Vehicle key communication parameter calibration method and vehicle-mounted digital key system

Abstract

The invention discloses a vehicle key communication parameter calibration method and a vehicle-mounted digital key system, relates to the technical field of intelligent automobile electronics, is used for realizing dynamic calibration of parameters, and provides the vehicle key communication parameter calibration method aiming at extra cost and power consumption caused by the fact that a traditional scheme depends on image acquisition equipment to recognize a scene. The scene where the vehicle is located is recognized by multiplexing the existing BLE module and UWB module on the vehicle. According to the method, no extra hardware equipment needs to be added, and extra implementation cost and implementation difficulty cannot be brought. And moreover, the power consumption of the BLE module and the UWB module is lower than that of image acquisition equipment such as a camera, so that the cruising ability of a vehicle battery can be effectively ensured. Besides, two different signals and two different features are used during scene recognition, and multi-dimensional and multi-breadth feature recognition is realized, so that the scene recognition precision is higher, and the method can adapt to the recognition requirements of more complex and more subdivided scenes.

Description

Technical Field

[0001] This application relates to the field of intelligent vehicle electronic technology, and in particular to a method for calibrating vehicle key communication parameters and an on-board digital key system. Background Technology

[0002] Currently, vehicle digital key systems generally rely on Bluetooth Low Energy (BLE) technology to enable communication between the vehicle and key devices such as mobile phones and smartwatches. The calibration parameters of digital key systems, such as transmission power, signal threshold, and positioning algorithm, are usually fixed preset values. However, in real-world applications, vehicles may be in various complex environments, such as open spaces, semi-open spaces, and densely populated areas. Fixed calibration parameters cannot be applied to different scenarios, leading to problems such as increased positioning errors, response delays, or false triggering.

[0003] In related technologies, the current scene is typically identified using image acquisition devices such as cameras, and then calibration parameters are configured by selecting appropriate fixed value combinations based on different scene conditions to solve the aforementioned problems. However, this approach relies on additional hardware, and image acquisition devices such as cameras are expensive and consume significant power, thus affecting the vehicle's battery life.

[0004] Therefore, those skilled in the art urgently need a vehicle key communication parameter calibration method to solve the problems caused by relying on image acquisition devices such as cameras to achieve scene recognition and complete parameter calibration. Summary of the Invention

[0005] The purpose of this application is to provide a method for calibrating vehicle key communication parameters and an in-vehicle digital key system, so as to realize the dynamic calibration of parameters of the in-vehicle digital key system.

[0006] To address the aforementioned technical issues, this application provides a vehicle key communication parameter calibration method, applicable to multiple anchor points deployed on a vehicle, each anchor point including: a low-power Bluetooth module and an ultra-wideband module;

[0007] The methods include:

[0008] Using any one of the anchor points as a simulated key end and the remaining anchor points as receiving ends, a test task is executed; wherein, the test task includes: controlling the simulated key end to simulate a remote control key, transmitting Bluetooth signals to each of the receiving ends through the low-power Bluetooth module to establish a Bluetooth connection, and sending pulse signals through the ultra-wideband module to perform ranging.

[0009] The signal characteristic data of the Bluetooth signal received by each of the receiving terminals and the ranging result data of the pulse signal are obtained.

[0010] The current scene is identified based on the signal feature data and the ranging result data, and a calibration parameter set corresponding to the current scene is matched.

[0011] The parameter calibration of the digital key system is completed by matching the calibration parameter set obtained.

[0012] In an optional embodiment, after performing the test task using any one of the anchor points as the simulated key end and the remaining anchor points as the receiving end, the method further includes:

[0013] Switch to the next anchor point as the new simulated key end, and the remaining anchor points as the new receiving ends, and execute the test task until all anchor points have executed the test task once as the simulated key end;

[0014] The acquisition of signal feature data of the Bluetooth signal received by each of the receiving terminals and ranging result data of the pulse signal includes:

[0015] When each anchor point is used as the simulated key terminal to perform the test task, the corresponding signal characteristic data and ranging result data received by each receiving terminal are obtained.

[0016] In one optional embodiment, the signal characteristic data of the Bluetooth signal includes: RSSI characteristic data;

[0017] The ranging result data of the pulse signal includes: time of flight data;

[0018] The current scene types include: open scene, dense scene, and semi-open scene.

[0019] Identifying the current scene based on the signal feature data and the ranging result data includes:

[0020] If all the conditions in the open space condition set are met, then the current scene is determined to be an open space scene;

[0021] The set of open conditions includes:

[0022] The mean of the RSSI feature data is within an open area, and the standard deviation is less than the first open area threshold; the time-of-flight data meets the preset first stability condition.

[0023] If all the conditions in the dense condition set are met, then the current scene is determined to be a dense scene;

[0024] The dense condition set includes:

[0025] The mean of the RSSI feature data is within the dense interval and the standard deviation is greater than the first dense threshold. The waveform of the RSSI feature data exhibits random jump characteristics. The mean of the time-of-flight data is greater than the second dense threshold and satisfies the second instability condition.

[0026] If all the conditions in the semi-open space condition set are met, then the current scene is determined to be a semi-open space scene.

[0027] The semi-empty condition set includes:

[0028] The mean of RSSI feature data is within the semi-open range, and the waveform shows a step-like decrease; the flight time data satisfies the first stability condition during the line-of-sight path and exhibits abrupt changes during the non-line-of-sight path.

[0029] The median values ​​of the open area, the semi-open area, and the dense area decrease sequentially.

[0030] In an optional embodiment, the scenario type of the current scenario further includes: a dynamic interference scenario;

[0031] Identifying the current scene based on the signal feature data and the ranging result data further includes:

[0032] If all the conditions in the dynamic interference condition set are met, then the current scenario is determined to be a dynamic interference scenario.

[0033] The dynamic interference condition set includes:

[0034] The fluctuation frequencies of both RSSI feature data and time-of-flight data are greater than the first dynamic interference threshold; the mean value of RSSI feature data is within the dynamic interference range, and the waveform exhibits sawtooth fluctuations and sudden drops with amplitudes greater than the third dynamic interference threshold.

[0035] In an optional embodiment, identifying the current scene based on the signal feature data and the ranging result data, and matching the calibration parameter set corresponding to the current scene includes:

[0036] Extract the data features from the signal feature data and the ranging result data, and construct the corresponding scene feature vector;

[0037] The scene feature vector is input into a pre-trained classification model, and the scene label output by the classification model is obtained.

[0038] The corresponding calibration parameter set is called from the pre-established dynamic parameter library according to the scene label;

[0039] The classification model is a machine learning model that performs supervised learning using a sample dataset, which includes multiple groups of sample data labeled with different scene types.

[0040] In an optional embodiment, after completing the parameter calibration of the digital key system through the calibration parameter set obtained by matching, the method further includes:

[0041] Obtain the Bluetooth connection establishment success rate and ranging error of each receiver;

[0042] The preset communication performance expectation standard is determined based on the Bluetooth connection establishment success rate and ranging error.

[0043] If not, then retrain the classification model or adjust the calibration parameter set.

[0044] In one alternative embodiment, retraining the classification model includes:

[0045] Acquire historical signal feature data, historical ranging result data, and corresponding historical scene labels from the current time to the last training time of the classification model;

[0046] Each set of historical signal feature data and historical ranging result data is used as a new sample data set, added to the sample dataset, and the retraining of the classification model is triggered; wherein, the historical scene label is used as the label for the corresponding sample data set.

[0047] In one alternative embodiment, adjusting the calibration parameter set includes:

[0048] Back up the original data of the calibration parameter set;

[0049] The parameters in the calibration parameter set are increased or decreased according to a preset ratio;

[0050] The parameter calibration of the digital key system is completed using the adjusted calibration parameter set.

[0051] Determine whether the Bluetooth connection establishment success rate and the ranging error have been improved;

[0052] If so, keep the adjustment direction unchanged and return to the step of increasing or decreasing the parameters in the calibration parameter set according to the preset ratio until the Bluetooth connection establishment success rate and the ranging error meet the expected communication performance standard;

[0053] If not, change the adjustment direction and return to the step of increasing or decreasing the parameters in the calibration parameter set according to the preset ratio;

[0054] If the Bluetooth connection establishment success rate and ranging error are not improved after the changes and adjustments, the calibration parameter set is restored using the backed-up original data.

[0055] To address the aforementioned technical issues, this application also provides an in-vehicle digital key system, comprising multiple anchor points deployed on the vehicle, each anchor point including: a low-power Bluetooth module and an ultra-wideband module;

[0056] Each of the anchor points is used to implement the steps of the vehicle key communication parameter calibration method as described above.

[0057] In an optional embodiment, the anchor points are connected via CANFD communication, whereby CANFD is used to transmit signal characteristic data and ranging result data.

[0058] The vehicle-mounted digital key system also includes: a host computer;

[0059] Among the anchor points, there is a main anchor point. The main anchor point is connected to the host computer via Bluetooth communication through the low-power Bluetooth module, and sends the summarized signal feature data and the ranging result data to the host computer.

[0060] The host computer is used to: store the signal feature data and the ranging result data, and train a classification model based on the signal feature data and the ranging result data.

[0061] This application provides a vehicle key communication parameter calibration method. It achieves scene recognition by reusing existing Bluetooth Low Energy (BLE) and Ultra Wideband (UWB) modules in the vehicle, and dynamically calibrates parameters in the vehicle digital key system based on the identified scene and matching the corresponding calibration parameter set. Specifically, this method simulates vehicle key communication through various anchor points in the vehicle, simulating any anchor point as a remote key to control it to transmit Bluetooth signals and pulse signals to other anchor points. Furthermore, transmitting Bluetooth signals to establish Bluetooth communication is an existing function of the BLE modules in each anchor point, and transmitting pulse signals for ranging is an existing function of the UWB modules in each anchor point. In other words, this method is entirely implemented using the existing functions of the original modules in the vehicle, without adding any additional hardware, thus avoiding additional implementation costs and power consumption. Furthermore, since the power consumption of BLE and UWB modules is lower than that of image acquisition devices such as cameras, and this method utilizes the signals emitted by the original functions of BLE and UWB modules and the ranging results to achieve scene recognition, it can be carried out on the basis of the vehicle's normal Bluetooth communication and ranging functions, further reducing additional power consumption and significantly improving the vehicle's battery life.

[0062] Furthermore, because the environment affects Bluetooth communication and UWB ranging differently depending on the vehicle's location, the received signals and ranging results will exhibit varying characteristics. Therefore, this method not only identifies the vehicle's current environment but also achieves multi-dimensional (multiple features: signal features and data features) and multi-breadth (multiple signals: Bluetooth signal and pulse signal) scene recognition by utilizing the signal characteristics of both signals and the data characteristics of the ranging results. Therefore, compared to scene recognition using only image information through image recognition technology, it offers higher accuracy, can identify more complex and diverse sub-scenes, and thus achieves better dynamic parameter calibration results.

[0063] The vehicle digital key system provided in this application corresponds to the above method and has the same effect. Attached Figure Description

[0064] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This invention provides a structural diagram of an in-vehicle digital key system;

[0066] Figure 2 A flowchart of a vehicle key communication parameter calibration method provided by the present invention;

[0067] Figure 3 A flowchart for polling and executing test tasks is provided by the present invention;

[0068] Figure 4 This is a schematic diagram of signal transmission for an in-vehicle digital key system provided by the present invention. Detailed Implementation

[0069] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0070] The core of this application is to provide a method for calibrating vehicle key communication parameters and an in-vehicle digital key system.

[0071] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0072] Currently, in the application of in-vehicle digital key systems, the vehicle typically acquires image information of the external environment through image acquisition devices such as cameras, and then uses image recognition technology to identify the current scene in which the vehicle is located. Based on the identified scene, corresponding fixed values ​​are selected to complete the parameter calibration of the in-vehicle digital key system. However, the above solution requires additional image acquisition equipment, which brings additional implementation costs. Furthermore, the high power consumption of image acquisition devices such as cameras can adversely affect the vehicle's battery life.

[0073] To address the aforementioned problems, this application provides a method for calibrating vehicle key communication parameters. For example... Figure 1 As shown, this method is applied to multiple anchor points (Anchor-0 to Anchor-7) deployed on the vehicle. Each anchor point includes a Bluetooth Low Energy (BLE) module and an Ultra Wide Band (UWB) module. It should be noted that this method is not limited to the specific number of anchor points; as long as there are more than one, it is acceptable. Figure 1 The example shown is of 8 anchor points deployed on the vehicle side. In real-world scenarios, vehicles are typically deployed with 2-6 anchor points.

[0074] The specific process of the method is as follows: Figure 2 As shown, the steps include:

[0075] S11: Use any anchor point as the simulated key end and the remaining anchor points as the receiving ends to execute the test task.

[0076] The test tasks include: controlling the simulated key terminal to simulate a remote control key, transmitting Bluetooth signals to each receiver terminal through the low-power Bluetooth module to establish a Bluetooth connection, and sending pulse signals through the ultra-wideband module to perform distance measurement.

[0077] S12: Acquire the signal characteristic data of the Bluetooth signal received by each receiver and the ranging result data of the pulse signal.

[0078] S13: Identify the current scene based on signal feature data and ranging result data, and match the calibration parameter set corresponding to the current scene.

[0079] S14: Complete the parameter calibration of the digital key system by matching the calibration parameter set.

[0080] In traditional vehicle digital key systems, anchor points deployed on the vehicle can communicate with other devices (such as remote keys) via Bluetooth signals sent through their BLE modules, and can perform ranging (locating the remote key) by sending pulse signals sent through their UWB modules. Based on this, as explained above, this method simulates one anchor point as the remote key, enabling Bluetooth communication and ranging / location between the remote key and the anchor point (i.e., transmitting and receiving Bluetooth and pulse signals). When the vehicle is in different scenarios (such as open, dense, or semi-open environments), the environment affects the Bluetooth and pulse signals differently, resulting in different signal characteristics and ranging results. Based on this, the current vehicle scenario can be identified based on signal characteristics and ranging results, allowing for the matching of the correct calibration parameter set to complete the dynamic calibration of the vehicle digital key system's parameters.

[0081] It should be noted that there are no restrictions on the selection of the anchor point as the simulated key in step S11; any anchor point deployed on the vehicle can be selected. Furthermore, using the anchor point as the simulated key means simulating the remote key and completing the interaction between the remote key and the anchor point. Or, more directly, the anchor point as the simulated key sends Bluetooth and pulse signals to other anchor points to simulate the Bluetooth communication and UWB positioning tasks originally performed between the remote key and the vehicle anchor point in the in-vehicle digital key system.

[0082] For step S12, this embodiment is not limited to the specific category of the scene that can be identified. As long as the specific characteristics of the Bluetooth signal, pulse signal, and ranging result corresponding to the scene can be clearly identified, any scene can be added to the range of scene types that can be identified in this step. Optionally, in the dynamic calibration of the parameters of the vehicle digital key system, the scene types that can be identified may include, but are not limited to: open scene, dense scene, and semi-open scene.

[0083] Furthermore, for the various scenario types described above, this embodiment also provides corresponding examples of signal and ranging result data characteristics. However, it should be noted that the specific parameters in the following examples are only reference values, and the parameter values ​​should be adapted to different actual or test scenarios. However, under different scenarios, the magnitude relationship between parameter values ​​reflecting the same characteristic, the stability of a certain characteristic parameter, and the waveform characteristics presented by certain data remain largely unchanged.

[0084] A. Open Scene: No large obstacles; signal propagation is primarily along the line-of-sight (LOS) path, with ground reflection present. Typical scenario: A vehicle parked in the center of an outdoor lawn; a user walks towards the vehicle holding their phone, such as in an open-air parking lot or on a suburban road.

[0085] A-1) Key characteristics: The proportion of LOS is greater than a certain value (e.g., >90%), at which point the main reflector of the signal is the ground (i.e., a single reflecting intersection).

[0086] (A-2) For Bluetooth signals, the Received Signal Strength Indication (RSSI) can reflect signal characteristics. The BLE RSSI characteristics are analyzed as follows: A high mean RSSI (e.g., -65±5dBm) and small RSSI fluctuations (e.g., standard deviation <3dBm). In this case, the RSSI exhibits a typical waveform of periodic oscillations (caused by ground reflection).

[0087] (A-3) For pulse signals, ranging results are generally calculated using Time of Flight (ToF), so the ToF value of the pulse signal can be used as ranging result data. UWB ToF characteristic analysis is as follows: ToF value is stable.

[0088] A-4) Ranging Influence: Under LOS conditions, ranging accuracy is high because the pulse signal mainly travels along a direct path, resulting in weak multipath effects (primarily ground reflection). However, ground transmission may cause dual-path effects, leading to ranging fluctuations at specific distances (critical distances).

[0089] A-5) Typical performance of the ranging value (calculated from the ToF value): The ranging value is stable and the error is small.

[0090] B. Dense Scenes: Dense obstacles lead to a significant multipath effect, resulting in a high proportion of non-line-of-sight (NLOS) paths. Typical scenarios include vehicles parked in underground garages surrounded by concrete pillars and metal pipes, such as underground parking garages or urban roads with tall buildings.

[0091] B-1) Key characteristics: NLOS percentage is greater than a certain value (e.g., >60%), and the number of multipaths is greater than or equal to a certain number (e.g., ≥4).

[0092] B-2) BLE RSSI characteristic analysis: low mean RSSI (e.g., -80±10dBm); drastic fluctuations (e.g., standard deviation >8dBm); typical waveform: random jumps (generated by multipath superposition).

[0093] B-3) UWB ToF Feature Analysis: The ToF value is too large (caused by the extension of the NLOS path).

[0094] B-4) Ranging Impact: With a high proportion of NLOS (Normally In-Loop) signals, the signal propagates through reflection and diffraction, resulting in a longer path (the measured distance is greater than the actual distance). Furthermore, it exhibits large fluctuations and severe multipath effects, potentially capturing non-direct paths and causing ranging jumps.

[0095] B-5) Typical characteristics: The distance measurement value (positive deviation) is too large, the error can reach several meters, and the stability is poor.

[0096] C. Semi-open scene: A scene that combines LOS and NLOS, with partial obstruction but a relatively open overall field of view. Typical scenario: A vehicle is parked under a tree, the tree trunk partially obstructs the signal, but there are no dense obstacles around it, such as a residential parking lot or a park parking lot.

[0097] C-1) Key feature: The proportion of LOS paths is within a certain range (e.g., 40%~70%). This range indicates that the proportion of LOS paths in a semi-open scene is less than that in an open scene. In this scene, signal reflectors include trees, low shrubs, etc.

[0098] C-2) BLE RSSI Characteristic Analysis: The mean RSSI value is moderate (e.g., -72±6dBm); the RSSI value exhibits intermittent sudden drops (e.g., a sudden drop of 10~15dBm when shading occurs). Typical waveform: Step-like decrease (e.g., a user walking into the shade).

[0099] C-3) UWB ToF feature analysis: ToF values ​​are segmentally stable (i.e., stable during LOS period, but exhibiting jumps during occlusion period).

[0100] C-4) Distance Measurement Effects: When LOS and NLOS are mixed, the distance measurement value will suddenly increase when there are obstructions such as trees (i.e., entering NLOS), and will return to normal after the obstruction disappears (i.e., entering LOS). The overall distance measurement accuracy depends on the proportion of the LOS path; the higher the proportion of LOS, the higher the overall distance measurement accuracy.

[0101] C-5) Typical performance: The ranging value is accurate at LOS, jumps at NLOS, and the overall error is moderate (e.g., ±0.5m~1m).

[0102] Based on the data characteristics of the signals and ranging results in the three scenarios provided above, this embodiment also provides a specific implementation plan for step S13:

[0103] The signal characteristic data of the Bluetooth signal includes: RSSI characteristic data; the ranging result data of the pulse signal includes: time of flight data; the scene type of the current scene includes: open scene, dense scene and semi-open scene.

[0104] Then step S13 includes:

[0105] S13-A: If all conditions in the open space condition set are met, then the current scene is determined to be an open space scene.

[0106] The set of open conditions includes:

[0107] The mean of the RSSI feature data is within the open range (such as -65±5dBm in the example above), and the standard deviation is less than the first open range threshold (such as 3dBm in the example above); the time-of-flight data meets the preset first stability condition (determined based on the actual scenario's requirement and standard for stable ToF values).

[0108] S13-B: If all conditions in the dense condition set are met, then the current scene is determined to be a dense scene.

[0109] The dense condition set includes:

[0110] The mean of the RSSI feature data is within the dense interval (such as -80±10dBm in the example above), and the standard deviation is greater than the first dense threshold (such as 8dBm in the example above). The waveform of the RSSI feature data exhibits random jump characteristics. The mean of the time-of-flight data is greater than the second dense threshold (the ToF value is larger in dense scenarios than in other scenarios. The critical threshold can be determined based on the characteristics of the ToF value in each scenario to make distinctions), and it meets the second instability condition (determined according to the requirements and standards for the instability of the ToF value in the actual scenario).

[0111] S13-C: If all conditions in the semi-open space condition set are met, then the current scene is determined to be a semi-open space scene.

[0112] The semi-empty condition set includes:

[0113] The mean of the RSSI characteristic data is within the semi-open range (such as -72±6dBm in the example above), and the waveform shows a step-like decrease (the RSSI value drops sharply by 10~15dBm when the obstruction occurs in the example above); the time-of-flight data meets the first stability condition during the line-of-sight path and shows a jump during the non-line-of-sight path.

[0114] The median values ​​for the open, semi-open, and dense areas decrease sequentially (i.e., as in the example above, the median value for the open area is -65dBm, the median value for the semi-open area is -72dBm, and the median value for the dense area is -80dBm).

[0115] Based on this embodiment, scene recognition can be achieved by utilizing the signal characteristics of both the Bluetooth signal emitted by the BLE module and the pulse signal emitted by the UWB module, as well as the ranging result data characteristics, across different dimensions and breadths of features. Compared to related technologies that rely solely on image information for scene recognition, this method offers higher recognition accuracy and interference resistance, and can adapt to the recognition needs of more complex scenes.

[0116] In summary, the vehicle key communication parameter calibration method provided in this application achieves the identification of the vehicle's environment by reusing the existing BLE and UWB modules on the vehicle. Based on the identified environment, it matches the corresponding calibration parameter set to complete the parameter calibration of the vehicle digital key system. This method requires no additional hardware, thus avoiding additional implementation costs and difficulties. Furthermore, when using the BLE and UWB modules for scene recognition, this method utilizes their existing functions, and the power consumption of the BLE and UWB modules is significantly lower than that of image acquisition devices such as cameras, effectively improving the vehicle's battery life. In addition, this method uses two different signals and two different features—signal characteristics and ranging results—as the discrimination criteria during scene recognition, achieving multi-dimensional and broad feature recognition. This results in higher scene recognition accuracy and can adapt to the recognition needs of more complex and subdivided scenes.

[0117] On the other hand, as can be seen from the above embodiments, this method simulates any anchor point as a remote control key and uses other anchor points as receiving ends to perform a test task, in order to obtain the signal characteristics of Bluetooth signals and pulse signals transmitted and received between anchor points, as well as ranging result data. Figure 1 It is easy to see that, due to the different placement positions of the anchor points on the vehicle end, and the varying relative positions of the obstructions and the vehicle in different scenarios (even in open scenes, the ground acts as an obstruction, and whether the anchor point is placed closer to the roof or the bottom of the vehicle will result in different features), this embodiment provides a further implementation scheme to minimize or even eliminate the impact of the anchor point placement position on scene recognition and further improve the accuracy of scene recognition. After step S11, the above method further includes:

[0118] S15: Switch to the next anchor point as the new simulated key end, and the remaining anchor points as the new receiving ends, and execute the test task until all anchor points have executed the test task once as simulated key ends.

[0119] At this point, step S12 specifically becomes:

[0120] When each anchor point is used as a simulated key to perform a test task, the signal characteristic data and ranging result data received by each receiving end are obtained.

[0121] That is to say Figure 3 As shown, this embodiment no longer executes test tasks using only a combination of FOB and receiver anchor points. Instead, it employs a polling test method, ensuring that each anchor point acts as an FOB point and executes a test task once. Furthermore, in step S12, scene recognition is performed using the data features obtained after all test tasks have been executed. This minimizes the impact of anchor point placement on data features, thereby improving the accuracy and reliability of scene recognition.

[0122] On the other hand, as described in the above embodiments, this method achieves scene recognition by combining the signal characteristics of both Bluetooth and pulse signals, as well as the data characteristics of the ranging results from the UWB module's positioning and ranging. Therefore, compared to using only one type of image information for scene recognition, it has higher recognition accuracy and can meet the needs of recognizing more complex scenes.

[0123] In response to this, this embodiment, based on the three scene types provided in the above embodiments, also provides a more complex scene type and a corresponding recognition scheme:

[0124] The current scenario type also includes: dynamic interference scenario. That is:

[0125] D. Dynamic Interference Scenarios: These are scenarios where there are moving interference sources (such as pedestrians, vehicles, etc.) in the environment surrounding the vehicle. The characteristic of this scenario is that the signal characteristics change rapidly over time. Typical scenarios include: temporary parking on the roadside, intermittent obstruction caused by vehicles traveling in adjacent lanes, etc.

[0126] D-1) Key characteristics: RSSI / ToF fluctuation frequency is greater than a certain value (e.g., >1Hz), and the moving speed of the interference source is greater than or equal to a certain value (e.g., ≥5km / h).

[0127] D-2) BLE RSSI Characteristic Analysis: RSSI exhibits high-frequency jitter (e.g., fluctuation frequency > 1Hz), and the mean value is similar to that of a semi-open scene (e.g., -72±6dBm). This scene exhibits pulse-like interference (e.g., a sudden -20dBm drop). Typical waveform: sawtooth fluctuations (due to momentary obstruction caused by passing vehicles).

[0128] D-3) UWB ToF Feature Analysis: ToF value fluctuates at high frequencies (e.g., fluctuation frequency > 1Hz).

[0129] D-4) Ranging effect: Moving objects (such as pedestrians and vehicles) cause the signal path to change over time, which may momentarily block the direct path (ranging value jump), and the ranging value exhibits high-frequency jitter (such as jitter frequency > 1Hz).

[0130] D-5) Typical characteristics: The ranging value fluctuates drastically, with a large error range (from centimeters to meters).

[0131] Then step S13 also includes:

[0132] S13-D: If all conditions in the dynamic interference condition set are met, then the current scenario is determined to be a dynamic interference scenario;

[0133] The dynamic interference condition set includes:

[0134] The fluctuation frequency of both RSSI feature data and time-of-flight data is greater than the first dynamic interference threshold (such as 1Hz in the example above); the mean value of RSSI feature data is within the dynamic interference range (such as -72±6dBm in the example above), and the waveform exhibits sawtooth fluctuations and sudden drops with amplitudes greater than the third dynamic interference threshold (such as 20dBm in the example above) (sudden -20dBm drop in the example above).

[0135] Therefore, this embodiment, based on the higher scene recognition accuracy of this method, adds new scene types for accurate identification. This allows the vehicle remote key system using this method to adapt to more subdivided scenarios and adaptively complete dynamic parameter calibration, resulting in better communication quality and positioning accuracy.

[0136] On the other hand, the above embodiments provide typical characteristics of the signals and ranging results corresponding to each scenario type. If the vehicle obtains signal characteristics and ranging result data characteristics after performing a test task, the scenario type in which the vehicle is currently located can be identified based on this. However, the above embodiments do not strictly limit how the data characteristics obtained by the vehicle are matched to the typical characteristics under different scenario types. The simplest approach is to compare each data feature item by item to see if it meets the condition set of different scenarios.

[0137] However, this approach may lack flexibility and have insufficient recognition efficiency. To address this issue, this application also provides a further implementation scheme for scene recognition in step S13. Step S13 further includes:

[0138] S131: Extract data features from signal feature data and ranging result data, and construct the corresponding scene feature vector.

[0139] S132: Input the scene feature vector into the pre-trained classification model and obtain the scene label output by the classification model.

[0140] S133: Call the corresponding calibration parameter set from the pre-established dynamic parameter library according to the scene label.

[0141] The classification model is a machine learning model that uses supervised learning through a sample dataset, which includes multiple groups of sample data labeled with different scene types.

[0142] As described above, this embodiment establishes a classification model using machine learning methods. After obtaining the signal features of the current vehicle and the distance measurement result data features, the features are extracted to construct a scene feature vector. The scene feature vector is input into the classification model, which can then classify the scene into the corresponding scene type and output the corresponding scene label to complete scene recognition. Finally, based on the identified scene label, the corresponding type of calibration parameter set is directly called from the pre-established dynamic parameter library to prepare for subsequent dynamic parameter calibration.

[0143] This method eliminates the need to compare each data feature item with the corresponding conditions in each condition set. Instead, the classification model uses machine learning technology to categorize data features and complete scene recognition, resulting in higher scene recognition efficiency and greater flexibility in implementation.

[0144] Furthermore, based on the above embodiment providing a scene recognition scheme using a pre-trained classification model, this embodiment also provides a further implementation scheme. After step S14, the method further includes:

[0145] S21: Obtain the Bluetooth connection establishment success rate and ranging error of each receiver;

[0146] S22: Determine whether the preset communication performance expectation standard is met based on the Bluetooth connection establishment success rate and ranging error; if not, proceed to step S23; if yes, exit this method.

[0147] S23: Retrain the classification model or adjust the calibration parameter set.

[0148] After parameter calibration, this embodiment further assesses the effectiveness of the calibration by evaluating the success rate of Bluetooth connection establishment via the BLE module and the accuracy of positioning and ranging via the UWB module. If the Bluetooth connection success rate is low or the ranging error is large after parameter calibration, it indicates that the identified scene type may be incorrect, or the calibration parameter set may be outdated and no longer applicable to the current situation. Therefore, this embodiment employs measures such as retraining the classification model and adjusting the calibration parameter set to optimize dynamic parameter calibration.

[0149] Furthermore, regarding how to retrain the aforementioned classification model, it can be done by replacing it with the latest sample dataset, or by using other methods to determine a new sample dataset for training. This embodiment does not impose any limitations on this. In an optional embodiment, this embodiment also provides a method for retraining a classification model, including:

[0150] S31: Obtain historical signal feature data, historical ranging result data, and corresponding historical scene labels from the current time to the last classification model training time.

[0151] S32: Add each set of corresponding historical signal feature data and historical ranging result data as a new sample data set to the sample dataset and trigger the retraining of the classification model.

[0152] Among them, historical scene labels serve as markers for the corresponding sample data groups.

[0153] In other words, this embodiment updates the sample dataset by adding all historical data (signal feature data, ranging result data, and historical scene labels output by the classification model) from the moment the classification model is retrained to the moment of the most recent training of the classification model as new sample data. Furthermore, since the training of the classification model is supervised, the historical data, in addition to the signal feature data and ranging result data reflecting scene characteristics, also includes the corresponding output historical scene labels. These labels can serve as markers for the signal feature data and ranging result data, eliminating the need for manual re-labeling and fulfilling the conditions for supervised learning, thus enabling the retraining of the classification model.

[0154] Similarly, the most direct way to adjust the calibration parameter set is for technicians to manually adjust the calibration parameter sets corresponding to each scenario based on the current needs, thus completing the update. This embodiment also provides another non-manual calibration parameter set adjustment scheme, which further includes:

[0155] S41: Backup the original data of the calibration parameter set.

[0156] S42: Increase or decrease the parameters in the calibration parameter set according to the preset ratio.

[0157] S43: Complete the parameter calibration of the digital key system using the adjusted calibration parameter set.

[0158] S44: Determine whether the Bluetooth connection establishment success rate and ranging error have been improved; if yes, proceed to step S45; if no, proceed to step S46.

[0159] S45: Keep the adjustment direction unchanged, return to the step of increasing or decreasing the parameters in the calibration parameter set according to the preset ratio, until the Bluetooth connection establishment success rate and ranging error meet the expected communication performance standards.

[0160] S46: Change the adjustment direction and return to the step of increasing or decreasing the parameters in the calibration parameter set according to the preset ratio.

[0161] S47: If the Bluetooth connection establishment success rate and ranging error are not improved after the adjustment, the calibration parameter set is restored by backing up the original data.

[0162] It should be noted that in step S42, when adjusting the parameters in the calibration parameter set, the adjustment direction for each parameter can be different. A default adjustment direction can be preset according to actual needs, thus determining the adjustment direction when the parameter is first adjusted in step S42. Furthermore, regarding the change in adjustment direction in step S46, that is, if the adjustment direction for a parameter in step S42 is to increase, then in step S46, that parameter will be decreased. Similarly, if the adjustment direction for a parameter in step S42 is to decrease, then that parameter will be increased in step S46.

[0163] Furthermore, as shown in steps S42 and S45, the calibration parameter set adjustment implemented by this method is a step adjustment, using a preset ratio as the step adjustment for each parameter. It should be noted that the preset ratio for each parameter can be different, determined according to actual needs. Based on this step adjustment scheme, once the adjustment direction is correct, the parameters can be gradually adjusted to appropriate values, achieving automatic optimization and adjustment of the calibration parameter set.

[0164] In the above embodiments, a method for calibrating vehicle key communication parameters has been described in detail. This application also provides an embodiment corresponding to an in-vehicle digital key system. Figure 1 As shown, this embodiment provides an in-vehicle digital key system, including: multiple anchor points (Anchor-0 to Anchor-7) deployed on the vehicle, each anchor point including: a low-power Bluetooth module and an ultra-wideband module.

[0165] Each anchor point is used to implement the steps of the vehicle key communication parameter calibration method provided in any of the above embodiments.

[0166] Since the embodiments of the system part correspond to the embodiments of the method part, please refer to the description of the embodiments of the method part for the embodiments of the system part, and they will not be repeated here.

[0167] However, furthermore, this embodiment also provides a further implementation scheme for the vehicle digital key system, such as... Figure 4 As shown:

[0168] The anchor points are connected via a Controller Area Network (CANFD, an upgraded version of the CAN bus) to transmit signal characteristic data and ranging result data.

[0169] The vehicle digital key system also includes: a host computer.

[0170] There is a main anchor point among the anchor points. The main anchor point is connected to the host computer via Bluetooth through a low-power Bluetooth module and sends the summarized signal characteristic data and ranging result data to the host computer.

[0171] The host computer is used to: store signal feature data and ranging result data, and to train a classification model based on the signal feature data and ranging result data.

[0172] It should be noted that this embodiment is not limited to the main anchor point being one of all anchor points deployed on the vehicle. In this embodiment, the main anchor point only serves to aggregate signal characteristic data and ranging results obtained from itself and other anchor points acting as receivers, and transmit this data to the host computer via Bluetooth communication using the BLE module. This function can be implemented by any anchor point, therefore no restrictions are placed on the selection of the main anchor point. In an optional embodiment, the main anchor point can be the FOB end of the last test task during the polling test process. After the last polling, the data from all anchor points is integrated and uploaded to the host computer. Or as... Figure 3 As shown, in another embodiment, the first anchor point (e.g.) is used based on the numbering order. Figure 1 and Figure 3 The Anchor-0 in the example serves as the main anchor point, but this embodiment does not impose any restrictions on this.

[0173] It should be noted that the host computer in this embodiment is used to store historical data and train the classification model. After the model training is complete, it can be deployed on the vehicle to achieve the parameter calibration expected by this method. Subsequent parameter calibrations will not require the use of the host computer (unless retraining is needed as in the above embodiment). Therefore, the host computer here can be other devices outside the vehicle, such as laptops, tablets, or other devices with certain data processing capabilities (at least sufficient to meet the needs of model training).

[0174] For model training, in practical applications, it can be performed before actual vehicle deployment. Using the previously collected dataset, offline analysis and machine learning methods are employed to train the model. Once the model training converges, a final model with fixed parameters is obtained, which is then imported into the product (BLE / UWB module) to complete model deployment. Subsequently, during real-vehicle operation, the model can receive real-time detected vehicle data and output the corresponding inference results (i.e., the parameters that need to be calibrated).

[0175] Furthermore, it should be noted that the communication method between the host computer and the main anchor point is not limited to Bluetooth communication implemented through the BLE module. This embodiment considers that the BLE module is already present in the vehicle and is inherently used for Bluetooth communication, and common devices that can serve as host computers, such as laptops and tablets, generally also have Bluetooth communication capabilities. Therefore, data transmission based on Bluetooth communication can be achieved without adding any additional communication equipment or cables, making it easier to implement. However, if the amount of historical data transmitted from the main anchor point to the host computer is found to be too large during model training, other methods can be used to achieve communication between the host computer and the main anchor point to reduce data loss. For example, wired communication methods such as CANFD in the above example are currently supported by all anchor points on the vehicle. It should also be noted that the specific implementation scenario of the host computer in this embodiment focuses on the early research and development stage, which differs from the parameter calibration method in the aforementioned embodiments. In actual vehicle deployment, the host computer component corresponding to this embodiment will no longer exist. The parameters obtained after the module completes training will be directly solidified and integrated into the product's software algorithm layer. The actual application process is as described in the above embodiment: the system collects environmental data around the vehicle in real time, and after the trained algorithm model performs scene classification and recognition, it automatically switches the calibration parameters corresponding to the module to achieve optimal adaptation under different scenarios.

[0176] The foregoing has provided a detailed description of a vehicle key communication parameter calibration method and an in-vehicle digital key system provided in this application. The various embodiments in the specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

[0177] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A method for calibrating vehicle key communication parameters, characterized in that, It is applied to multiple anchor points deployed on the vehicle end, each anchor point including: a low-power Bluetooth module and an ultra-wideband module; The methods include: Using any one of the anchor points as a simulated key end and the remaining anchor points as receiving ends, a test task is executed; wherein, the test task includes: controlling the simulated key end to simulate a remote control key, transmitting Bluetooth signals to each of the receiving ends through the low-power Bluetooth module to establish a Bluetooth connection, and sending pulse signals through the ultra-wideband module to perform ranging. Acquire the signal characteristic data of the Bluetooth signal received by each of the receiving terminals, and the ranging result data of the pulse signal; The current scene is identified based on the signal feature data and the ranging result data, and a calibration parameter set corresponding to the current scene is matched. The parameter calibration of the digital key system is completed by matching the calibration parameter set obtained.

2. The vehicle key communication parameter calibration method according to claim 1, characterized in that, After executing the test task using any one of the anchor points as the simulated key end and the remaining anchor points as the receiving end, the method further includes: Switch to the next anchor point as the new simulated key end, and the remaining anchor points as the new receiving ends, and execute the test task until all anchor points have executed the test task once as the simulated key end; The acquisition of signal feature data of the Bluetooth signal received by each of the receiving terminals and ranging result data of the pulse signal includes: When each anchor point is used as the simulated key terminal to perform the test task, the corresponding signal characteristic data and ranging result data received by each receiving terminal are obtained.

3. The vehicle key communication parameter calibration method according to claim 1, characterized in that, The signal characteristic data of the Bluetooth signal includes: RSSI characteristic data; The ranging result data of the pulse signal includes: time of flight data; The current scene types include: open scene, dense scene, and semi-open scene. Identifying the current scene based on the signal feature data and the ranging result data includes: If all the conditions in the open space condition set are met, then the current scene is determined to be an open space scene; The set of open conditions includes: The mean of the RSSI feature data is within an open area, and the standard deviation is less than the first open area threshold; the time-of-flight data meets the preset first stability condition. If all the conditions in the dense condition set are met, then the current scene is determined to be a dense scene; The dense condition set includes: The mean of the RSSI feature data is within the dense interval and the standard deviation is greater than the first dense threshold. The waveform of the RSSI feature data exhibits random jump characteristics. The mean of the time-of-flight data is greater than the second dense threshold and satisfies the second instability condition. If all the conditions in the semi-open space condition set are met, then the current scene is determined to be a semi-open space scene. The semi-empty condition set includes: The mean of RSSI feature data is within the semi-open range, and the waveform shows a step-like decrease; the flight time data satisfies the first stability condition during the line-of-sight path and exhibits abrupt changes during the non-line-of-sight path. The median values ​​of the open area, the semi-open area, and the dense area decrease sequentially.

4. The vehicle key communication parameter calibration method according to claim 3, characterized in that, The current scenario type also includes: dynamic interference scenario; Identifying the current scene based on the signal feature data and the ranging result data further includes: If all the conditions in the dynamic interference condition set are met, then the current scenario is determined to be a dynamic interference scenario. The dynamic interference condition set includes: The fluctuation frequencies of both RSSI feature data and time-of-flight data are greater than the first dynamic interference threshold; the mean value of RSSI feature data is within the dynamic interference range, and the waveform exhibits sawtooth fluctuations and sudden drops with amplitudes greater than the third dynamic interference threshold.

5. The vehicle key communication parameter calibration method according to any one of claims 1 to 4, characterized in that, Identifying the current scene based on the signal feature data and the ranging result data, and matching the calibration parameter set corresponding to the current scene includes: Extract the data features from the signal feature data and the ranging result data, and construct the corresponding scene feature vector; The scene feature vector is input into a pre-trained classification model, and the scene label output by the classification model is obtained. The corresponding calibration parameter set is called from the pre-established dynamic parameter library according to the scene label; The classification model is a machine learning model that performs supervised learning using a sample dataset, which includes multiple groups of sample data labeled with different scene types.

6. The vehicle key communication parameter calibration method according to claim 5, characterized in that, After completing the parameter calibration of the digital key system through the calibration parameter set obtained by matching, the process also includes: Obtain the Bluetooth connection establishment success rate and ranging error of each receiver; The preset communication performance expectation standard is determined based on the Bluetooth connection establishment success rate and ranging error. If not, then retrain the classification model or adjust the calibration parameter set.

7. The vehicle key communication parameter calibration method according to claim 6, characterized in that, Retraining the classification model includes: Acquire historical signal feature data, historical ranging result data, and corresponding historical scene labels from the current time to the last training time of the classification model; Each set of historical signal feature data and historical ranging result data is used as a new sample data set, added to the sample dataset, and the retraining of the classification model is triggered; wherein, the historical scene label is used as the label for the corresponding sample data set.

8. The vehicle key communication parameter calibration method according to claim 6, characterized in that, Adjusting the calibration parameter set includes: Back up the original data of the calibration parameter set; The parameters in the calibration parameter set are increased or decreased according to a preset ratio; The parameter calibration of the digital key system is completed using the adjusted calibration parameter set. Determine whether the Bluetooth connection establishment success rate and the ranging error have been improved; If so, keep the adjustment direction unchanged and return to the step of increasing or decreasing the parameters in the calibration parameter set according to the preset ratio until the Bluetooth connection establishment success rate and the ranging error meet the expected communication performance standard; If not, change the adjustment direction and return to the step of increasing or decreasing the parameters in the calibration parameter set according to the preset ratio; If the Bluetooth connection establishment success rate and ranging error are not improved after the changes and adjustments, the calibration parameter set is restored using the backed-up original data.

9. A vehicle-mounted digital key system, characterized in that, It includes multiple anchor points deployed on the vehicle, each anchor point including: a low-power Bluetooth module and an ultra-wideband module; Each of the anchor points is used to implement the steps of the vehicle key communication parameter calibration method as described in any one of claims 1 to 8.

10. The vehicle-mounted digital key system according to claim 9, characterized in that, The anchor points are connected via CANFD communication, which is used to transmit signal characteristic data and ranging result data. The vehicle-mounted digital key system also includes: a host computer; Among the anchor points, there is a main anchor point. The main anchor point is connected to the host computer via Bluetooth communication through the low-power Bluetooth module, and sends the summarized signal feature data and the ranging result data to the host computer. The host computer is used to: store the signal feature data and the ranging result data, and train a classification model based on the signal feature data and the ranging result data.

Images