be performed at least in part by a computer program(s).

[0159] An apparatus may include means for performing any aspect of the method(s) described herein. According to an example embodiment, the means include at least one processor and a memory including program code, the at least one processor and the program code being configured to, when executed by the at least one processor, cause performance of any aspect of the method.

[0160] The functionality described herein may be performed, at least in part, by one or more computer program product components (such as software components). According to an example embodiment, the client device 200 includes a processor that, when executed, is configured by program code to perform the described operations and example embodiments of the functionality. Alternatively, or in addition, the functionality described herein may be performed, at least in part, by one or more hardware logic components. For example, but not limited to, illustrative types of hardware logic components that may be used include field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and graphics processing units (GPUs).

[0161] Any range or device value given herein may be expanded or altered without losing the effect sought. Furthermore, any exemplary embodiment may be combined with another exemplary embodiment unless expressly prohibited.

[0162] Although the subject matter has been described in language specific to structural features and / or acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as examples of implementing the claims, and other equivalent features and acts are intended to fall within the scope of the claims.

[0163] It is to be understood that the benefits and advantages described above may relate to one example embodiment or may relate to several example embodiments. The example embodiments are not limited to those that solve any or all of the problems described or have any or all of the benefits and advantages described. It is also to be understood that reference to 'an' item refers to one or more of these items.

[0164] The steps of the methods described herein can be performed in any suitable order, or simultaneously where appropriate. Additionally, individual blocks can be deleted from any method without departing from the spirit and scope of the subject matter described herein. Aspects of any of the above-described example embodiments can be combined with aspects of any other example embodiment described to form further example embodiments without losing the desired effect.

[0165] The term 'comprising' is used herein to mean including identified method blocks or elements, but that such blocks or elements do not comprise an exclusive list and the method or apparatus may contain additional blocks or elements.

[0166] It is to be understood that the above description is given by way of example only and that various modifications may be made by those skilled in the art. The above description, examples, and data provide a complete description of the structure and use of the exemplary embodiments. Although various exemplary embodiments have been described above with a certain degree of particularity or with reference to one or more individual exemplary embodiments, those skilled in the art may make many changes to the disclosed exemplary embodiments without departing from the spirit or scope of this specification.

Claims

1. A client device (200) for communication, comprising: at least one processor (202); as well as at least one memory (204), the at least one memory comprising computer program code; The at least one memory and the computer program code are configured to, with the at least one processor, cause the client device to: receiving at least one sidelink signal in a time slot (511) on a plurality of subchannels (510); identifying, using at least one detection model, a plurality of potentially active demodulation reference signal configurations among a plurality of possible demodulation reference signal configurations (700, 800) in the time slot across the plurality of subchannels based on the at least one sidelink signal, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot; identifying a plurality of payload transmissions in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot (501, 502, 503, 504); as well as At least one decoding path is created for the plurality of payload transmissions in the time slot based on the identified plurality of payload transmissions in the time slot and the identified plurality of potentially active demodulation reference signal configurations.

2. The client device (200) of claim 1, wherein the demodulation reference signal transmission in the time slot comprises at least a demodulation reference signal transmission for a control channel and a demodulation reference signal transmission for a data channel, and wherein the plurality of payload transmissions comprises at least the control channel and the data channel.

3. The client device (200) according to claim 1 or 2, wherein the plurality of payload transmissions comprises physical sidelink control channel (PSCCH) transmissions and / or physical sidelink shared channel (PSSCH) transmissions, and the demodulation reference signal transmission comprises a PSCCH demodulation reference signal (DMRS) transmission and / or a PSSCH DMRS transmission.

4. The client device (200) of claim 1, wherein the at least one detection model comprises at least one machine learning model (901, 902).

5. The client device (200) of claim 4, wherein the at least one machine learning model comprises at least one of: a multi-label multi-class classifier, a neural network, a deep neural network, or a decision forest.

6. The client device (200) of claim 4, wherein the at least one machine learning model is configured to take as input (903) modulated symbols or in-phase, quadrature-phase (IQ) samples of a subchannel of the plurality of subchannels and output a probability value (904) for each of the plurality of possible demodulation reference signal configurations active in the subchannel, wherein each of the plurality of possible demodulation reference signal configurations defines an active demodulation reference signal transmission in the subchannel.

7. The client device (200) of claim 1 , wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to: identify the payload transmission in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot by correlating between symbols across subchannels of the plurality of subchannels so as to identify payload transmission occurring across subchannels of the plurality of subchannels.

8. The client device (200) of claim 4, wherein the at least one machine learning model comprises a first machine learning model (1104) trained for low mobility and a second machine learning model (1105) trained for high mobility.

9. The client device (200) of claim 4, wherein the at least one memory and the computer program code are further configured to, together with the at least one processor, cause the client device to create the at least one decoding path for the payload transmission in such a manner that the at least one decoding path maximizes the number of decoded transmissions.

10. The client device (200) of claim 4, wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to create the at least one decoding path for the payload transmission by: calculating a reference signal received power (RSRP) for each demodulation reference signal configuration of the plurality of potentially active demodulation reference signal configurations; arranging the plurality of potentially active demodulation reference signal configurations according to the calculated RSRP of each potentially active demodulation reference signal configuration; as well as The at least one decoding path is created based on the permutation of the plurality of potentially active demodulation reference signal configurations.

11. The client device (200) of claim 10, wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to: create at least one decoding path based on the permutation by excluding from the at least one decoding path potentially active demodulation reference signal configurations having an RSRP below a preconfigured threshold RSRP.

12. The client device (200) of claim 10, wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the client device to decode the payload transmission according to the at least one created decoding path.

13. The client device (200) according to any one of claims 4 to 12, wherein the at least one memory and the computer program code are further configured to, together with the at least one processor, cause the client device to: responsive to the at least one decoding path successfully decoding the plurality of payload transmissions in the time slot, storing inputs and outputs of the at least one machine learning model; and Retrain the at least one machine learning model using the stored inputs and the stored outputs.

14. A method of communication (1300), comprising: receiving (1301) at least one sidelink signal in a time slot on a plurality of subchannels; identifying (1302) a plurality of potentially active demodulation reference signal configurations among a plurality of possible demodulation reference signal configurations in the time slot across the plurality of subchannels based on the at least one sidelink signal using at least one detection model, wherein each demodulation reference signal configuration defines a demodulation reference signal transmission in the time slot; identifying (1303) a plurality of payload transmissions in the time slot based on the identified plurality of potentially active demodulation reference signal configurations in the time slot; as well as At least one decoding path is created (1304) for the plurality of payload transmissions in the time slot based on the identified plurality of payload transmissions in the time slot and the identified plurality of potentially active demodulation reference signal configurations.

15. A computer program product comprising a program code configured to perform the method according to claim 14 when the computer program product is executed on a computer.

Images