# US 2026/0150086 A1 · Method, apparatus, and system for semantic communications

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0150086 A1
- Document ID: US20260150086A1
- Status: Published application
- Kind: A1
- Application number: 19/427,058
- Filed: 2025-12-19
- Published: 2026-05-28
- Earliest claimed: 2023-06-21
- Assignees: Huawei Technologies Co., Ltd.
- Applicants: Huawei Technologies Co., Ltd.
- Inventors: Mengyao Ma; Yiqun Ge; Jianglei Ma; Qifan Zhang
- Web page: https://patentgrep.com/patent/US20260150086A1

## CPC

- H04W 68/02: Arrangements for increasing efficiency of notification or paging channel
- H03M 13/6312: {Error control coding in combination with data compression}
- H04W 76/14: Direct-mode setup
- H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72)
- H03M: CODING; DECODING; CODE CONVERSION IN GENERAL (using fluidic means F15C4/00; optical analogue/digital converters G02F7/00; coding, decoding or code conversion, specially adapted for particular applications, see the relevant subclasses, e.g. G01D, G01R, G06F, G06T, G09G, G10L, G11B, G11C, H04B, H04L, H04M, H04N; ciphering or deciphering for cryptography or other purposes involving the need for secrecy G09C)

## Abstract

Provided are a sensing communication method, apparatus, and system. An apparatus such as a central device can send tokenization configuration and broadcast or multi-cast or unicast query message(s), so that other apparatus(es) such as one or more sensing devices can obtain the query message(s) and respond with sensing result(s) in response to the obtained query message(s). The sensing result(s) may include at least one piece of sensed data and/or at least one sensing semantic corresponding to sensing token(s) that is determined from at least one sensing token and matches query message(s), where the at least one sensing token is tokenized from at least one sensing semantic based on the tokenization configuration.

## Claims

### Claim 1

A method, comprising:
obtaining a tokenization configuration;
obtaining at least one query message;
tokenizing at least one sensing semantic to at least one sensing token based on the tokenization configuration, wherein each sensing semantic of the at least one sensing semantic is tokenized to a corresponding sensing token of the at least one sensing token;
determining, from the at least one sensing token, one or more sensing tokens matching one or more query messages of the at least one query message; and
sending a sensing result, wherein the sensing result indicates at least one of:
at least one piece of sensed data or one or more sensing semantics corresponding to the one or more sensing tokens.

### Claim 2 (depends on 1)

The method according to claim 1, wherein the tokenization configuration indicates at least one of:
at least one tokenization model;
at least one function;
at least one projection matrix;
at least one graph-based or topology-based pruning; or
at least one compression approach.

### Claim 3 (depends on 2)

The method according to claim 2, wherein at least one of a set of tokenization models, a set of functions, a set of projection matrixes, a set of graph-based or topology-based prunings, or a set of compression approaches is configured for at least one of multiple tasks, multiple modalities, or multiple combinations of a task and a modality.

### Claim 4 (depends on 3)

The method according to claim 3, wherein each tokenization model of the set of tokenization models, each function of the set of functions, each projection matrix of the set of projection matrixes, each graph-based or topology-based pruning of the set of graph-based or topology-based prunings, or each compression approach of the set of compression approaches corresponds to at least one respective identifier indicating a respective task identifier, a respective modality identifier, or both the respective task identifier and the respective modality identifier.

### Claim 5 (depends on 2)

The method according to claim 2, wherein the tokenization configuration indicates the at least one tokenization model, and
the tokenizing the at least one sensing semantic to the at least one sensing token based on the tokenization configuration comprises:
tokenizing each sensing semantic of the one or more sensing semantics to the corresponding sensing token based on a corresponding tokenization model of the at least one tokenization model.

### Claim 6 (depends on 5)

The method according to claim 5, wherein each tokenization model of the at least one tokenization model is a corresponding deep learning model of at least one deep learning model, and
the tokenizing the at least one sensing semantic to the at least one sensing token based on the at least one tokenization model comprises:
inputting each sensing semantic of the one or more sensing semantics into the corresponding deep learning model of the at least one deep learning model to obtain the corresponding sensing token.

### Claim 7

An apparatus comprising:
at least one processor coupled with a memory storing instructions, wherein when the instructions executed by the at least one processor, the apparatus is caused to perform:
obtaining a tokenization configuration; and obtain at least one query message;
tokenizing at least one sensing semantic to at least one sensing token based on the tokenization configuration, wherein each sensing semantic of the at least one sensing semantic is tokenized to a corresponding sensing token of the at least one sensing token;
determining, from the at least one sensing token, one or more sensing tokens matching one or more query messages of the at least one query message; and
sending a sensing result, wherein the sensing result indicates at least one of:
at least one piece of sensed data or one or more sensing semantics corresponding to the one or more sensing tokens.

### Claim 8 (depends on 7)

The apparatus according to claim 7, wherein the tokenization configuration indicates at least one of:
at least one tokenization model;
at least one function;
at least one projection matrix;
at least one graph-based or topology-based pruning; or
at least one compression approach.

### Claim 9 (depends on 8)

The apparatus according to claim 8, wherein at least one of a set of tokenization models, a set of functions, a set of projection matrixes, a set of graph-based or topology-based pruning, or a set of compression approaches is configured for at least one of multiple tasks, multiple modalities, or multiple combinations of a task and a modality.

### Claim 10 (depends on 9)

The apparatus according to claim 9, wherein each tokenization model of the set of tokenization models, each function of the set of functions, each projection matrix of the set of projection matrixes, each graph-based or topology-based pruning of the set of graph-based or topology-based prunings, or each compression approach of the set of compression approaches corresponds to at least one respective identifier indicating a respective task identifier, a respective modality identifier, or both the respective task identifier and the respective modality identifier.

### Claim 11 (depends on 8)

The apparatus according to claim 8, wherein the tokenization configuration indicates the at least one tokenization model, and
the tokenizing the at least one sensing semantic to the at least one sensing token based on the tokenization configuration comprises:
tokenizing each sensing semantic of the one or more sensing semantics to the corresponding sensing token based on a corresponding tokenization model of the at least one tokenization model.

### Claim 12 (depends on 11)

The apparatus according to claim 11, wherein each tokenization model of the at least one tokenization model is a corresponding deep learning model of at least one deep learning model, and
the tokenizing the at least one sensing semantic to the at least one sensing token based on the at least one tokenization model comprises:
inputting each sensing semantic of the one or more sensing semantics into the corresponding deep learning model of the at least one deep learning model to obtain the corresponding sensing token.

### Claim 13 (depends on 8)

The apparatus according to claim 8, wherein the tokenization configuration indicates the at least one function, and
the tokenizing the at least one sensing semantic to the at least one sensing token based on the at least one tokenization model comprises:
applying, with each sensing semantic of the one or more sensing semantics as a function parameter, a corresponding function of the at least one function to obtain the corresponding sensing token.

### Claim 14

A apparatus, comprising:
at least one processor coupled with a memory storing instructions, wherein when the instructions executed by the at least one processor, the apparatus is caused to perform:
sending a tokenization configuration;
sending at least one query message; and
obtaining one or more sensing results, wherein each sensing result of the one or more sensing results comprises at least one of at least one piece of sensed data or at least one sensing semantic corresponding to at least one sensing token that is tokenized from at least one sensing semantic based on the tokenization configuration and matches one or more query messages of the at least one query message, wherein each sensing semantic of the at least one sensing semantic is tokenized to a corresponding sensing token.

### Claim 15 (depends on 14)

The apparatus according to claim 14, wherein the tokenization configuration indicates at least one of:
at least one tokenization model;
at least one function;
at least one projection matrix;
at least one graph-based or topology-based pruning; or
at least one compression approach.

### Claim 16 (depends on 15)

The apparatus according to claim 15, wherein at least one of a set of tokenization models, a set of functions, a set of projection matrixes, a set of graph-based or topology-based prunings, or a set of compression approaches is configured for at least one of multiple tasks, multiple modalities, or multiple combinations of a task and a modality.

### Claim 17 (depends on 16)

The apparatus according to claim 16, wherein each tokenization model of the set of tokenization models, each function of the set of functions, each projection matrix of the set of projection matrixes, each graph-based or topology-based pruning of the set of graph-based or topology-based prunings, or each compression approach of the set of compression approaches corresponds to at least one respective identifier indicating a respective task identifier, a respective modality identifier, or both the respective task identifier and the respective modality identifier.

### Claim 18 (depends on 15)

The apparatus according to claim 15, wherein the tokenization configuration indicates the at least one tokenization model, and each sensing token of one or more sensing tokens is tokenized from a corresponding sensing semantic based on a corresponding tokenization model of the at least one tokenization model.

### Claim 19 (depends on 18)

The apparatus according to claim 18, wherein each tokenization model of the at least one tokenization model is a corresponding deep learning model of at least one deep learning model, and each sensing token of the one or more sensing tokens is tokenized from the corresponding sensing semantic by inputting the corresponding sensing semantic into the corresponding deep learning model of the at least one deep learning model to obtain each sensing token of the at least one sensing token.

### Claim 20 (depends on 15)

The apparatus according to claim 15, wherein the tokenization configuration indicates the at least one function, and each sensing token of one or more sensing tokens is tokenized from a corresponding sensing semantic by applying a corresponding function of the at least one function with the corresponding sensing semantic as a function parameter to obtain each sensing token of the at least one sensing token.
