Method for precise tracking of a car key based on short-range wireless signals

Abstract

The application provides a decimeter-level vehicle key positioning and tracking method based on short-range wireless signals, belongs to the technical field of intelligent networked vehicle positioning, and can realize real-time and accurate tracking of the position of a vehicle key, meeting actual application requirements such as unlocking a vehicle. In order to improve tracking accuracy, an anchor node selection method is designed, which can effectively identify LoS anchor nodes and eliminate nodes with significant ranging errors, thereby improving the reliability and accuracy of the system. In addition, an AEPF algorithm is proposed, combined with a localized extended EKF, which significantly reduces the particle calculation complexity and optimizes the real-time performance, providing technical support for efficient positioning in complex scenarios. The application has wide application prospects and can be widely applied to the safety management, precise control and personalized services of intelligent networked vehicles. For example, it can be used to realize keyless entry and start of a vehicle, accurate positioning of a parking position, and rapid tracking of a lost vehicle key in a complex scenario, and provides key technical support for the development of intelligent transportation and vehicle networking technology, and has significant market value and promotion potential.

Description

Technical Field

[0001] This invention belongs to the fields of wireless communication and positioning technology and intelligent connected new energy vehicles, and is a method for accurate tracking of car keys based on short-range wireless signals. Background Technology

[0002] With the rapid development of intelligent connected new energy vehicles, vehicles are no longer isolated embedded systems, but rather an important component of intelligent transportation networks. The application of wireless communication technology makes vehicle operation more convenient, allowing users to remotely control and manage their vehicles through smart devices. Against this backdrop, car key tracking systems have gradually become a research hotspot. Precisely tracking the location of the car key enhances security, ensuring the vehicle only unlocks when the key is within a specific range. Simultaneously, accurately knowing the key's location improves user convenience, such as automatically unlocking the doors when the user approaches the vehicle, or activating specific vehicle settings based on the key's distance.

[0003] Currently, there are many positioning technologies based on short-range wireless signals, such as those based on WiFi, Bluetooth, Ultra-Wideband (UWB), and Radio Frequency Identification (RFID). However, these technologies all have limitations when applied to car key location tracking systems. For example, WiFi has a large coverage area, is suitable for medium- to long-range positioning, especially in scenarios with existing WiFi networks, and is relatively easy to deploy and use. However, WiFi has low positioning accuracy and high power consumption, making it unsuitable for devices powered by batteries for extended periods. Bluetooth Low Energy (BLE) offers low power consumption and broad device compatibility, making it suitable for short-range positioning and enabling large-scale applications at low cost. However, its positioning accuracy is relatively low, typically relying on Received Signal Strength Indicator (RSSI) to estimate distance, leading to significant errors. UWB positioning technology can provide centimeter-level ranging accuracy, making it particularly suitable for high-precision applications, and has strong anti-interference capabilities, maintaining stable positioning in complex environments. However, its main drawback is its high cost. RFID, on the other hand, is a low-cost, low-power technology suitable for short-range object tracking and authentication. Passive RFID tags require no battery support and can be used for extended periods, but their positioning range is very short, typically limited to a few meters. In summary, these short-range wireless positioning technologies are limited in their application for car key location and tracking.

[0004] This invention proposes a decimeter-level car key tracking system based on single-carrier frequency-hopping signals, enabling real-time positioning and tracking of car keys. For frequency-hopping systems, a novel anchor node selection method is proposed, which can reliably identify line-of-sight (LoS) anchor nodes while excluding anchor nodes with large ranging errors due to non-line-of-sight (NLoS) conditions. By analyzing the variation trend of the channel impulse response (CIR) amplitude, this method can effectively distinguish between LoS and NLoS anchor nodes. Furthermore, an Adaptive Extended Kalman Particle Filter (AEKPF) algorithm is proposed, which combines the efficiency of Extended Kalman Filter (EKF) with the flexibility of Particle Filtering (PF). EKF can provide local state estimation and optimize particle distribution, thereby reducing computational complexity and improving real-time performance. This hybrid method maintains tracking accuracy even under unpredictable pedestrian movement conditions. Summary of the Invention

[0005] The present invention aims to provide a car key tracking method based on short-range wireless signals, which can achieve accurate tracking of the car key's location.

[0006] The present invention discloses a method for accurate tracking of car keys based on short-range wireless signals, comprising the following steps:

[0007] Step 1: Arrange the anchor nodes according to... Figure 1 The layout is done in a certain way;

[0008] Step 2: Remove NLoS anchor nodes and anchor nodes with large ranging errors to maximize tracking accuracy. The specific steps are as follows:

[0009] Step 2 (1) Preprocess the CIR amplitude, including moving average smoothing and normalization operations, to remove noise and standardize the data;

[0010] Step 2 (2): Using the binary hypothesis testing formula, distinguish between LoS and NLoS anchor nodes based on the characteristic of peak quantity;

[0011] Step 2 (3): Further eliminate anchor nodes with large distance measurement errors by using the triangle inequality to ensure that valid anchor nodes that meet the distance constraints are selected;

[0012] Step 3: Use the iterative least squares method to obtain the initial position of the tracking algorithm;

[0013] Step 4: To further improve tracking accuracy and enhance the real-time performance of the algorithm, the AEKPF algorithm is used to track the target position. The specific steps of the AEKPF algorithm are as follows:

[0014] Step 4 (1) Initialize particle state. Initialize the particle set based on the known initial position points;

[0015] Step 4(2): Adaptively adjust the noise matrix. The observation noise matrix is ​​adaptively adjusted based on the anchor nodes selected in Step 2.

[0016] Step 4 (3) State Prediction. Use EKF to predict the state and estimate the target's position at the current moment based on the system's state transition equation;

[0017] Step 4 (4) Particle weight update. Calculate the likelihood of each particle using the observation data and update the particle weights to ensure that the particle state reflects the true position;

[0018] Step 4 (5) Resampling. The particles are resampled to avoid the degradation problem of low-weight particles, and a new particle set is generated based on the high-weight particles;

[0019] Step 4 (6) Position estimation. The estimated position of the target is calculated based on the state of all particles using a weighted average method, thereby achieving accurate tracking of the vehicle key.

[0020] Beneficial effects

[0021] The present invention has the following advantages:

[0022] 1. High efficiency and reliability: The anchor node selection method reliably identifies Loss of Position (LoS) anchor nodes and can eliminate points with large ranging errors, thereby improving the positioning accuracy of the system;

[0023] 2. Strong real-time performance: The AEKPF algorithm combines the efficiency of EKF with the flexibility of PF, which not only reduces computational complexity but also enhances real-time performance.

[0024] 3. No need to model pedestrian trajectories: Accurate tracking can be achieved regardless of the direction or trajectory of the pedestrian approaching the vehicle;

[0025] 4. Low cost and high precision: By utilizing short-range wireless frequency hopping signals, accurate tracking is achieved on a low-cost hardware platform, providing feasibility for the large-scale application of car key positioning systems. Attached Figure Description

[0026] Figure 1 This indicates the placement location of the anchor nodes in this system;

[0027] Figure 2 Comparison of CIR amplitude spectra before and after preprocessing;

[0028] Figure 3 A comparison of the CIR amplitude spectrum fluctuations in LoS and NLoS. Detailed Implementation Plan

[0029] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings. The present invention provides a method for precise tracking of car keys based on short-range wireless signals, specifically comprising the following steps:

[0030] Step 1, according to Figure 1 The layout shown for anchor nodes ensures that at least one anchor node maintains a LoS connection with the car key regardless of the direction from which a pedestrian carrying the car key approaches the vehicle. However, this layout may also result in some anchor nodes having an NLoS connection with the car key. Directly using the ranging data from all anchor nodes for tracking may introduce significant errors due to NLoS conditions and multipath effects. Therefore, selecting appropriate anchor nodes is essential before performing car key tracking.

[0031] Step 2: Remove NLoS anchor nodes and anchor nodes with large ranging errors to maximize tracking accuracy. The specific steps are as follows:

[0032] Step 2 (1) In frequency hopping systems, the change of center frequency over time introduces different channel conditions, which in turn affects path delay and gain. Under LoS conditions, due to the small attenuation of the direct path, the CIR amplitude remains relatively consistent between adjacent frequencies, and the influence of the reflection path is relatively weak. However, under NLoS conditions, the multipath effect causes significant fluctuations in the CIR amplitude. This difference can be used to distinguish between LoS anchor nodes and NLoS anchor nodes, specifically: under NLoS conditions, there are more detectable peaks in the CIR amplitude curve. However, before peak detection, a series of preprocessing steps are required for CIR to remove interference and improve detection accuracy. To reduce noise-induced fluctuations, moving average smoothing is used to process the CIR amplitude. The CIR amplitude after moving average smoothing can be expressed as:

[0033]

[0034] Where w is the moving average smoothing coefficient, |h i (t) represents the amplitude signal of the i-th frequency point CIR, and S represents the total number of frequency points. Let |h i The expression for (t) after maximum amplitude normalization is:

[0035]

[0036] After preprocessing, the amplitude spectrum of the CIR becomes smoother, such as... Figure 2 As shown.

[0037] Step 2 (2): Using the binary hypothesis test formula, distinguish between LoS and NLoS anchor nodes based on the characteristic of peak quantity. After preprocessing, under LoS conditions, the CIR amplitude becomes smoother, while under NLoS conditions, the CIR amplitude exhibits greater fluctuations, resulting in more peaks, such as... Figure 3 As shown. To maximize the accuracy of peak detection, only peaks with a prominence greater than 0.05 are detected.

[0038] Based on the above analysis, the difference in the number of detected peaks can be used to identify Loss anchor nodes. Specifically, the Loss anchor node identification problem is formulated as a binary hypothesis test, where the Loss condition is H0 and the NLoS condition is H1. As mentioned earlier, the number of peaks under the Loss condition is significantly less than the number of peaks under the NLoS condition. Therefore, we define the hypothesis test as follows:

[0039]

[0040] Where P represents the number of detected peaks, P th This represents the recognition threshold.

[0041] Step 2 (3): After processing the CIR, Loss-of-Sight (LoS) anchor nodes can be identified. However, even under LoS conditions, the distance of the direct path may still be difficult to measure accurately due to the multipath effect. Therefore, anchor nodes with large distance measurement errors need to be eliminated based on the principle of triangle inequalities. Only anchor nodes that satisfy at least two of the following triangle inequalities are selected:

[0042] d i +a ij ≥d j (4)

[0043] Where, d i and d j These represent the distance measurement results for the i-th and j-th anchor nodes, respectively. ij The distance between the i-th and j-th anchor nodes.

[0044] Step 3: Use iterative least squares to obtain the initial position of the tracking algorithm. The core of the tracking model is to predict the target's current state based on its previous state. However, there is no prior state to refer to before the target starts moving, so an initial point needs to be determined first. Anchor node (x, y) and target (x i ,y i The distance between them should be:

[0045]

[0046] First, an initial point (x0, y0) is determined. This point does not need to be very accurate; it is mainly used for algorithm iteration. The estimated distance to the i-th anchor node is:

[0047]

[0048] Then, a first-order Taylor expansion is used to linearize the error between the true distance and the estimated distance:

[0049]

[0050] The coefficient matrix H is as follows:

[0051]

[0052] Finally, through multiple iterations, the most accurate initial estimated position is obtained:

[0053]

[0054] Here, (x0, y0) is the initial position of the tracking algorithm.

[0055] Step 4: By selecting the anchor nodes as described above, even a single accurate ranging result is usually sufficient for tracking. However, tracking the trajectory of a car key still faces many challenges. The movement trajectory of a car key is often difficult to model; in other words, the trajectory of a pedestrian is difficult to predict. For example, a pedestrian approaching a vehicle may suddenly change direction or even stop due to unforeseen circumstances. Furthermore, their movement speed may vary from slow walking to fast running. Therefore, for pedestrian movements with unpredictable direction and speed, EKF struggles to achieve optimal tracking results. Using PF tracking requires setting a large particle range and number, which significantly increases the algorithm's complexity and computation time, making it unsuitable for meeting the high real-time requirements of a car key tracking system. To address this, a method combining EKF in a local region is proposed to optimize particle generation and distribution, thereby reducing computational complexity and improving real-time performance. Simultaneously, the observation noise matrix is ​​adaptively adjusted based on the selection of anchor nodes. This algorithm is called the AEPF algorithm.

[0056] In reality, although it is impossible to accurately model pedestrian trajectories, their motion can be simplified to uniform motion because walking speed is usually relatively stable.

[0057] x t =[x t v x,t y t v y,t ] T (10)

[0058] Where, x t and y tThe two-dimensional coordinates of the target, v x,t and v y,t Let be the corresponding velocity components. The state transition equation is as follows:

[0059] x t =Fx t-1 +w t (11)

[0060] Among them, w t The noise is a process noise that follows a Gaussian distribution. The state transition matrix F is:

[0061]

[0062] Where Δt is the sampling period. Assume the positions of the four anchor nodes are (x1, y1), (x2, y2), (x3, y3), and (x4, y4). The observation vector z... t for:

[0063] z t =[d1 d2 d3 d4] T (13)

[0064] Wherein, each observation distance is:

[0065]

[0066] Here, the observation noise follows a Gaussian distribution.

[0067] The specific steps of the AEKPF algorithm are as follows:

[0068] Step 4 (1) Initialize particle state. Initialize the particle set based on the known initial position points. Assume the first position point is x0 = [x0 0 y0 0] T Set the initial state of all particles to this location point:

[0069]

[0070] Step 4(2): Adaptively adjust the noise matrix. The observation noise matrix is ​​adaptively adjusted based on the anchor nodes selected in Step 2. First, the observation noise matrix is ​​initialized:

[0071]

[0072] Then, define a 4×4 diagonal matrix S with diagonal elements as follows: This represents the selection status of the anchor node. If the i-th anchor node is selected, Set to a smaller value m; otherwise This is set to a large value n (n >> m). To adaptively adjust the observation noise matrix, simply multiply the observation noise matrix by the coefficient matrix:

[0073] R new =R t-1 S (17)

[0074] Step 4 (3), State Prediction. State prediction is performed using EKF, estimating the target's current position based on the system's state transition equation. First, the posterior distribution x is calculated using the standard EKF procedure. t|t The covariance matrix P t|t This provides the particles with a density distribution conforming to the Gaussian recommendation. The state of each particle is updated as follows:

[0075]

[0076] Step 4 (4) Particle Weight Update. Calculate the likelihood of each particle using the observation data and update the particle weights to ensure that the particle state reflects the true position. The weight of each particle is updated based on the observed likelihood value:

[0077]

[0078] in, Based on the state of the i-th particle Predicted observations R represents the weight of the i-th particle at time t. new This is the adjusted observation noise matrix.

[0079] Step 4 (5) Resampling. The particles are resampled to avoid the degradation problem of low-weight particles, and a new particle set is generated based on high-weight particles. After resampling, the weights of all particles are reset to equal values:

[0080]

[0081] Step 4 (6) Position estimation. The estimated position of the target is calculated based on the states of all particles using a weighted average method, thereby achieving accurate tracking of the vehicle key.

[0082]

[0083] The above description is merely a specific embodiment of the present invention. Any feature disclosed in this specification may be replaced by other equivalent or similar features unless otherwise specified. All disclosed features, or steps in all methods or processes, may be combined in any way except for mutually exclusive features and / or steps.

Claims

1. A decimeter-level car key tracking method based on short-range wireless signals, characterized in that: Step 1: Install positioning anchor points on the vehicle; Step 2: Remove non-line-of-sight (NLoS) anchor nodes and anchor nodes with large ranging errors; Step 3: Use iterative least squares method to obtain the initial position of the tracking algorithm; Step four involves tracking the target position using the Adaptive Extended Kalman Particle Filter (AEKPF) algorithm. This includes removing NLoS anchor nodes and anchor nodes with large ranging errors, including: (1) Under NLoS conditions, the amplitude of the Channel Impulse Response (CIR) is processed using moving average smoothing. The CIR amplitude after moving average smoothing is expressed as: （1） in, It is the moving average smoothing coefficient. Representing the The amplitude of CIR at each frequency point Represents the total number of frequency points; The expression after maximum amplitude normalization is: （2） (2) Using the binary hypothesis testing formula, LoS and NLoS anchor nodes are distinguished based on the characteristics of peak quantity; The difference in the number of detected peak values ​​is used to identify Loss anchor nodes; the Loss anchor node identification problem is formulated as a binary hypothesis test, where the Loss condition is... NLoS conditions are The number of peaks under LoS conditions is less than the number of peaks under NLoS conditions; the hypothesis test is defined as follows: （3） in, This represents the number of detected peak values. Represents the recognition threshold; (3) After processing the CIR, identify the Loss anchor nodes; eliminate anchor nodes with large ranging errors according to the principle of triangle inequality; select only at least two anchor nodes that satisfy the following triangle inequality: （4） in, and Representing the first The and the first Distance measurement results for each anchor node For the first The and the first The distance between anchor nodes; The target position is tracked using the Adaptive Extended Kalman Particle Filter (AEKPF) algorithm, which simplifies pedestrian motion into uniform motion. （10） in, and These are the two-dimensional coordinates of the target. and For the corresponding velocity components; the state transition equation is as follows: （11） in, The noise is a process noise that follows a Gaussian distribution. ; State transition matrix for: （12） in, The sampling period is given; assuming the positions of the four anchor nodes are respectively... , , and ; Observation vector for: （13） Wherein, each observation distance is: （14） The observation noise follows a Gaussian distribution. ; The specific steps of the AEKPF algorithm are as follows: (1) Initialize particle state; initialize particle set based on known initial position points; assume the first positioning point is ,in, , This is the initial position of the tracking algorithm; the initial state of all particles is set to the first positioning point. （15） (2) Adaptively adjust the noise matrix; adaptively adjust the observation noise matrix according to the anchor node selected in step two; first initialize the observation noise matrix: （16） Then, define a Coefficient diagonal matrix Its diagonal element is , representing the case where the anchor node is selected; if the first One anchor node was selected. Set as ;otherwise Set as value m; To adaptively adjust the observation noise matrix, multiply the observation noise matrix by the coefficient diagonal matrix: （17） (3) State prediction; Extended Kalman Filter (EKF) is used for state prediction. The position of the target at the current time is estimated based on the state transition equation of the system. First, the posterior distribution xt|t and covariance matrix are calculated through the standard EKF process. This ensures that each particle follows a Gaussian distribution; the state of each particle is updated as follows: （18） (4) Particle weight update; calculate the likelihood of each particle using the observation data and update the particle weight to ensure that the particle state reflects the true position; the weight of each particle is updated according to the observed likelihood value: in, Based on the Particle state Predicted observations express Time of the first The weight of each particle, The adjusted observation noise matrix; (5) Resampling; resampling of particles avoids the degradation problem of low-weight particles, and regenerates the particle set based on high-weight particles; after resampling, the weights of all particles are reset to equal values: （20） (6) Position estimation; The estimated position of the target is calculated based on the state of all particles using a weighted average method, thereby enabling the tracking of the vehicle key: （21）。 2. The decimeter-level car key tracking method based on short-range wireless signals according to claim 1, characterized in that, The initial position of the tracking algorithm is obtained using the iterative least squares method, including: Anchor Node With the goal The distance between them is: （5） First, determine an initial point. Used for algorithm iteration; and the first The estimated distances to each anchor node are: （6） Then, a first-order Taylor expansion is used to linearize the error between the true distance and the estimated distance: （7） Wherein, the coefficient matrix for: （8） Finally, through multiple iterations, the most accurate initial estimated position is obtained: （9）。

Images