Swallowing and breathing coordination evaluation method based on millimeter wave radar

Abstract

The invention discloses a swallowing breathing coordination evaluation method based on a millimeter wave radar, and belongs to the technical field of physiological signal non-contact detection and intelligent health evaluation. The method comprises the following steps: arranging a millimeter wave radar in front of a detected object, detecting a neck front-upper chest area of the detected object, establishing a throat swallowing area and a chest breathing area based on distance window separation and angular domain energy separation, respectively extracting throat transient displacement characteristics and chest breathing displacement characteristics, and constructing quality evaluation parameters; and synchronously identifying a swallowing starting point, a swallowing peak value, a swallowing ending point, an inspiration phase, an expiration phase and an apnea section, carrying out time sequence alignment, and constructing a swallowing-breathing event chain. And performing aggregation evaluation based on a multi-time swallowing-breathing event chain, and outputting a mistaken inhalation risk grade, a swallowing coordination score and a rehabilitation training feedback index in combination with quality gating. The method can be used for swallowing function screening, rehabilitation evaluation and follow-up visit monitoring, and has a good clinical application prospect.

Description

Technical Field

[0001] This invention belongs to the field of non-contact detection of physiological signals and intelligent health assessment technology. Specifically, it relates to a swallowing and breathing coordination assessment method based on millimeter-wave radar, and more particularly to a swallowing function assessment method that uses millimeter-wave radar to jointly perceive, align, aggregate, and assess the swallowing movements of the larynx and the breathing movements of the chest of the test subject. Background Technology

[0002] Swallowing function assessment is a crucial component of aspiration risk screening and rehabilitation training guidance. During normal swallowing, there is a relatively stable coordination between the swallowing action and the breathing process. The pre-swallowing respiratory phase, apnea during swallowing, and post-swallowing respiratory recovery pattern are all closely related to aspiration risk. Therefore, accurately assessing the degree of coordination between swallowing and breathing is a key issue in swallowing disorder monitoring.

[0003] Current methods for assessing swallowing function mainly include video fluoroscopy swallowing examination, fiberoptic endoscopy, surface electromyography, neck vibration detection, respiratory chest strap monitoring, and manual observation and assessment. While these methods can obtain swallowing or respiratory information to some extent, they generally suffer from problems such as complex equipment, strong contact dependence, insufficient continuous monitoring capabilities, and limited home applications. Millimeter-wave radar, with its non-contact and continuous monitoring characteristics, has been used in recent years for human respiration, heart rate, and body movement detection, and some studies have also attempted to apply it to swallowing monitoring.

[0004] However, existing technologies still have the following shortcomings. First, existing millimeter-wave radar swallowing detection schemes mostly focus on acquiring signals from a single site, making it difficult to simultaneously and stably acquire larynx swallowing information and chest breathing information in the same detection chain, thus making it difficult to support joint analysis of swallowing and breathing. Second, even if existing schemes can detect swallowing events, they mostly stop at whether swallowing occurred or the number of swallows is counted, lacking a mechanism to accurately align the swallowing start point, swallowing peak, swallowing end point with the inspiratory phase, expiratory phase, and apnea segment on a unified time axis, making it difficult to establish a complete swallowing-breathing event chain. Third, existing risk assessments are mostly based on single swallowing results, lacking aggregation analysis of multiple swallowing results, making it difficult to reflect the stability and recovery pattern of swallowing coordination; at the same time, when there is significant body movement, obvious head deflection, interference from multiple people, or insufficient signal quality, existing schemes usually lack effective quality gating mechanisms, which can easily reduce the credibility of the assessment results.

[0005] Therefore, there is an urgent need to propose a swallowing and breathing coordination assessment method based on millimeter-wave radar to achieve joint perception of the larynx swallowing area and the chest breathing area, unified time axis alignment of swallowing events and breathing stages, multiple swallowing aggregation assessments, and risk output under quality gating. Summary of the Invention

[0006] To address the shortcomings of existing millimeter-wave radar swallowing detection schemes, such as the difficulty in simultaneously and stably acquiring laryngeal swallowing information and chest breathing information in the same detection link, the difficulty in establishing a complete swallowing-breathing event chain under a unified time axis, and the lack of aggregated analysis of multiple swallows and quality gating mechanisms for risk assessment, this invention provides a swallowing and breathing coordination assessment method based on millimeter-wave radar. This method achieves dual-region joint perception of laryngeal swallowing movements and chest breathing movements, temporal alignment of swallowing events and breathing stages, aggregated assessment of multiple swallowing results, and output of aspiration risk and swallowing function score under quality-controlled conditions, thereby improving the completeness, stability, and reliability of swallowing function assessment.

[0007] To achieve the above objectives, the present invention specifically adopts the following technical solution: A method for assessing swallowing and breathing coordination based on millimeter-wave radar includes the following steps: S1: The millimeter-wave radar is deployed at the detection position corresponding to the front of the neck and upper chest of the subject, and the radar echo signal of the region is collected; based on range window separation and angular domain energy separation, the larynx swallowing region and the chest breathing region are established, the transient displacement features of the larynx and the respiratory displacement features of the chest are extracted respectively, and quality evaluation parameters are constructed. S2: Based on the transient displacement features of the larynx, identify the swallowing initiation point, swallowing peak and swallowing end point; based on the respiratory displacement features of the chest, identify the inspiratory phase, expiratory phase and apnea segment; and align the swallowing events and respiratory phases on a unified time axis to obtain the pre-swallowing respiratory phase, swallowing apnea length, post-swallowing recovery delay and post-swallowing abnormal inspiratory events, and construct a swallowing-breathing event chain for a single swallowing. S3: Based on the swallowing-breathing event chain corresponding to multiple swallowings, extract swallowing coordination stability features, post-swallowing recovery pattern features, abnormal inhalation ratio features, and weak residual movement features after swallowing. Combined with the quality evaluation parameters, the results of the gating are used to output the aspiration risk level, swallowing coordination score, and rehabilitation training feedback indicators.

[0008] Furthermore, the millimeter-wave radar in S1 is positioned facing the front of the neck and upper chest region of the subject, so that the radar detection field of view simultaneously covers the throat region and the chest region, so as to simultaneously acquire the micro-movement information of the throat swallowing and the respiratory motion information of the chest in the same detection link.

[0009] Furthermore, the establishment of the larynx swallowing region and the chest breathing region in S1 is specifically as follows: the radar echo signal is windowed in the range direction to determine the candidate larynx distance window corresponding to the larynx depth range and the candidate chest distance window corresponding to the upper chest depth range; the angular energy distribution within each candidate distance window is analyzed to determine the larynx energy concentration region and the chest energy concentration region respectively; and the larynx swallowing region and the chest breathing region are formed by combining the range position constraint and the angular energy continuity constraint respectively.

[0010] Furthermore, the extraction of the transient displacement features of the larynx and the respiratory displacement features of the chest in S1 is specifically as follows: based on at least one of the phase change features, energy centroid change features, and peak trajectory change features in the larynx swallowing region, the transient displacement features of the larynx representing the transient motion of swallowing are extracted; based on the periodic displacement change features in the chest respiratory region, the respiratory displacement features of the chest representing the fluctuations of breathing are extracted; and a unified sampling time axis is established for the transient displacement features of the larynx and the respiratory displacement features of the chest.

[0011] Furthermore, the quality evaluation parameters in S1 are specifically as follows: based on the radar echo signal and the tracking results of the larynx swallowing region and the chest breathing region, the intensity of body movement interference, the degree of head deflection, the degree of multi-target interference, the tracking stability of the larynx swallowing region, the tracking stability of the chest breathing region, the significance of the swallowing signal, and the significance of the breathing signal are obtained to characterize the validity and reliability of the current detection data.

[0012] Furthermore, the identification of the swallowing initiation point, swallowing peak point, and swallowing end point in S2 is specifically as follows: based on the amplitude change, slope change, local extreme value characteristics, and duration constraints of the transient displacement characteristics of the larynx, the swallowing initiation point, swallowing peak point, and swallowing end point are determined, and the time interval between the swallowing initiation point and the swallowing end point is used as the swallowing time window corresponding to a single swallowing event.

[0013] Furthermore, the respiratory phase identification and timing alignment in S2 specifically involves: identifying the inspiratory phase, expiratory phase, and apnea segment based on the displacement direction, periodic turning point, and amplitude changes of the chest respiratory displacement characteristics; mapping the swallowing time window to the respiratory phase corresponding to the chest respiratory displacement characteristics to determine the pre-swallowing respiratory phase; calculating the apnea length during swallowing and the post-swallowing recovery delay based on the changes in respiratory state within the swallowing time window and the preset observation windows before and after it; and identifying abnormal premature inspiration and abnormal respiratory phase switching that occur within a preset time after swallowing as abnormal inspiratory events after swallowing.

[0014] Furthermore, the swallowing-breathing event chain in S2 is specifically as follows: taking a single swallowing event as the analysis unit, the pre-swallowing respiratory phase, the length of the swallowing pause, the post-swallowing recovery delay, and the post-swallowing abnormal inhalation event are associated according to a unified time axis to form a swallowing-breathing event chain that characterizes the coordination relationship of a single swallowing event.

[0015] Furthermore, the S3 multiple swallowing aggregation assessment specifically involves: constructing swallowing-breathing event chains for each swallowing; statistically analyzing the pre-swallowing respiratory phase distribution, swallowing pause length distribution, post-swallowing recovery delay distribution, the proportion of abnormal inspiratory events after swallowing, and the occurrence of weak residual movements after swallowing for each swallowing; forming swallowing coordination stability index, recovery pattern consistency index, and aspiration risk aggregation index based on the above statistical results; and outputting aspiration risk level and swallowing coordination score based on the indicators.

[0016] Furthermore, the result gating in S3 specifically involves: determining whether the current detection result meets the preset quality conditions based on the quality evaluation parameters; when the quality evaluation parameters meet the preset quality conditions, outputting the corresponding aspiration risk level and swallowing coordination score; when the quality evaluation parameters do not meet the preset quality conditions, outputting a low confidence indicator and performing rejection processing, pausing the output of the high confidence aspiration risk level and high confidence swallowing coordination score.

[0017] Furthermore, the rehabilitation training feedback indicators in S3 are specifically composed of at least two of the following: safe swallowing percentage, average pause length during swallowing, average recovery time after swallowing, abnormal inspiratory cut-off percentage, degree of fluctuation in swallowing coordination, and frequency of weak residual movements after swallowing. These indicators are used to characterize the trend of swallowing function changes and training effects of the tested subject.

[0018] Furthermore, the aforementioned weak residual movements after swallowing specifically refer to low-amplitude additional micro-movements that occur in the laryngeal region within a preset observation time after the main swallowing event has ended, which are used to help characterize abnormal manifestations related to post-swallowing residues.

[0019] Compared with the prior art, the present invention has the following beneficial effects: First, this invention establishes a larynx swallowing region and a chest breathing region under the same millimeter-wave radar detection system, and extracts transient displacement features of the larynx and respiratory displacement features of the chest respectively, thereby simultaneously acquiring swallowing action information and breathing process information under non-contact conditions. This overcomes the problem of information fragmentation and difficulty in joint analysis caused by single-site detection in the prior art, and provides a unified and stable data foundation for swallowing and breathing coordination assessment.

[0020] Secondly, this invention can perform corresponding analysis of the swallowing start point, swallowing peak, swallowing end point and inspiratory phase, expiratory phase and apnea segment under a unified time axis, to obtain the pre-swallowing respiratory phase, swallowing apnea length, post-swallowing recovery delay and post-swallowing abnormal inspiratory events, thereby establishing a complete swallowing-breathing event chain. Compared with existing schemes that only perform swallowing detection or swallowing counting, it can more comprehensively reflect the coordination state between swallowing and breathing.

[0021] Third, this invention statistically aggregates the swallowing-breathing event chain corresponding to multiple swallowings to form swallowing coordination stability index, recovery pattern consistency index, and aspiration risk aggregation index, thereby reducing the impact of occasional abnormalities, instantaneous body movements, and single noises on the assessment results, making the output results more reflective of the overall swallowing function status of the tested subject, and more in line with the clinical assessment needs based on comprehensive judgment of multiple performances.

[0022] Fourth, this invention incorporates parameters such as the intensity of body movement interference, the degree of head deflection, the degree of multi-target interference, the tracking stability of the larynx swallowing area and the chest breathing area, and the signal significance into the quality evaluation system. When the detection quality meets the preset conditions, the corresponding evaluation results are output. When the detection quality does not meet the preset conditions, a low confidence indicator is output and rejection processing is performed. This avoids outputting unreliable high-confidence-risk conclusions under low-quality data conditions, thereby improving the safety and reliability of the system in medical screening, rehabilitation monitoring, and home application scenarios. Attached Figure Description

[0023] Figure 1 Overall Flowchart of the Method Figure 2 Detection scene and dual regions of interest schematic diagram Figure 3 Schematic diagram of dual ROI establishment and dual displacement feature extraction Figure 4 Schematic diagram of swallowing-breathing event chain time alignment Figure 5 Schematic diagram of multiple swallowing aggregation assessment and quality gating Detailed Implementation The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0024] This embodiment uses a bedside sip-swallowing screening scenario as an example to illustrate a swallowing-breathing coordination assessment method based on millimeter-wave radar. It should be noted that this invention is not limited to the drinking and swallowing scenario, but is also applicable to dry swallowing tasks and spontaneous swallowing monitoring scenarios under natural resting conditions; as long as there are identifiable swallowing events within the observation period, the swallowing-breathing coordination relationship can be analyzed and assessed according to the method of this invention.

[0025] like Figure 1 As shown, the method in this embodiment includes three stages: joint sensory modeling of the laryngeal and thoracic regions, alignment of swallowing events with respiratory stages, and aspiration risk and swallowing function scoring; as shown Figure 2 As shown, the millimeter-wave radar is positioned directly in front of the object being measured, with its field of view covering the area from the front of the neck to the upper chest; as Figure 3 As shown, range window separation and angular domain energy separation are performed on the radar echo signal to establish the larynx swallowing region and the chest breathing region, and the transient displacement features of the larynx and the respiratory displacement features of the chest are extracted respectively; Figure 4 As shown, the timing of key swallowing events and the breathing phase is aligned on a unified timeline; as... Figure 5 As shown, multiple swallowing outcomes are aggregated and evaluated, and under quality gating conditions, the aspiration risk level, swallowing coordination score, and rehabilitation training feedback indicators are output.

[0026] S1, Joint Sensing Modeling of Throat and Chest Regions In this embodiment, the subject assumes a natural sitting posture with the head and neck kept basically upright. The millimeter-wave radar is positioned in front of the subject, with the radar detection direction aimed at the front of the neck and upper chest area. During the detection process, the subject completes multiple small sips of water swallowing tasks as prompted, maintaining short periods of natural breathing between each swallow to collect complete information on respiratory changes before, during, and after swallowing.

[0027] In this step, radar echo signals from the anterior neck to upper chest region of the subject are first acquired. Then, the echo signals are windowed in the range direction to determine candidate range windows for the larynx at a corresponding depth range and candidate range windows for the upper chest at a corresponding depth range. Next, the angular energy distribution within each candidate range window is analyzed to determine the energy concentration regions for the larynx and chest, respectively. Finally, by combining range position constraints and angular energy continuity constraints, the larynx swallowing region and the chest breathing region are formed.

[0028] Based on at least one of the phase change characteristics, energy centroid change characteristics, and peak trajectory change characteristics in the larynx swallowing region, transient displacement characteristics of the larynx that characterize the transient motion of swallowing are extracted; based on the periodic displacement change characteristics in the chest breathing region, chest respiratory displacement characteristics that characterize the fluctuations of breathing are extracted; and a unified sampling time axis is established for the two types of displacement characteristics to provide a consistent time reference for subsequent time alignment.

[0029] Simultaneously, this embodiment constructs quality evaluation parameters to assess the validity and reliability of the current detection data. These parameters include the intensity of body motion interference, the degree of head deflection, the degree of multi-target interference, the tracking stability of the laryngeal swallowing region, the tracking stability of the chest breathing region, the significance of the swallowing signal, and the significance of the respiratory signal. These quality evaluation parameters can be used to remove obviously abnormal segments and also for subsequent result gating.

[0030] S2, Swallowing events aligned with respiratory phases In this embodiment, the swallowing initiation point, swallowing peak point, and swallowing end point are identified based on the transient displacement characteristics of the larynx. Specifically, the start time, peak time, and end time of a single swallowing event are determined according to the amplitude abrupt change, slope change, local extreme value characteristics, and duration constraints of the transient displacement characteristics of the larynx, and the time interval between the swallowing initiation point and the swallowing end point is defined as the swallowing time window corresponding to that swallowing event.

[0031] To facilitate implementation, in the first During a single swallow, the swallowing time window is defined as: in, For the first The swallowing time window corresponding to each swallow; For the first The moment when swallowing begins; For the first The moment at which swallowing ends. If necessary, the peak swallowing time is recorded as... ,in .

[0032] Simultaneously, based on the displacement direction, periodic turning points, and amplitude changes of the chest respiratory displacement characteristics, the inspiratory phase, expiratory phase, and apnea segment are identified. Subsequently, the swallowing time window is mapped to the respiratory phase corresponding to the chest respiratory displacement characteristics, completing the temporal alignment of swallowing events and respiratory phases on a unified time axis.

[0033] Based on this, define the first The length of the pausing phase during each swallow is: in, For the first The length of the swallowing pause during each swallow; In order to be with the first The onset of the first swallowing-related apnea; In order to be with the first The moment of cessation of swallowing-related breathing apnea; and and It is determined by changes in respiratory status within the swallowing time window and within the preset observation windows before and after it.

[0034] Furthermore, define the first The recovery time after swallowing is: in, For the first Post-swallowing recovery time delay; For the first The moment at the end of swallowing; For the first The moment when breathing returns to a normal rhythm after swallowing. The "normal rhythm" can be determined by the return of chest respiratory displacement characteristics to a fluctuation pattern consistent with the adjacent respiratory cycle before swallowing.

[0035] In this embodiment, the pre-swallowing respiratory phase, swallowing pause length, post-swallowing recovery delay, and post-swallowing abnormal inspiratory events of a single swallow are correlated along a unified time axis to form a swallowing-breathing event chain corresponding to that swallow, which is used to characterize the coordination relationship between the swallowing action and the breathing process. A post-swallowing abnormal inspiratory event refers to abnormal premature inspiratory behavior or abnormal respiratory phase switching behavior that occurs within a preset time range after swallowing.

[0036] S3, Aspiration Risk and Swallowing Function Score In this embodiment, after the subject completed multiple swallowing tasks, a swallowing-breathing event chain was constructed for each swallowing and then aggregated for evaluation. Specifically, the distribution of pre-swallowing respiratory phase, swallowing pause length, post-swallowing recovery delay, proportion of abnormal inspiratory events after swallowing, and occurrence of weak residual movements after swallowing were statistically analyzed in multiple swallowings. Based on the above statistical results, swallowing coordination stability index, recovery pattern consistency index, and aspiration risk aggregation index were formed.

[0037] To characterize the overall incidence of abnormal inspiratory events after swallowing in multiple swallowing events, the abnormal inspiratory cutoff ratio is defined as: To characterize the overall incidence of abnormal inspiratory events after swallowing in multiple swallowing events, the abnormal inspiratory cutoff ratio is defined as: in, This refers to the abnormal intake cut-off ratio; The number of swallowing events that occurred during the detection period; The total number of swallows included in the statistics, and .

[0038] To characterize the consistency of the swallowing-breathing timing relationship during multiple swallowing processes, the swallowing coordination stability index is defined as: in, As an indicator of swallowing coordination stability; The degree of dispersion of the respiratory phase before multiple swallows; The standard deviation of the length of multiple swallowing pauses; The standard deviation of the recovery time after multiple swallows; , , is a non-negative weighting coefficient used to adjust the contribution of the three types of discrete quantities to the stability index. From the above formula, it can be seen that... , and The smaller, the better The larger the value, the more stable the coordination of multiple swallows.

[0039] In this embodiment, the weak residual motion after swallowing refers to the low-amplitude additional micro-movements that appear in the laryngeal region within a preset observation time after the main swallowing event, used to help characterize abnormal manifestations related to post-swallowing residue. This feature can be used together with features such as abnormal inspiratory cut-off ratio and recovery delay to participate in the construction of an aggregated index for aspiration risk.

[0040] After completing the aggregate evaluation, a gating decision is made based on the aforementioned quality evaluation parameters. Therefore, the comprehensive quality score is defined as: in, For overall quality scoring; The mass sub-item corresponding to the intensity of bodily motion disturbance; The quality sub-item corresponding to the degree of head deflection; This refers to the quality sub-item corresponding to the degree of multi-target interference; The quality sub-item corresponding to the tracking stability of the larynx swallowing region; The quality sub-item corresponding to the tracking stability of the chest respiratory region; The quality sub-item corresponding to the significance of the swallowing signal; The quality sub-item corresponding to the significance of the respiratory signal; The weighting coefficients corresponding to each quality sub-item satisfy the following conditions: .

[0041] Furthermore, the quality gating condition can be expressed as: in, This is a preset quality threshold. When... When, output the corresponding aspiration risk level and swallowing coordination score; when When this occurs, a low-confidence flag is output, and rejection processing is performed, pausing the output of high-confidence aspiration risk level and high-confidence swallowing coordination score.

[0042] When the quality evaluation parameters meet the preset quality conditions, this embodiment further outputs rehabilitation training feedback indicators. The rehabilitation training feedback indicators may consist of at least two of the following: safe swallowing ratio, average pause length during swallowing, average recovery delay after swallowing, abnormal inspiratory cutoff ratio, degree of fluctuation in swallowing coordination, and frequency of weak residual movements after swallowing. These indicators are used to characterize the trend of swallowing function changes and training effects of the tested subject.

[0043] Through the above steps, this embodiment can simultaneously acquire swallowing and breathing information of the subject under non-contact conditions, and establish a swallowing-breathing event chain on a unified timeline, thereby achieving multiple swallowing aggregation assessments and aspiration risk output under quality gating. Compared with existing methods that only identify swallowing events or only monitor respiratory signals, this embodiment can more comprehensively reflect the coordination state between swallowing and breathing of the subject, and is more suitable for bedside screening, rehabilitation training assessment, and long-term follow-up monitoring scenarios.

Claims

1. A method for assessing swallowing and breathing coordination based on millimeter-wave radar, characterized in that, Includes the following steps: S1: Place the millimeter-wave radar in front of the person being tested and collect the radar echo signal of the front of the neck to the upper chest region of the subject; based on range window separation and angular domain energy separation, establish the larynx swallowing region and the chest breathing region, extract the transient displacement features of the larynx and the respiratory displacement features of the chest respectively, and construct quality evaluation parameters; S2: Based on the transient displacement features of the larynx, identify the swallowing initiation point, swallowing peak and swallowing end point; based on the respiratory displacement features of the chest, identify the inspiratory phase, expiratory phase and apnea segment; and align the swallowing event with the respiratory phase in time to obtain the pre-swallowing respiratory phase, swallowing apnea length, post-swallowing recovery delay and post-swallowing abnormal inspiratory event, and construct a single swallowing-breathing event chain; S3: Based on multiple swallowing-breathing event chains, extract swallowing coordination stability features, post-swallowing recovery pattern features, abnormal inhalation ratio features, and post-swallowing movement features, and combine the quality evaluation parameters with the execution result gating to output aspiration risk level, swallowing coordination score, and rehabilitation training feedback indicators.

2. The swallowing and breathing coordination assessment method based on millimeter-wave radar according to claim 1, characterized in that, The step S1, which establishes the larynx swallowing region and the chest breathing region, includes: performing range windowing on the radar echo signal to determine candidate larynx range windows corresponding to the larynx depth range and candidate chest range windows corresponding to the chest depth range; analyzing the angular energy distribution within each candidate range window to determine the larynx energy concentration region and the chest energy concentration region respectively; and combining the range position constraint and the angular energy continuity constraint to form the larynx swallowing region and the chest breathing region.

3. The swallowing and breathing coordination assessment method based on millimeter-wave radar according to claim 1, characterized in that, Step S1 involves extracting transient laryngeal displacement features and chest respiratory displacement features, including: extracting transient laryngeal displacement features characterizing swallowing transient motion based on at least one of phase change features, energy centroid change features, and peak trajectory change features within the laryngeal swallowing region; extracting chest respiratory displacement features characterizing respiratory fluctuations based on periodic displacement change features within the chest respiratory region; and establishing a unified time axis for the transient laryngeal displacement features and chest respiratory displacement features.

4. The swallowing and breathing coordination assessment method based on millimeter-wave radar according to claim 1, characterized in that, The method for identifying the swallowing initiation point, swallowing peak point, and swallowing end point in step S2 is as follows: based on the amplitude change, slope change, local extreme value characteristics, and duration constraints of the transient displacement characteristics of the larynx, the swallowing initiation point, swallowing peak point, and swallowing end point are determined; and the time interval between the swallowing initiation point and the swallowing end point is used as the swallowing time window corresponding to a single swallowing event.

5. The swallowing and breathing coordination assessment method based on millimeter-wave radar according to claim 1, characterized in that, Step S2 involves aligning the swallowing event with the respiratory phase, including: identifying the inspiratory phase, expiratory phase, and apnea segment based on the displacement direction, periodic turning point, and amplitude changes of the chest respiratory displacement characteristics; mapping the swallowing time window to the respiratory phase corresponding to the chest respiratory displacement characteristics to determine the pre-swallowing respiratory phase; calculating the apnea length during swallowing and the post-swallowing recovery delay based on the changes in respiratory state within the swallowing time window and within preset observation windows before and after it; and identifying any abnormal premature transition from a non-inspiratory state to an inspiratory state that occurs within a preset time after swallowing as an abnormal inspiratory event after swallowing.

6. The swallowing and breathing coordination assessment method based on millimeter-wave radar according to claim 1, characterized in that, Step S3 involves a comprehensive assessment based on multiple swallows, including: constructing swallowing-breathing event chains for each swallow; statistically analyzing the pre-swallowing respiratory phase distribution, swallowing pause length distribution, post-swallowing recovery delay distribution, the proportion of abnormal inspiratory events after swallowing, and the occurrence of weak residual movements after swallowing for each swallow; generating swallowing coordination stability indicators, recovery pattern consistency indicators, and aspiration risk aggregation indicators based on the above statistical results; and outputting aspiration risk level and swallowing coordination score based on the indicators.

7. The swallowing and breathing coordination assessment method based on millimeter-wave radar according to claim 1, characterized in that, The quality evaluation parameters include the intensity of body movement interference, the degree of head deflection, the tracking stability of the larynx swallowing area, the tracking stability of the chest breathing area, the significance of the swallowing signal, and the significance of the breathing signal. When the quality evaluation parameters meet the preset quality conditions, the corresponding aspiration risk level, swallowing coordination score, and rehabilitation training feedback index are output. When the quality evaluation parameters do not meet the preset quality conditions, a low confidence indicator is output, and rejection processing is performed, pausing the output of the high confidence aspiration risk level, high confidence swallowing coordination score, and rehabilitation training feedback index.

8. The swallowing and breathing coordination assessment method based on millimeter-wave radar according to claim 7, characterized in that, The rehabilitation training feedback indicators consist of at least two of the following: safe swallowing percentage, average pause length during swallowing, average recovery time after swallowing, proportion of abnormal inspiratory events after swallowing, degree of fluctuation in swallowing coordination, and frequency of weak residual movements after swallowing. These rehabilitation training feedback indicators are used to characterize the trend of swallowing function changes and training effects of the tested subjects.

Images