Augmented Reality Glasses Method and System for Identifying and Alerting Impact Risks While Cycling

By collecting inertial measurement data in head-mounted augmented reality glasses, using head posture and wearing vibration characteristics to suppress false alarms, and combining cycling speed and vehicle information to confirm road impact, the problem of real road impact identification and warning methods is solved, achieving more reliable cycling impact risk identification and low-interference warning.

CN123113022APending Publication Date: 2026-11-17SHIYE TECHNOLOGY (SHENZHEN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610996219.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-06
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately distinguish between real road impacts during cycling and false impacts caused by user movement or glasses shaking in head-mounted augmented reality glasses. Furthermore, existing alert methods may obstruct vision or distract attention.

Method used

By collecting inertial measurement data, false alarms are suppressed by utilizing changes in head posture and vibration characteristics of the device. The expected window of subsequent impacts is determined by combining the current riding speed and vehicle profile. Related impact events are detected, the road impact confidence level is determined, and a low-interference reminder is output based on the confidence level.

Benefits of technology

It improves the reliability of recognizing impacts on real road surfaces, reduces false alarms, and the alert method does not obstruct the rider's view, making it suitable for use in cycling scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113022A_ABST
    Figure CN123113022A_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for identifying and alerting cycling impact risks using augmented reality glasses. The method includes collecting inertial measurement data from head-mounted augmented reality glasses and obtaining the current cycling speed; extracting impact candidate events based on the inertial measurement data; suppressing false alarms due to head movement and loosening of the glasses based on head posture change characteristics and wearing vibration characteristics; determining the expected subsequent impact window corresponding to the impact candidate event; determining the road impact confidence level of the impact candidate event based on the time matching relationship, direction matching relationship, and attenuation matching relationship between the impact candidate event and related impact events; determining the cycling impact risk level based on the road impact confidence level; and outputting a low-interference alert corresponding to the cycling impact risk level in the display area of ​​the head-mounted augmented reality glasses. This invention improves the reliability of measurement data and reduces the obstruction of the cycling view by the alert information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of smart cycling glasses, and in particular to a method and system for recognizing and alerting the risk of impact while cycling using augmented reality glasses. Background Technology

[0002] With the development of augmented reality display technology and sports safety monitoring technology, head-mounted augmented reality glasses are gradually being applied in cycling scenarios. Users can view speed, navigation, exercise status, or safety alerts while cycling, thus reducing the frequency of looking down at devices such as phones and cycling computers, and improving the convenience of information access during cycling.

[0003] During cycling, road surface disturbances such as potholes, speed bumps, transverse cracks, manhole cover edges, and gravel sections are transmitted to the user's head through the wheels, frame, and body, causing corresponding vibration or impact signals to be generated by the inertial measurement units in the head-mounted augmented reality glasses. In existing technologies, some solutions identify abnormal road surfaces using accelerometers, speed information, or positioning information, and issue alerts to the user upon detection. However, these solutions typically use the bicycle, mobile phone, or fixedly installed cycling equipment as the signal acquisition target, resulting in a relatively direct correlation between the acquired vibration signals and road surface disturbances.

[0004] For head-mounted augmented reality glasses, inertial measurement data does not simply reflect road impacts. When a user turns their head to observe, looks down to check, looks up to avoid obstacles, adjusts their center of gravity, or changes their riding posture while cycling, these actions can all create signal changes in the head-mounted inertial measurement data that resemble impacts. Simultaneously, the tightness of the glasses, the elastic vibration of the temples, and the cushioning effect on the head and neck can also alter the signal pattern of road disturbances transmitted to the head. If methods such as single-acceleration peak values, fixed thresholds, or ordinary vibration intensity are still used for judgment, it is easy to misinterpret active head movements or the glasses' own shaking as road impacts, and it is also easy to miss real road impacts whose amplitude changes after being transmitted through vehicles and the human body.

[0005] Furthermore, in cycling scenarios, users need to continuously observe the road ahead. If the alerts displayed in the augmented reality headset are large pop-ups, long-duration text displays, or frequent repetitions, they may obstruct the field of vision or distract attention. Especially at high speeds, downhill, or on complex road conditions, the alert method itself can also affect the user's observation and reaction to road conditions. Therefore, cycling impact risk recognition not only needs to determine whether the impact signal in the head-mounted inertial measurement data originates from actual road surface disturbances, but also needs to combine the cycling status to output alerts in a low-interference manner. Summary of the Invention

[0006] The purpose of this application is to propose an augmented reality glasses cycling impact risk identification and reminder method and system, which aims to enable the glasses to more reliably distinguish between real road impact and false impact caused by active head movement and wearing shaking from head-mounted inertial measurement data, and to remind the user in a way that does not obstruct the cycling vision after identifying the cycling impact risk.

[0007] To address the aforementioned technical problems, this application provides a method for recognizing and alerting the impact risk of cycling using augmented reality glasses, employing the following technical solution: Collect inertial measurement data from the head-mounted augmented reality glasses and obtain the current cycling speed; Impact candidate events are extracted based on the inertial measurement data, and active head movement false alarm suppression and loosening false alarm suppression are performed on the impact candidate events according to the head posture change characteristics and wearing vibration characteristics. Based on the current riding speed and the front and rear wheel spacing in the vehicle profile, determine the subsequent impact expectation window corresponding to the impact candidate event after false alarm suppression; Within the subsequent impact expectation window, associated impact events are detected, and the road impact confidence of the impact candidate events is determined based on the time matching relationship, direction matching relationship, and attenuation matching relationship between the impact candidate events and the associated impact events. The cycling impact risk level is determined based on the road impact confidence level. Based on the cycling impact risk level, a low-interference reminder corresponding to the cycling impact risk level is output in the display area of ​​the head-mounted augmented reality glasses.

[0008] To address the aforementioned technical problems, this application also provides an augmented reality glasses cycling impact risk recognition and alert system, which employs the following technical solution: The data acquisition module is used to collect inertial measurement data from the head-mounted augmented reality glasses and obtain the current cycling speed; The extraction module is used to extract impact candidate events based on the inertial measurement data, and to perform active head movement false alarm suppression and wearing loosening false alarm suppression on the impact candidate events according to the head posture change characteristics and wearing vibration characteristics. The first determining module is used to determine the subsequent impact expectation window corresponding to the impact candidate event after false alarm suppression, based on the current riding speed and the front and rear wheel spacing in the vehicle profile; The second determining module is used to detect associated impact events within the subsequent impact expectation window, and determine the road impact confidence of the impact candidate events based on the time matching relationship, direction matching relationship and attenuation matching relationship between the impact candidate events and the associated impact events; The third determining module is used to determine the cycling impact risk level based on the road impact confidence level. The output module is used to output a low-interference reminder corresponding to the cycling impact risk level in the display area of ​​the head-mounted augmented reality glasses.

[0009] Compared with the prior art, the embodiments of this application have the following main advantages: The augmented reality glasses cycling impact risk identification and alert method disclosed in this application collects inertial measurement data of the head-mounted augmented reality glasses and the current cycling speed. Impact candidate events are extracted from the head-mounted inertial measurement data. Before confirming a road impact, the method actively suppresses false alarms related to head movement and loosening of the glasses by utilizing head posture change characteristics and wearing vibration characteristics. This reduces the likelihood of misinterpreting changes in inertial signals caused by the user turning their head, looking down, looking up, adjusting their riding posture, or shaking the glasses as road impacts. Based on this, the method determines the expected impact window by combining the current cycling speed with the front and rear wheel spacing in the vehicle profile. Finally, it considers the time interval between the impact candidate events and related impact events. The confidence level of road impact is determined by the inter-row matching relationship, directional matching relationship, and attenuation matching relationship. It can further confirm the real road impact in the head-mounted inertial measurement data by utilizing the sequential impact characteristics formed by the front and rear wheels passing through the same road disturbance in turn, thereby improving the reliability of head-mounted augmented reality glasses in identifying cycling impact risks. At the same time, the cycling impact risk level is determined based on the road impact confidence level, and corresponding low-interference reminders are output in the display area according to the display position, display area, continuous display time, and repetitive reminder restriction rules. This can reduce the obstruction of the cycling line of sight by the reminder information while identifying cycling impact risks, making the risk identification and reminder method more suitable for the cycling usage scenario of head-mounted augmented reality glasses. Attached Figure Description

[0010] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart of an embodiment of the augmented reality glasses cycling impact risk identification and reminder method according to this application; Figure 2 This is a structural schematic diagram of an embodiment of the augmented reality glasses cycling impact risk recognition and alert system according to this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0013] refer to Figure 1 The diagram illustrates a flowchart of an embodiment of the augmented reality glasses cycling impact risk identification and warning method according to this application. The augmented reality glasses cycling impact risk identification and warning method includes the following steps: Step S101: Collect inertial measurement data from the head-mounted augmented reality glasses and obtain the current cycling speed.

[0014] Step S102: Based on the inertial measurement data, extract impact candidate events, and based on the head posture change characteristics and wearing vibration characteristics, perform active head movement false alarm suppression and wearing loosening false alarm suppression on the impact candidate events.

[0015] Step S103: Based on the current riding speed and the front and rear wheel spacing in the vehicle profile, determine the subsequent impact expectation window corresponding to the impact candidate event after false alarm suppression.

[0016] Step S104: Detect associated impact events within the subsequent impact expectation window, and determine the road impact confidence of the impact candidate events based on the time matching relationship, direction matching relationship, and attenuation matching relationship between the impact candidate events and the associated impact events.

[0017] Step S105: Determine the cycling impact risk level based on the road impact confidence level.

[0018] Step S106: Based on the cycling impact risk level, output a low-interference reminder corresponding to the cycling impact risk level in the display area of ​​the head-mounted augmented reality glasses.

[0019] The method provided in this embodiment can be executed by a processor within the head-mounted augmented reality glasses, or it can be executed collaboratively by the head-mounted augmented reality glasses and a mobile terminal or cycling computer. The head-mounted augmented reality glasses are equipped with an inertial measurement unit and a display module. The inertial measurement unit is used to collect acceleration and angular velocity data generated by the user's head during cycling, and the display module is used to output prompt information in the user's field of vision. The current cycling speed can be obtained by the positioning module within the head-mounted augmented reality glasses, or it can be provided by a wheel speed sensor, cadence device, mobile terminal, or cycling computer that is communicatively connected to the head-mounted augmented reality glasses.

[0020] During cycling, the head-mounted augmented reality glasses continuously collect inertial measurement data and obtain the current cycling speed. Since the glasses are worn on the user's head, the inertial measurement data includes not only the impact response of road disturbances transmitted through the wheels, frame, body, and glasses, but also signal changes caused by the user's active head movements (turning head, looking down, looking up, observing vehicles to the side and rear), body weight adjustments, and glasses movement. Therefore, this implementation does not directly use a single acceleration peak as the road impact. Instead, it first extracts impact candidate events from the inertial measurement data, then performs active head movement false alarm suppression and loosening false alarm suppression on the impact candidate events based on head posture change characteristics and wearing vibration characteristics. Finally, it further determines whether the impact candidate events after false alarm suppression conform to the sequential impact pattern formed by the front and rear wheels of the vehicle passing through the same road disturbance in turn.

[0021] Specifically, the processor can detect short-term impact peaks in inertial measurement data and combine the peak duration, peak energy, and main impact direction to form impact candidate events. For each impact candidate event, the processor first determines whether there are continuous yaw, pitch, or roll attitude changes within its corresponding time window, and whether the high-frequency vibration energy, impact direction stability, and main peak trailing vibration characteristics match the characteristics of loose fitting. If the impact candidate event matches the characteristics of active head movement or loose fitting, the retention weight of the impact candidate event as a road impact is reduced; if the impact candidate event does not match the characteristics of active head movement or loose fitting, it is treated as an impact candidate event after false alarm suppression and continues to be used for subsequent impact confirmation.

[0022] The processor determines the expected window for subsequent impacts based on the current riding speed and the distance between the front and rear wheels. That is, when the front wheels of the vehicle pass over a speed bump, lateral crack, pothole edge, or raised surface, the rear wheels will typically pass over the same surface disturbance within a short period afterward, thus potentially creating a second impact in the head-mounted inertial measurement data that is temporally and characteristically correlated with the first impact. Within this expected window, the processor continues to detect associated impact events and compares impact candidate events with the associated impact events.

[0023] The comparison process includes time matching, direction matching, and attenuation matching. Time matching determines whether the time interval between two impact events matches the travel time corresponding to the current riding speed and wheelbase. Direction matching determines whether the main impact directions of the two impact events are consistent or similar in the riding reference coordinate system. Attenuation matching determines whether the energy attenuation processes of the two impact events after the peak have similar transmission characteristics. If both the impact candidate event and the associated impact event satisfy the above matching relationships, the processor determines that the impact candidate event is more likely to originate from a real road surface disturbance and increases its road impact confidence. If no associated impact event is detected within the subsequent impact expectation window, or if the detected associated impact event does not match the impact candidate event in time, direction, or attenuation, the processor decreases the road impact confidence of the impact candidate event.

[0024] The processor determines the cycling impact risk level based on the road impact confidence level. When the road impact confidence level is low, the event can be recorded without alerting the user; when the road impact confidence level is high, the impact intensity, current cycling speed, and impact continuity can be further combined to determine the cycling impact risk level. Subsequently, based on the cycling impact risk level, the processor outputs a corresponding low-interference alert in the display area of ​​the head-mounted augmented reality glasses. The low-interference alert can be presented as a short-lived icon, a short-lived color bar, brief text, or brightness variation at the edge of the display area to avoid significantly obstructing the user's view of the road directly in front.

[0025] In this embodiment, instead of directly judging road impact based on a single peak value in head-mounted inertial measurement data, the current riding speed and the distance between the front and rear wheels are used to establish a subsequent impact expectation window. The confidence level of road impact is confirmed by the time, direction and attenuation matching relationship between impact candidate events and related impact events, thereby improving the reliability of head-mounted augmented reality glasses in recognizing real road impacts in cycling scenarios.

[0026] In some optional implementations of this embodiment, before extracting impact candidate events based on the inertial measurement data, the method further includes: When the current cycling speed is within the cycling speed range and the inertial measurement data indicates that the change in the user's head posture is less than the posture stability threshold, a stable cycling time period is determined. Based on the inertial measurement data during the stable riding period, the headband vibration reference and wearing stability reference are determined.

[0027] Specifically, the processor continuously acquires current riding speed and inertial measurement data. When the current riding speed is within the recommended riding speed range, and the inertial measurement data indicates that the change in the user's head posture is less than the posture stability threshold, the processor determines that the user is in a stable riding state. The riding speed range can be set according to the speed range of a regular bicycle, e-bike, or other cycling vehicle. For example, when the current riding speed is higher than a low-speed threshold but lower than a high-speed anomaly threshold, and the changes in the user's head yaw angle, pitch angle, and roll angle are all small within a preset time period, this time period can be determined as a stable riding period.

[0028] During a stable riding period, the processor statistically analyzes the baseline vibration level of inertial measurement data to obtain the head-mounted vibration baseline. The head-mounted vibration baseline can include the baseline energy of linear acceleration, the normal peak range, the normal vibration frequency band, and the normal impact duration range. The processor can also determine a wearing stability baseline based on high-frequency vibration energy, the amplitude of minute head posture swaying, and the dispersion of acceleration direction. The wearing stability baseline characterizes the current stability of the glasses on the user's head; a tighter fit typically results in lower high-frequency vibration energy and directional dispersion, while a looser fit typically results in higher high-frequency vibration energy and directional dispersion.

[0029] In the subsequent impact recognition process, the processor can compare real-time inertial measurement data with head-mounted vibration and wearing stability benchmarks. Only when the real-time data shows a significant short-term anomaly relative to the head-mounted vibration benchmark, and this anomaly does not conform to the characteristics of loose wearing, will further impact candidate events be extracted. This avoids using fixed thresholds to adapt to all users and vehicles, and also reduces misjudgments caused by differences in head posture, vehicle rigidity, tire condition, and the tightness of eyeglasses.

[0030] In some optional implementations of this embodiment, the extraction of impact candidate events based on the inertial measurement data includes: The inertial measurement data is subjected to gravity removal processing to obtain the linear acceleration of the head; The linear acceleration of the head is filtered or decomposed into frequency bands to obtain the road impact-related components. The signal segments that satisfy the impact intensity and duration conditions are extracted from the road impact correlation components, and the signal segments are identified as the impact candidate events.

[0031] Specifically, the acceleration data collected by the inertial measurement unit includes both gravitational and motion components. The processor can combine the acceleration and angular velocity data to perform attitude calculations, estimate the current direction of gravity, and remove the gravitational component from the acceleration data to obtain the linear head acceleration. Linear head acceleration is more suitable for characterizing dynamic impacts on the user's head.

[0032] After obtaining the linear head acceleration, the processor filters or decomposes the head acceleration to obtain the road impact-related component. The road impact-related component can be the mid-frequency signal after removing low-frequency posture changes and high-frequency wearing vibrations. Low-frequency posture changes are usually related to the user's active head tilting, tilting, or turning, while high-frequency wearing vibrations are usually related to the elastic vibration of the temples or loosening of the glasses. Real road impacts are usually transient responses with short duration and concentrated energy. Therefore, filtering or frequency band decomposition can enhance the signal components corresponding to road impacts and reduce interference caused by active head movements and wearing vibrations.

[0033] The processor sets a sliding detection window in the road impact correlation component and calculates the impact intensity, duration, main impact direction, and post-peak attenuation characteristics within each sliding detection window. When the impact intensity of a signal segment exceeds the impact intensity condition and the duration is within a preset duration range, the processor identifies that signal segment as a candidate impact event. The impact intensity condition can be determined based on the head-mounted vibration reference, and the duration condition can be determined based on the characteristic that real road impacts typically have a short-duration main peak. For signal segments with excessively long durations, the processor may consider them more likely to originate from changes in user posture; for signal segments with excessively short durations and excessively high-frequency energy, the processor may consider them more likely to originate from wearing vibrations.

[0034] In some optional implementations of this embodiment, determining the subsequent impact expectation window corresponding to the impact candidate event based on the current riding speed and the front-to-rear wheel spacing in the vehicle profile includes: Based on the front and rear wheel spacing and the current riding speed, determine the expected time interval between the front wheel impact and the rear wheel impact; Based on the fluctuation range of the current riding speed and the impact transmission error range corresponding to the vehicle profile, the expected time interval is extended to obtain the subsequent impact expectation window.

[0035] Specifically, the vehicle profile stores the distance between the front and rear wheels. The higher the current riding speed, the shorter the time interval between the front and rear wheel impacts; the lower the current riding speed, the longer the time interval between the front and rear wheel impacts. The processor can determine the expected time interval between the front and rear wheel impacts based on the correspondence between the front and rear wheel distance and the current riding speed.

[0036] Since speed may fluctuate in a short period of time during actual riding, and road disturbances will have certain delays and shape changes after being transmitted through the wheels, frame, human body and augmented reality glasses, the processor should not set the expected time interval to a single point in time. Instead, it should expand the expected time interval according to the fluctuation range of the current riding speed and the impact transmission error range corresponding to the vehicle image to obtain the subsequent impact expectation window.

[0037] For example, when the current riding speed is stable, the subsequent impact prediction window can be relatively narrow to improve matching accuracy; when the current riding speed fluctuates significantly, or the vehicle profile shows a large impact transmission error, the subsequent impact prediction window can be appropriately widened to prevent the actual rear wheel impact from falling outside the window. The impact transmission error range corresponding to the vehicle profile can be determined based on vehicle type, tire cushioning capacity, frame rigidity, and historical riding data. In this way, the subsequent impact prediction window can adapt to different vehicles and different riding speeds, avoiding fixing the front and rear wheel sequential impact judgment to a single time threshold.

[0038] In some optional implementations of this embodiment, determining whether the associated impact event and the impact candidate event satisfy time matching, direction matching, and attenuation matching includes: When the time interval between the associated impact event and the impact candidate event falls within the subsequent impact expectation window, and the difference between the main impact direction of the associated impact event and the impact candidate event in the riding reference coordinate system is less than the direction matching threshold, and the difference between the energy attenuation after the main peak of the associated impact event and the impact candidate event is less than the attenuation matching threshold, it is determined that the associated impact event and the impact candidate event satisfy the time matching relationship, the direction matching relationship and the attenuation matching relationship.

[0039] Specifically, after detecting a related impact event within the subsequent impact expectation window, the processor first determines whether the time interval between the related impact event and the impact candidate event falls within the subsequent impact expectation window. If the time interval does not fall within the subsequent impact expectation window, it indicates that the related impact event is weakly related to the front and rear wheels passing through the same road surface disturbance in sequence, and the processor may not treat it as a matching event.

[0040] After the time matching is satisfied, the processor establishes a cycling reference coordinate system. This system can be based on the direction of gravity and the current cycling direction, where the direction of gravity determines the vertical direction, and the current cycling direction can be determined by the velocity direction, heading change, or short-time displacement direction. The processor transforms the principal impact directions of the impact candidate events and associated impact events into the cycling reference coordinate system and calculates the difference between their principal impact directions. When the difference is less than the direction matching threshold, the processor determines that the two events satisfy the direction matching relationship. This avoids incorrectly identifying impact peaks caused by the user turning their head, swaying their head to the side, or lateral movement of their glasses as associated impacts to the rear wheel.

[0041] After directional matching is satisfied, the processor further compares the post-peak energy decay process of the candidate impact event and the associated impact event. Post-peak energy decay can be represented by the energy decrease rate, decay duration, or remaining vibration energy within a preset time after the peak. When the difference in post-peak energy decay between the two impact events is less than the decay matching threshold, the processor determines that they satisfy the decay matching relationship. Since the impact decay patterns of the front and rear wheels after passing through the same road surface disturbance are usually similar, while the signal decay patterns caused by loose glasses or active head movements are usually unstable, decay matching can further improve the accuracy of road impact confirmation.

[0042] When the time matching relationship, direction matching relationship, and attenuation matching relationship are all satisfied, the processor determines that the associated impact event and the impact candidate event satisfy the front and rear wheel sequence impact consistency, and improves the road impact confidence of the impact candidate event.

[0043] In some optional implementations of this embodiment, the active head movement false alarm suppression includes: When there is a continuous change in yaw, pitch, or roll attitude within the time window corresponding to the impact candidate event, and the change in yaw, pitch, or roll attitude is monotonically changing or the direction of change is continuous and consistent within a preset time, the retention weight of the impact candidate event as a road impact is reduced.

[0044] Specifically, the processor calculates the changes in the user's head yaw, pitch, and roll attitudes based on inertial measurement data. Yaw attitude changes typically correspond to the user turning their head left or right, pitch attitude changes typically correspond to the user tilting their head up or down, and roll attitude changes typically correspond to the user tilting their head or adjusting their body's center of gravity. When there are continuous changes in yaw, pitch, or roll attitudes within the time window corresponding to the impact candidate event, and these attitude changes are monotonically changing or have a continuous and consistent direction of change within a preset time, the processor considers the signal to be more likely to originate from the user's active head movement rather than a short-term road impact.

[0045] For example, when a user looks back at a vehicle behind them, the head yaw angle will continuously change in the same direction for a period of time, and acceleration peaks may appear in the inertial measurement data. If only the acceleration peaks are used for judgment, this action can easily be misjudged as a road impact. In this embodiment, the processor will identify that the yaw attitude change has the characteristics of being continuous and directional, and thereby reduce the road impact confidence of the impact candidate event. Similarly, when a user looks down at the handlebars or looks up at the road conditions in the distance, the pitch attitude change usually also has the characteristics of continuous change, and the processor can also reduce the road impact confidence of the corresponding impact candidate event.

[0046] In this way, the processor can use the head posture change information naturally collected by the head-mounted augmented reality glasses to suppress false impacts caused by active head movements, thereby reducing the misjudgment of turning the head, looking down, looking up, or tilting the head as road impacts.

[0047] In some optional implementations of this embodiment, the suppression of false alarms regarding loosening of the garment includes: The high-frequency vibration energy, impact direction stability, and main peak tail vibration characteristics of the impact candidate event are determined based on the inertial measurement data. When the deviation of the high-frequency vibration energy from the wearing stability benchmark is higher than the wearing vibration threshold, the impact direction stability is lower than the direction stability threshold, and the main vibration frequency of the main peak trailing vibration characteristic remains in a fixed frequency band within a continuous window, and the main vibration frequency does not change accordingly with the change of the current riding speed, the weight of the impact candidate event as a road impact is reduced.

[0048] Specifically, the processor determines the high-frequency vibration energy, impact direction stability, and tail vibration characteristics after the main peak of the impact candidate event based on inertial measurement data. The high-frequency vibration energy is used to characterize the strength of the high-frequency vibration components in the impact candidate event, the impact direction stability is used to characterize whether the main impact direction is concentrated within a continuous window, and the tail vibration characteristics after the main peak are used to characterize whether there is continuous vibration after the impact on the main peak.

[0049] The processor compares the high-frequency vibration energy of the impact candidate event with a wearing stability benchmark. When the deviation of the high-frequency vibration energy from the wearing stability benchmark is higher than the wearing vibration threshold, it indicates that there may be abnormal wearing vibration in the current impact candidate event. The processor further judges the impact direction stability and the tail vibration characteristics after the main peak. If the impact direction stability is lower than the direction stability threshold, it means that the impact direction is dispersed within a continuous window and it is difficult to correspond to a clear road disturbance direction; if the main vibration frequency of the tail vibration characteristics after the main peak remains in a fixed frequency band within a continuous window, and the main vibration frequency does not change accordingly with the current riding speed, it means that the vibration is more likely to originate from loose glasses or inherent frame vibration.

[0050] Real-world continuous road surface disturbances are typically affected by the current cycling speed. As cycling speed changes, the time interval between adjacent road surface disturbances passing the wheel and the corresponding vibration frequency also change accordingly. In contrast, loose eyeglasses or frame elastic vibrations often exhibit relatively fixed high-frequency vibration characteristics, with their dominant vibration frequency not significantly changing with cycling speed. Therefore, this implementation method uses high-frequency vibration energy, directional stability, trailing vibration characteristics, and speed correlation to jointly determine looseness, which can more accurately distinguish between loose eyeglasses and real-world road surface impacts than simply setting a high-frequency threshold.

[0051] When the above conditions are met, the processor reduces the road impact confidence of the impact candidate event. The processor can also output a low-priority prompt to remind the user to adjust their glasses when the current riding impact risk is low and the visual load on the display area is low, but this prompt is not considered a road impact risk warning.

[0052] In some optional implementations of this embodiment, determining the cycling impact risk level based on the road impact confidence level includes: When the road impact confidence level reaches the effective impact threshold, the impact candidate event is determined to be an effective impact event; Within the observation window following the effective impact event, the cycling impact risk level is determined based on the decrease in cycling speed, head posture recovery time, and continuous impact density.

[0053] Specifically, when the road impact confidence level reaches the effective impact threshold, the processor determines the impact candidate event as a valid impact event. The effective impact threshold can be determined based on vehicle profiles, head-mounted vibration benchmarks, historical riding data, or user-defined safety preferences. For impact candidate events whose road impact confidence level does not reach the effective impact threshold, the processor may simply record the event without triggering a risk alert.

[0054] After a valid impact event is identified, the processor enters the observation window. This window can be set for a short period after the valid impact event to observe its subsequent impact on the riding state. Within the observation window, the processor acquires the rate of decrease in riding speed, head posture recovery time, and consecutive impact density. The rate of decrease in riding speed reflects whether the user slowed down or braked suddenly due to the road impact; the head posture recovery time reflects whether the user's head quickly returned to a stable riding state after the impact; and the consecutive impact density reflects whether multiple valid impact events occurred within a short period.

[0055] When the speed drop after a valid impact event is small, the head posture recovery time is short, and the density of consecutive impacts is low, the processor can determine a low level of riding impact risk. In this case, even if a real road impact is detected, the impact has a small impact on riding safety, and only recording or a slight warning can be issued. When the speed drop after a valid impact event is large, the head posture recovery time is long, or the density of consecutive impacts is high, the processor can determine a high level of riding impact risk. In this case, it indicates that the road impact may have already affected the user's stable riding, requiring a more obvious but still limited low-interference warning.

[0056] In some optional implementations of this embodiment, the step of outputting a low-interference reminder corresponding to the cycling impact risk level in the display area of ​​the head-mounted augmented reality glasses includes: The current visual load level is determined based on the current riding speed, slope, head posture stability, and the current occupancy status of the display area. Based on the cycling impact risk level and the current visual load level, determine the edge display position, display area, and continuous display time of the low-interference reminder in the display area; When the current visual load level is higher than the preset load level and the cycling impact risk level does not reach the strong reminder level, the low interference reminder will be downgraded to an edge reminder or simply recorded. During the same continuous impact process, if the risk level of the cycling impact does not increase, the repeated output of the low-interference reminder is restricted within a preset freeze time.

[0057] Specifically, the processor determines the current visual load level based on the current cycling speed, gradient, head posture stability, and the current occupancy of the display area. Higher cycling speeds typically require greater attention to the road ahead; gradient indicates whether the current road segment is downhill, uphill, or has a changing gradient; downhill or steep gradient segments generally require higher levels of focus; head posture stability indicates whether the user is in a stable observation state; and the current occupancy of the display area indicates whether navigation information, incoming call information, activity data, or other notifications have already occupied the display area.

[0058] The processor determines the display position, area, and duration of low-interference alerts within the display area based on the cycling impact risk level and the current visual load level. The display position is preferentially chosen at the edge of the display area or in a location that does not obstruct the view of the center of the road. The display area is limited according to the cycling impact risk level and the current visual load level; a smaller display area is used for lower risk levels, while a larger display area can be used for higher risk levels, but it must still avoid obstructing the user's main observation area. The duration of the display is also controlled based on the cycling impact risk level and the current visual load level; low-risk alerts can be displayed briefly and then disappear automatically, while high-risk alerts can have a longer display time, but still have a maximum limit.

[0059] When the current visual load level is higher than the preset load level, and the cycling impact risk level does not reach the strong alert level, the processor will downgrade the low-interference alert to an edge alert or simply record it. Edge alerts can take the form of short-term brightness changes at the edge of the display area, short-term color bars, or small icon prompts. When the cycling impact risk level reaches the strong alert level, even if the current visual load level is high, the processor can output brief and limited alert content, but still control its display position, display area, and duration.

[0060] During the same series of impacts, if the risk level of the cycling impact does not increase, the processor limits the repeated output of low-interference alerts within a preset freeze time. The same series of impacts can be determined based on the time interval between consecutive impacts, the road impact confidence level, and the continuity of the cycling position. By setting a preset freeze time, the user can avoid repeatedly receiving the same level of alerts on the same road segment or during the same series of bumps, reducing the interference of frequent flashing or repeated display of alerts on the rider's vision and attention.

[0061] In some implementations, the processor can also set a limit on repeated reminders. During the same continuous impact process, if the risk level of the cycling impact does not increase, the processor can refrain from repeatedly outputting strong reminders for a preset time to avoid frequent flashing of the reminder information affecting the user's observation. The processor can also be set to de-hysteresis, de-displaying the reminder only after the cycling impact risk level has decreased and remained so for a period of time, to prevent the reminder information from frequently appearing and disappearing in a critical state.

[0062] In this way, the low-interference reminder does not simply pop up a message after a risk is detected. Instead, it combines the cycling impact risk level and the current visual load level to schedule the display position, display area, display time and reminder intensity, thereby reducing obstruction of the cyclist's line of sight while alerting them to the cycling impact risk.

[0063] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0064] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0065] Further reference Figure 2 As a response to the above Figure 1 The implementation of the method shown in this application provides an embodiment of an augmented reality glasses cycling impact risk recognition and reminder system. This system embodiment is similar to... Figure 1 Corresponding to the method embodiments shown, the system can be specifically applied to various electronic devices.

[0066] like Figure 2 As shown, the augmented reality glasses cycling impact risk recognition and reminder system 200 described in this embodiment includes: a data acquisition module 201, an extraction module 202, a first determination module 203, a second determination module 204, a third determination module 205, and an output module 206. Wherein: The acquisition module 201 is used to acquire inertial measurement data from the head-mounted augmented reality glasses and obtain the current cycling speed; Extraction module 202 is used to extract impact candidate events based on the inertial measurement data, and to perform active head movement false alarm suppression and wearing loosening false alarm suppression on the impact candidate events according to head posture change characteristics and wearing vibration characteristics; The first determining module 203 is used to determine the subsequent impact expectation window corresponding to the impact candidate event after false alarm suppression based on the current riding speed and the front and rear wheel spacing in the vehicle profile; The second determining module 204 is used to detect associated impact events within the subsequent impact expectation window, and determine the road impact confidence of the impact candidate event based on the time matching relationship, direction matching relationship and attenuation matching relationship between the impact candidate event and the associated impact event; The third determining module 205 is used to determine the cycling impact risk level based on the road impact confidence level; The output module 206 is used to output a low-interference reminder corresponding to the cycling impact risk level in the display area of ​​the head-mounted augmented reality glasses.

[0067] The augmented reality glasses cycling impact risk identification and reminder system provided in this embodiment of the invention can realize all the processes of the augmented reality glasses cycling impact risk identification and reminder method in the above embodiments. The functions and technical effects of each module in the device are the same as those of the augmented reality glasses cycling impact risk identification and reminder method in the above embodiments, and will not be repeated here.

[0068] The above are merely preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for recognizing and alerting the impact risk of cycling using augmented reality glasses, characterized in that, include: Collect inertial measurement data from the head-mounted augmented reality glasses and obtain the current cycling speed; Impact candidate events are extracted based on the inertial measurement data, and active head movement false alarm suppression and loosening false alarm suppression are performed on the impact candidate events according to the head posture change characteristics and wearing vibration characteristics. Based on the current riding speed and the front and rear wheel spacing in the vehicle profile, determine the subsequent impact expectation window corresponding to the impact candidate event after false alarm suppression; Within the subsequent impact expectation window, associated impact events are detected, and the road impact confidence of the impact candidate events is determined based on the time matching relationship, direction matching relationship, and attenuation matching relationship between the impact candidate events and the associated impact events. The cycling impact risk level is determined based on the road impact confidence level. Based on the cycling impact risk level, a low-interference reminder corresponding to the cycling impact risk level is output in the display area of ​​the head-mounted augmented reality glasses.

2. The method according to claim 1, characterized in that, Before extracting impact candidate events based on the inertial measurement data, the method further includes: When the current cycling speed is within the cycling speed range and the inertial measurement data indicates that the change in the user's head posture is less than the posture stability threshold, a stable cycling time period is determined. Based on the inertial measurement data during the stable riding period, the headband vibration reference and wearing stability reference are determined.

3. The method according to claim 2, characterized in that, The extraction of impact candidate events based on the inertial measurement data includes: The inertial measurement data is subjected to gravity removal processing to obtain the linear acceleration of the head; The linear acceleration of the head is filtered or decomposed into frequency bands to obtain the road impact-related components. The signal segments that satisfy the impact intensity and duration conditions are extracted from the road impact correlation components, and the signal segments are identified as the impact candidate events.

4. The method according to claim 1, characterized in that, The step of determining the subsequent impact expectation window corresponding to the impact candidate event after false alarm suppression, based on the current riding speed and the front and rear wheel spacing in the vehicle profile, includes: Based on the front and rear wheel spacing and the current riding speed, determine the expected time interval between the front wheel impact and the rear wheel impact; Based on the fluctuation range of the current riding speed and the impact transmission error range corresponding to the vehicle profile, the expected time interval is extended to obtain the subsequent impact expectation window.

5. The method according to claim 4, characterized in that, The step of determining whether the associated impact event and the impact candidate event satisfy the time matching relationship, direction matching relationship, and attenuation matching relationship includes: When the time interval between the associated impact event and the impact candidate event falls within the subsequent impact expectation window, and the difference between the main impact direction of the associated impact event and the impact candidate event in the riding reference coordinate system is less than the direction matching threshold, and the difference between the energy attenuation after the main peak of the associated impact event and the impact candidate event is less than the attenuation matching threshold, it is determined that the associated impact event and the impact candidate event satisfy the time matching relationship, the direction matching relationship and the attenuation matching relationship.

6. The method according to claim 1, characterized in that, The active head movement false alarm suppression includes: When there is a continuous change in yaw, pitch, or roll attitude within the time window corresponding to the impact candidate event, and the change in yaw, pitch, or roll attitude is monotonically changing or the direction of change is continuous and consistent within a preset time, the retention weight of the impact candidate event as a road impact is reduced.

7. The method according to claim 2, characterized in that, The suppression of false alarms regarding looseness of the device includes: The high-frequency vibration energy, impact direction stability, and main peak tail vibration characteristics of the impact candidate event are determined based on the inertial measurement data. When the deviation of the high-frequency vibration energy from the wearing stability benchmark is higher than the wearing vibration threshold, the impact direction stability is lower than the direction stability threshold, and the main vibration frequency of the main peak trailing vibration characteristic remains in a fixed frequency band within a continuous window, and the main vibration frequency does not change accordingly with the change of the current riding speed, the weight of the impact candidate event as a road impact is reduced.

8. The method according to claim 1, characterized in that, The determination of the cycling impact risk level based on the road impact confidence level includes: When the road impact confidence level reaches the effective impact threshold, the impact candidate event is determined to be an effective impact event; Within the observation window following the effective impact event, the cycling impact risk level is determined based on the decrease in cycling speed, head posture recovery time, and continuous impact density.

9. The method according to claim 8, characterized in that, The step of outputting a low-interference alert corresponding to the cycling impact risk level in the display area of ​​the head-mounted augmented reality glasses includes: The current visual load level is determined based on the current riding speed, slope, head posture stability, and the current occupancy status of the display area. Based on the cycling impact risk level and the current visual load level, determine the edge display position, display area, and continuous display time of the low-interference reminder in the display area; When the current visual load level is higher than the preset load level and the cycling impact risk level does not reach the strong reminder level, the low interference reminder will be downgraded to an edge reminder or simply recorded. During the same continuous impact process, if the risk level of the cycling impact does not increase, the repeated output of the low-interference reminder is restricted within a preset freeze time.

10. An augmented reality glasses-based system for identifying and alerting the impact risk of cycling, characterized in that, include: The data acquisition module is used to collect inertial measurement data from the head-mounted augmented reality glasses and obtain the current cycling speed; The extraction module is used to extract impact candidate events based on the inertial measurement data, and to perform active head movement false alarm suppression and wearing loosening false alarm suppression on the impact candidate events according to the head posture change characteristics and wearing vibration characteristics. The first determining module is used to determine the subsequent impact expectation window corresponding to the impact candidate event after false alarm suppression, based on the current riding speed and the front and rear wheel spacing in the vehicle profile; The second determining module is used to detect associated impact events within the subsequent impact expectation window, and determine the road impact confidence of the impact candidate events based on the time matching relationship, direction matching relationship and attenuation matching relationship between the impact candidate events and the associated impact events; The third determining module is used to determine the cycling impact risk level based on the road impact confidence level. The output module is used to output a low-interference reminder corresponding to the cycling impact risk level in the display area of ​​the head-mounted augmented reality glasses.