Dynamic and static coding olfactory recognition platform and gas sensing method thereof

Abstract

The application discloses a dynamic and static coding olfactory recognition platform for embodied intelligence and a gas sensing method thereof, and the platform relies on a multi-element sensitive material to construct an electronic nose mode of a sensor array, and proposes active physical coding measurement of gas-phase molecules based on broken transport symmetry. The core comprises a programmable air flow dynamic coding controller, a dynamic and static coding electronic nose and a data acquisition and processing unit. By combining and configuring the dynamic and static coding of the air outlet, the symmetry of gas transport is actively broken, and a characteristic air flow path is shaped in the chamber, so that a dynamic response mode corresponding to the gas type is excited on a single sensitive material sensor array. The application deeply fuses physical flow field regulation and chemical sensing, realizes active physical coding measurement of gas-phase molecules, and can be used as the fifth mode of the embodied intelligent system, namely, an olfactory interface, cooperates with modes such as vision, hearing, touch and proprioceptive perception, and realizes multi-modal environmental perception and information fusion.

Description

Technical Field

[0001] This invention relates to the field of gas sensing and olfactory computing technology, specifically to a dynamic and static coding olfactory recognition platform for embodied intelligence and its gas sensing method, and in particular to an olfactory sensing system based on the active physical dynamic coding measurement principle of gas phase molecules with broken transport symmetry, which enables a single sensitive material to identify multiple volatile organic compounds. Background Technology

[0002] Embodied intelligence emphasizes that intelligent agents acquire perception and cognition capabilities through physical interaction with their environment, with its core being the deep integration of multimodal perception and real-time decision-making. Currently, visual, auditory, tactile, and even proprioceptive technologies have been gradually applied in intelligent agents such as robots, while the integration of olfaction—an important modality for biological organisms to perceive environmental chemical information—in embodied intelligence systems is still in its early stages. Traditional electronic nose technology often relies on constructing sensor arrays composed of various gas-sensitive materials, utilizing the differences in material chemical selectivity to generate cross-responses to different gases, forming a characteristic fingerprint spectrum. However, this strategy, which relies on material diversity, faces problems such as high cost, complex manufacturing processes, poor long-term stability, and difficult calibration, making it difficult to meet the requirements of embodied intelligence for sensor miniaturization, low power consumption, high reliability, and real-time performance.

[0003] In recent years, some studies have attempted to enhance sensor selectivity through physical structure control, such as designing flow channels with different geometries or using temperature modulation techniques. However, a systematic theory and methodology have not yet been formed, and most studies are limited to passive control and lack the ability to actively programmable dynamic coding. Summary of the Invention

[0004] This invention introduces the concept of transport symmetry breaking into the field of gas sensing for the first time, revealing the intrinsic relationship between the transport behavior of gas molecules in an asymmetric flow field and the dynamic response of the sensor. Based on this, a new principle of active physical dynamic coding measurement of gas phase molecules is proposed, and a dynamic and static coding olfactory recognition platform for embodied intelligence is constructed, providing a brand-new "fifth mode" - olfactory interface - for intelligent agents.

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a dynamic and static coded olfactory recognition platform and its gas sensing method for embodied intelligence. By actively controlling the symmetry of the airflow transport path, it can achieve diverse dynamic responses on a single sensitive material, thereby achieving accurate identification of multiple gases in a low-cost and highly stable manner.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A dynamic coding olfactory recognition platform for embodied intelligence, comprising:

[0008] A programmable airflow dynamic coding controller is used to generate airflow dynamic coding instructions according to preset parameters;

[0009] The dynamic coding electronic nose is connected to the programmable airflow dynamic coding controller. It has a gas flow path and a single sensitive material gas sensor array inside, and several air outlets on the cavity. The air outlets are selectively opened or closed according to the airflow dynamic coding command to introduce a transport symmetry disruption in the gas path, so that the airflow flowing through the single sensitive material gas sensor array forms different airflow paths.

[0010] The data acquisition and processing unit is connected to the gas sensor array of the dynamically coded electronic nose and is used to acquire the dynamic response signals generated by the gas sensor array under different airflow paths, and to identify the gas type or concentration based on the dynamic response signals.

[0011] A statically encoded olfactory recognition platform for embodied intelligence, comprising:

[0012] The statically coded electronic nose has a cavity containing a parallel gas sensor array of a single sensitive material. Several air outlets are located at different positions on the cavity. Due to the inconsistent positions of the air outlets, the transport symmetry is broken, causing the airflow through the parallel gas sensor array of the single sensitive material to form different airflow paths.

[0013] The data acquisition and processing unit is connected to the single-sensitive material parallel gas sensor array of the statically coded electronic nose, and is used to acquire the dynamic response signals generated by the sensor array under different airflow paths, and identify the gas type or concentration based on the dynamic response signals.

[0014] Furthermore, the dynamically coded electronic nose includes a top cavity, a rubber sealing ring, a gas sensor array, a bottom cover, a flexible interface, and fasteners. The top cavity is integrally formed from 316L stainless steel powder using selective laser melting technology. One side integrates a groove as a gas flow path, and the other side integrates several hollow connecting tubes as air inlets and outlets. The four corners of the cavity have through-holes with first screw holes. The gas sensor array is a sensor array based on a flexible printed circuit board, and the sensitive material coated on it is a single sensitive material. The bottom cover is a flat plate matching the size of the top cavity, with first screw holes at corresponding positions. The rubber sealing ring is a rectangular rubber film matching the size of the top cavity, with an opening of the same shape as the gas flow path and a through hole corresponding to the first screw hole laser-cut on it. The top cavity, rubber sealing ring, gas sensor array, and bottom cover are stacked sequentially from top to bottom and fixed with fasteners.

[0015] Furthermore, the programmable airflow dynamic coding controller realizes independent on / off control of the air outlet by controlling the valve array, and the valve array is connected one-to-one with several air outlets of the dynamic coding electronic nose.

[0016] Furthermore, the data acquisition and processing unit includes a multi-channel data acquisition module, a feature extraction module, and a pattern recognition module; the multi-channel data acquisition module is used to acquire dynamic response signals; the feature extraction module is used to extract time-domain, frequency-domain, or time-frequency-domain features related to gas types from the dynamic response signals; the pattern recognition module performs classification or regression analysis on the extracted features based on machine learning algorithms to achieve gas type identification or concentration prediction.

[0017] Furthermore, the statically coded electronic nose includes: a parallel top cavity, a parallel rubber sealing ring, a parallel gas sensor array of a single sensitive material, a parallel bottom cover, and a parallel flexible interface;

[0018] The parallel top cavity is integrally formed from 316L stainless steel powder using 3D printing technology; one side of its main body is provided with an air inlet connecting pipe and several independent air outlet connecting pipes, and the other side is integrated with a groove as a parallel gas flow path.

[0019] The parallel bottom cover is a flat plate that matches the size of the parallel top cavity; the parallel rubber sealing ring is a rectangular rubber film that matches the size of the parallel top cavity, with through holes laser-cut out in the same shape as the parallel gas flow path and several second screw holes to ensure the airtightness between the gas sensor array and the parallel top cavity.

[0020] The parallel top cavity, parallel rubber sealing ring, single sensitive material parallel gas sensor array, and parallel bottom cover are stacked sequentially from top to bottom, with their several second screw holes aligned one by one; several screws are inserted from the second screw holes of the top cavity, and then nuts are tightened on one side of the parallel bottom cover to form a statically coded electronic nose.

[0021] A gas sensing method based on the aforementioned dynamic coding olfactory recognition platform includes the following steps:

[0022] S1: The programmable airflow dynamic coding controller generates airflow dynamic coding instructions to selectively open or close multiple air outlets of the dynamic coding electronic nose, introducing a transport symmetry disruption in the gas flow path.

[0023] S2: The gas to be tested is introduced into the dynamic coded electronic nose. The gas flows through a single sensitive material gas sensor array under the action of the airflow path, which excites a dynamic response signal.

[0024] S3: The dynamic response signal of the gas sensor array is acquired in real time through the data acquisition and processing unit and preprocessed.

[0025] S4: Extract feature parameters from the preprocessed dynamic response signal to construct a gas fingerprint feature vector;

[0026] S5: Input the gas fingerprint feature vector into the pattern recognition module and output gas type or concentration information.

[0027] A gas sensing method based on the static coding olfactory recognition platform includes the following steps:

[0028] S1: The gas to be measured is introduced into the statically encoded electronic nose. The gas flows through a parallel gas sensor array with a single sensitive material under the action of the airflow path, which excites a dynamic response signal.

[0029] S2: The data acquisition and processing unit acquires the dynamic response signal of the parallel gas sensor array of a single sensitive material in real time and performs preprocessing.

[0030] S3: Extract feature parameters from the preprocessed dynamic response signal to construct a gas fingerprint feature vector;

[0031] S4: Input the gas fingerprint feature vector into the pattern recognition module and output gas type or concentration information.

[0032] An application of the aforementioned dynamic and static coding olfactory recognition platform in the identification of various volatile organic compounds, which is used to distinguish different types of volatile organic compounds or detect their concentration.

[0033] The application integrates the platform into robots, drones, wearable devices, or IoT terminals as an olfactory sensing module for embodied intelligence.

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0035] 1) Theoretical innovation: For the first time, the breaking of transport symmetry was introduced into the field of gas sensing, revealing the intrinsic relationship between the transport behavior of gas molecules in an asymmetric flow field and the dynamic response of the sensor, providing a brand-new physical mechanism for gas identification;

[0036] 2) Breakthrough in principle: A new principle of active physical dynamic coding measurement of gas phase molecules was proposed. By actively controlling the gas flow transport path, the chemical information of the gas is dynamically encoded into a dynamic response signal, realizing a paradigm shift from "passive response" to "active dynamic coding".

[0037] 3) Simplified hardware: The sensor array requires only a single sensitive material, which greatly reduces the complexity of material preparation and batch variation, and improves system stability and consistency;

[0038] 4) Programmability: The introduction of a programmable airflow dynamic coding controller can dynamically adjust the dynamic coding strategy according to the identification task, realizing flexible and adaptive gas sensing;

[0039] 5) Cutting-edge applications: Integrating olfaction as the fifth modality into embodied intelligence systems fills the gap in robot olfactory perception and provides a new dimension for multimodal environmental perception and intelligent decision-making. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the dynamic coding olfactory recognition platform of the present invention;

[0041] Figure 2 This is an exploded view of the dynamically coded electronic nose of the present invention;

[0042] Figure 3 This is a schematic diagram of the combination of the dynamic coded electronic nose of the present invention;

[0043] Figure 4 This is a schematic diagram of the top cavity of the present invention;

[0044] Figure 5 This is a schematic diagram of the gas sensing array of the present invention;

[0045] Figure 6 This is a schematic diagram of the static coding olfactory recognition platform of the present invention;

[0046] Figure 7 This is an exploded view of the statically encoded electronic nose of the present invention;

[0047] Figure 8 This is a schematic diagram of the statically coded electronic nose of the present invention;

[0048] Figure 9 This is a schematic diagram of the parallel top cavity of the present invention;

[0049] Figure 10 This is a schematic diagram of the single-sensitive material parallel gas sensor array of the present invention;

[0050] Figure 11 This is a linear discriminant analysis graph of the dynamically coded electronic nose of the present invention for various volatile organic compounds;

[0051] Figure 12 This is a normalized confusion matrix diagram of the validation set of wheat samples for the statically encoded electronic nose of the present invention.

[0052] Figure 13 This is a schematic diagram illustrating an application scenario of the embodied intelligent olfactory interface of the present invention;

[0053] In the diagram: 101 - Programmable airflow dynamic coding controller; 103 - Data acquisition and processing unit; 100 - Valve array;

[0054] 102-Dynamic Coding Electronic Nose: 1-Screw; 2-Top Cavity; 3-Rubber Sealing Ring; 4-Gas Sensor Array; 5-Bottom Cover; 6-Nut; 7-Gas Flow Path; 8-Hollow Connecting Tube; 9-First Screw Hole; 23-Flexible Interface;

[0055] 104-Static Encoded Electronic Nose: 10-Parallel Top Cavity; 11-Parallel Rubber Sealing Ring; 12-Single Sensitive Material Parallel Gas Sensor Array; 13-Parallel Bottom Cover; 14-Inlet Connecting Pipe; 15-First Outlet Connecting Pipe; 16-Second Outlet Connecting Pipe; 17-Third Outlet Connecting Pipe; 18-Fourth Outlet Connecting Pipe; 19-Fifth Outlet Connecting Pipe; 20-Second Screw Hole; 21-Circular Microgroove; 22-Parallel Flexible Interface; 24-Parallel Gas Flow Path. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be described in complete, clear, and detailed below with reference to the accompanying drawings and specific embodiments. The described embodiments are some, but not all, of the embodiments of the present invention. Based on the described embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0057] See Figure 1 The present invention provides a dynamic coding olfactory recognition platform for embodied intelligence, comprising:

[0058] The programmable airflow dynamic coding controller 101 is used to generate airflow dynamic coding instructions according to a preset;

[0059] The dynamic coding electronic nose 102 is connected to the programmable airflow dynamic coding controller 101. It has a gas flow path and a single sensitive material gas sensor array 4 inside, and several air outlets are provided on the cavity. The air outlets are selectively opened or closed according to the airflow dynamic coding command to introduce a transport symmetry disruption in the gas path, so that the airflow flowing through the single sensitive material gas sensor array forms different airflow paths.

[0060] The data acquisition and processing unit 103 is connected to the gas sensor array of the dynamic coded electronic nose 102, and is used to acquire the dynamic response signals generated by the gas sensor array under different airflow paths, and identify the gas type or concentration based on the dynamic response signals.

[0061] The programmable airflow dynamic coding controller 101 achieves independent on / off control of the air outlet by controlling the valve array 100, and the valve array 100 is connected one-to-one with several air outlets of the dynamic coding electronic nose 102.

[0062] See Figure 6 The present invention provides a statically encoded olfactory recognition platform for embodied intelligence, comprising:

[0063] The statically encoded electronic nose 104 has a single-sensitive material parallel gas sensor array 12 inside its cavity. Several air outlets are provided at different positions on the cavity. Due to the inconsistency of the positions of the air outlets, the transport symmetry is broken, causing the airflow flowing through the single-sensitive material parallel gas sensor array to form different airflow paths.

[0064] The data acquisition and processing unit 103 is connected to the single-sensitive material parallel gas sensor array 12 of the statically encoded electronic nose, and is used to acquire the dynamic response signals generated by the sensor array under different airflow paths, and identify the gas type or concentration based on the dynamic response signals.

[0065] See Figures 2-5 The dynamic coding electronic nose 102 of the present invention includes: screw 1, nut 6, top cavity 2, rubber sealing ring 3, gas sensor array 4, bottom cover 5, and 10-channel flexible interface 23.

[0066] The top cavity 2 is integrally formed from 316L stainless steel powder using selective laser melting technology. Its main dimensions are 33.35 mm × 18.4 mm × 2 mm. One side integrates a 1 mm deep groove as a gas flow path 7, and the other side integrates six hollow connecting pipes 8 with an inner diameter of 1 mm and an outer diameter of 3 mm as air inlets / outlets. Four corners have 3 mm diameter through-hole screw holes 9. The hollow connecting pipes 8 on the outlet side can be selectively blocked by valves or a programmable controller to achieve different spatial airflow dynamic coding configurations (i.e., "fingering"), thereby introducing transport symmetry breaking and realizing active physical dynamic coding measurement of gas phase molecules.

[0067] The rubber sealing ring 3 is a rectangular rubber film, with a portion of the same shape as the gas flow path 7 and four through holes corresponding to the first screw hole 9 removed by laser cutting.

[0068] Stack the top cavity 2, rubber sealing ring 3, gas sensor array 4, and bottom cover 5 sequentially from top to bottom, aligning the four first screw holes 9 one-to-one. Insert the four screws 1 into the first screw holes 9 of the top cavity 2, and then tighten the nuts 6 on one side of the bottom cover 5 to secure it. This completes the assembly of the dynamic coded electronic nose. Figure 2 , Figure 3As shown.

[0069] Example 1: Assembly of a dynamically coded electronic nose, see [link / reference] Figures 2-5 Align and stack the top cavity 2, rubber sealing ring 3, gas sensor array 4, and bottom cover 5 in sequence. Pass the four M3 screws 1 through the first screw hole 9 and tighten them with M3 nuts 6 to complete the assembly.

[0070] Example 2: Application of dynamically coded electronic nose in the identification of various volatile organic compounds, see [reference]. Figure 11 2-1: First, connect the air inlet pipe of the top cavity 2 to the gas path. Multiple connecting pipes on the outlet side are selectively blocked by the programmable airflow dynamic coding controller 101 according to a preset dynamic coding scheme. Rinse the cavity with high-purity nitrogen to stabilize the baseline of the gas sensor array 4.

[0071] 2-2: The gas sensor array 4 was exposed to various prepared headspace volatile gases containing different volatile organic compounds. During sample testing, different dynamic encoding configurations of the gas outlets were switched sequentially, and the dynamic response signals of the single-sensitive material gas sensor array 4 under different airflow modes were recorded. The resistance changes of each sensor were recorded using a multi-channel data acquisition device. The collected sensor response signals were then standardized.

[0072] The data generated linear discriminant analysis plots for various volatile organic compounds, such as... Figure 11 As shown.

[0073] Example 3: Assembly and application of a statically coded electronic nose based on an asymmetric air outlet layout

[0074] See Figures 7-10 Its structure adopts a "parallel" design, including: a parallel top cavity 10, a parallel rubber sealing ring 11, a single sensitive material parallel gas sensor array 12, a parallel bottom cover 13, and a 20-channel parallel flexible interface 22.

[0075] The parallel top cavity 10 is integrally formed from 316L stainless steel powder using 3D printing technology. One side of its main body has an air inlet connecting pipe 14 and five independent air outlet connecting pipes 15, 16, 17, 18, and 19. These five air outlets are spatially asymmetrically arranged diagonally: the first air outlet connecting pipe 15 is located directly above the rightmost electrode of the first row of electrodes; the second air outlet connecting pipe 16 is located directly above the rightmost electrode of the second row; the third air outlet connecting pipe 17 is located directly above the third electrode of the third row; the fourth air outlet connecting pipe 18 is located directly above the second electrode of the fourth row; and the fifth air outlet connecting pipe 19 is located directly above the first electrode of the fifth row. The other side has parallel gas flow paths 24. This asymmetrical layout has been optimized through computational fluid dynamics simulation to maximize the diversity of the flow field distribution and maximize the breaking of transport symmetry. Through-hole second screw holes 20 are located at the four corners of the cavity.

[0076] The single-sensitive-material parallel gas sensor array 12 is a flexible printed circuit board based on a six-layer stacked structure, consisting of, from top to bottom: a top PI reinforcing plate, a top solder mask layer, a top circuit layer, a PI substrate insulating layer, a bottom solder mask layer, and a bottom PI reinforcing plate. Precise windows are made in the top PI reinforcing plate and the top solder mask layer to form 20 circular microgrooves 21 in five rows and four columns. Each circular microgroove 21 integrates a set of interdigitated electrodes consisting of six electrode fingers at its center. Using a drop-coating process, 2 μL of a single gas sensing material is drop-coated onto the interdigitated electrodes within each microgroove, forming a single-sensitive-material sensing unit.

[0077] The parallel bottom cover 13 is a flat plate that matches the size of the parallel top cavity 10, and a second screw hole is provided at the corresponding position. The parallel rubber sealing ring 11 is a rectangular rubber film that matches the size of the parallel top cavity 10, and an opening corresponding to the microgroove array area and four through holes corresponding to the second screw hole 20 are laser-cut on it to ensure the airtightness between the sensor array and the parallel top cavity.

[0078] The parallel top cavity 10, parallel rubber sealing ring 11, single sensitive material parallel gas sensor array 12, and parallel bottom cover 13 are stacked sequentially from top to bottom, with their four second screw holes 20 aligned one by one; four screws are inserted from the second screw holes 20 of the parallel top cavity 10, and then nuts are tightened on one side of the parallel bottom cover 13 to complete the assembly of the static coded electronic nose.

[0079] Due to the asymmetry of the air outlet layout, the static coded electronic nose of this embodiment forms a unique velocity field, pressure field, and concentration gradient field distribution in the chamber, thereby exciting a characteristic dynamic response mode on a single sensitive material sensor array.

[0080] See Figure 12The electronic nose was applied to a complex wheat odor fingerprint recognition experiment: A total of 205 wheat samples were collected, covering 12 odor categories (corresponding to x and y coordinates: 0 for bitterness, 1 for sourness, 2 for rancidity, 3 for putridity, 4 for pesticide odor, 5 for mold, 6 for musty odor, 7 for insect infestation, 8 for artificial flavoring, 9 for storage pesticide odor, 10 for normal fresh wheat odor, and 11 for normal aged wheat odor). The dynamic resistance response of 20 sensor channels was continuously recorded. After preprocessing, time-domain features (response peak, response time, recovery time, etc.) were extracted. A random forest classifier was used for 5-fold cross-validation, achieving an average accuracy of 92.0%. The confusion matrix is ​​shown below. Figure 12 As shown, the electronic nose in this embodiment can effectively achieve accurate identification of wheat odor and construction of fingerprint spectrum.

[0081] Example 4: Constructing a dynamic and statically encoded olfactory recognition platform and an embodied intelligent olfactory interface

[0082] The dynamically coded electronic nose 102 assembled in Example 1, together with the programmable airflow dynamic coding controller 101 and the data acquisition and processing unit 103, constitutes a dynamic coding olfactory recognition platform. This platform can dynamically adjust the dynamic coding strategy of the air outlet to adapt to different recognition tasks. The statically coded electronic nose 104 assembled in Example 3, together with the data acquisition and processing unit 103, constitutes a static coding olfactory recognition platform. Furthermore, both platforms can be integrated into a robot system as the fifth modality of embodied intelligence—the olfactory interface—enabling the robot to perceive odor information in the environment in real time and fuse it with visual, auditory, and other modal data, thereby improving the robot's perception and decision-making capabilities in complex environments.

[0083] Industrial applicability

[0084] The dynamic coding olfactory recognition platform and its gas sensing method for embodied intelligence provided by this invention have advantages such as simple structure, low cost, high recognition accuracy, and strong programmability, and can be widely used in (e.g.) Figure 13 (as shown)

[0085] Robotic environmental perception and hazardous gas detection;

[0086] UAV-based air pollution monitoring;

[0087] Wearable devices for personal health management;

[0088] IoT-based smart home air quality monitoring;

[0089] Applications include industrial process control and food safety testing.

[0090] The implementation of this invention will greatly promote the integrated application of olfactory perception technology in embodied intelligence systems, providing intelligent agents with more comprehensive environmental perception capabilities.

Claims

1. A dynamic coding olfactory recognition platform for embodied intelligence, characterized in that, include: A programmable airflow dynamic coding controller (101) is used to generate airflow dynamic coding instructions; The dynamic coding electronic nose (102) is connected to the programmable airflow dynamic coding controller. It has a gas flow path and a single sensitive material gas sensor array inside, and several air outlets are provided on the cavity. The air outlets are selectively opened or closed according to the airflow dynamic coding command to introduce a transport symmetry disruption in the gas path, so that the airflow flowing through the single sensitive material gas sensor array forms different airflow paths. The data acquisition and processing unit (103) is connected to the gas sensor array of the dynamic coded electronic nose (102) and is used to acquire the dynamic response signals generated by the gas sensor array under different airflow paths, and identify the gas type or concentration based on the dynamic response signals; the data acquisition and processing unit (103) includes a multi-channel data acquisition module, a feature extraction module and a pattern recognition module.

2. The dynamic coding olfactory recognition platform according to claim 1, characterized in that, The dynamic coded electronic nose (102) includes a top cavity (2), a rubber sealing ring (3), a gas sensor array (4), a bottom cover (5), a flexible interface (23), and fasteners; the top cavity (2) is integrally formed from 316L stainless steel powder using selective laser melting technology, with a groove integrated on one side as a gas flow path (7), and several hollow connecting tubes (8) integrated on the other side as air inlets and outlets, and through first screw holes (9) at the four corners of the cavity; the gas sensor array (4) is based on flexible printed circuit board... The sensor array on the circuit board is coated with a single sensitive material; the bottom cover (5) is a flat plate that matches the size of the top cavity (2) and has screw holes at corresponding positions; the rubber sealing ring (3) is a rectangular rubber film that matches the size of the top cavity (2) and has an opening with the same shape as the gas flow path (7) and a through hole corresponding to the first screw hole (9) cut out by laser; the top cavity (2), the rubber sealing ring (3), the gas sensor array (4), and the bottom cover (5) are stacked from top to bottom and fixed by fasteners.

3. The dynamic coding olfactory recognition platform according to claim 1, characterized in that, The programmable airflow dynamic coding controller (101) realizes independent on / off control of the air outlet by controlling the valve array (100), and the valve array (100) is connected one-to-one with several air outlets of the dynamic coding electronic nose (102).

4. The dynamic coding olfactory recognition platform according to claim 1, characterized in that, The multi-channel data acquisition module is used to acquire dynamic response signals; the feature extraction module is used to extract time-domain, frequency-domain, or time-frequency-domain features related to gas types from the dynamic response signals; the pattern recognition module performs classification or regression analysis on the extracted features based on machine learning algorithms to achieve gas type identification or concentration prediction.

5. A statically encoded olfactory recognition platform for embodied intelligence, characterized in that, include: The static coded electronic nose (104) has a single sensitive material parallel gas sensor array (12) in its cavity. The cavity has several air outlets, which are distributed in an asymmetrical diagonal pattern. Due to the inconsistent positions of the air outlets, the transport symmetry is broken, causing the airflow through the single sensitive material parallel gas sensor array (12) to form different airflow paths. The data acquisition and processing unit (103) is connected to the single-sensitive material parallel gas sensor array (12) of the statically encoded electronic nose (104) and is used to acquire the dynamic response signals generated by the sensor array under different airflow paths, and identify the gas type or concentration based on the dynamic response signals; the data acquisition and processing unit (103) includes a multi-channel data acquisition module, a feature extraction module and a pattern recognition module.

6. The statically encoded olfactory recognition platform according to claim 5, characterized in that, The statically coded electronic nose (104) includes: a parallel top cavity (10), a parallel rubber sealing ring (11), a single sensitive material parallel gas sensor array (12), a parallel bottom cover (13), and a parallel flexible interface (22). The parallel top cavity (10) is integrally formed from 316L stainless steel powder using 3D printing technology; one side of its main body is provided with an air inlet connecting pipe (14) and several independent air outlet connecting pipes, and the other side is integrated with a groove as a parallel gas flow path (24). The parallel bottom cover (13) is a flat plate that matches the size of the parallel top cavity (10); the parallel rubber sealing ring (11) is a rectangular rubber film that matches the size of the parallel top cavity (10), and laser-cuts out through holes with the same shape as the parallel gas flow path (24) and several second screw holes (20). The parallel top cavity (10), parallel rubber sealing ring (11), single sensitive material parallel gas sensor array (12), and parallel bottom cover (13) are stacked from top to bottom, with their several second screw holes (20) aligned one by one; several screws are inserted from the second screw holes (20) of the parallel top cavity (10), and then nuts are tightened on one side of the parallel bottom cover (13) to form a static coded electronic nose.

7. A gas sensing method based on the dynamic coding olfactory recognition platform of claim 1, characterized in that, Includes the following steps: S1: The programmable airflow dynamic coding controller generates airflow dynamic coding instructions to selectively open or close multiple air outlets of the dynamic coding electronic nose, introducing a transport symmetry disruption in the gas flow path. S2: The gas to be measured is introduced into the dynamic coded electronic nose. The gas flows through a single sensitive material gas sensor array under the action of the airflow path, which excites a dynamic response signal. S3: The dynamic response signal of the gas sensor array is acquired in real time through the data acquisition and processing unit and preprocessed. S4: Extract feature parameters from the preprocessed dynamic response signal to construct a gas fingerprint feature vector; S5: Input the gas fingerprint feature vector into the pattern recognition module and output gas type or concentration information.

8. A gas sensing method based on the static coding olfactory recognition platform of claim 5, characterized in that, Includes the following steps: S1: The gas to be measured is introduced into the statically encoded electronic nose. The gas flows through a parallel gas sensor array with a single sensitive material under the action of the airflow path, which excites a dynamic response signal. S2: The data acquisition and processing unit acquires the dynamic response signal of the parallel gas sensor array of a single sensitive material in real time and performs preprocessing. S3: Extract feature parameters from the preprocessed dynamic response signal to construct a gas fingerprint feature vector; S4: Input the gas fingerprint feature vector into the pattern recognition module and output gas type or concentration information.

9. An application of the coded olfactory recognition platform as described in claim 1 or 5 in the identification of various volatile organic compounds, characterized in that, Used to distinguish different types of volatile organic compounds or to detect their concentration.

10. The application according to claim 9, characterized in that, The platform is integrated into robots, drones, wearable devices, or IoT terminals as an olfactory sensing module of embodied intelligence.

Images