Low-altitude aircraft cooperative positioning method based on Bluetooth ad hoc network

Abstract

The invention provides a low-altitude aircraft cooperative positioning method based on a Bluetooth ad hoc network, and the method comprises the steps: enabling low-altitude aircraft nodes to find neighbors through Bluetooth broadcasting and scanning, and constructing the ad hoc network; carrying out Bluetooth RSSI distance measurement on all nodes in the network and neighbor nodes, smoothing RSSI through Kalman filtering, converting the RSSI into distance by using an online calibrated path loss model, and recording a result; selecting three aircraft nodes which meet a triangular inequality and are neighbor to one another, and establishing a global coordinate system by taking the aircraft nodes as initial anchor nodes; selecting three anchor nodes adjacent to the non-anchor node, performing trilateral positioning by using the distance measurement results of the three anchor nodes and the node to be positioned, calculating the coordinates of the three anchor nodes, and upgrading the three anchor nodes into the anchor nodes; and iteratively executing the steps until all nodes in the network are positioned. According to the invention, the low-cost and low-power-consumption Bluetooth module is utilized, the distributed and high-robustness cooperative positioning of the aircraft cluster is realized through an innovative signal processing and modeling method in an environment without GPS signals, the support of a ground base station is not needed, and the method is suitable for emergency investigation, cluster performance and other scenes.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of wireless communication and navigation positioning, and specifically relates to a cooperative positioning method for low-altitude aircraft based on Bluetooth ad hoc networks. Background Technology

[0002] With the widespread application of drones in logistics, surveying, emergency communications, and other fields, accurate positioning is a crucial prerequisite for completing missions. Currently, drones mainly rely on GPS for positioning, but in urban canyons, indoor spaces, forests, or environments subject to interference, GPS signals weaken or even fail, causing drones to malfunction.

[0003] To address the problems of existing technologies, several solutions have been proposed, including: UWB positioning technology: While highly accurate, it suffers from high hardware costs and power consumption, making it unsuitable for large-scale, low-cost drone swarms. Visual / LiDAR SLAM: Relies on environmental features, its performance drops sharply in feature-sparse or adverse weather conditions, and it requires significant computational power and airborne computing capabilities. Traditional base station positioning: In remote areas or at low altitudes, base station coverage is insufficient, resulting in limited positioning accuracy. Bluetooth technology, with its advantages of low cost, low power consumption, and high integration, has become the preferred choice for short-range communication. However, its direct application to dynamic aircraft positioning faces challenges: Low ranging accuracy: Bluetooth Received Signal Strength Indication (RSSI) is susceptible to multipath effects, antenna orientation, and environmental disturbances, resulting in ranging stability and accuracy far lower than UWB. The raw RSSI value fluctuates greatly, leading to significant ranging errors when used directly. Difficult dynamic topology management: Low-altitude aircraft nodes move rapidly, and network topology changes frequently, making traditional static or quasi-static positioning algorithms difficult to apply.

[0004] Therefore, there is an urgent need in this field for a solution that can leverage the advantages of Bluetooth hardware and overcome its accuracy deficiencies through innovative signal processing, model calibration, and networking algorithms to achieve low-cost, highly robust cooperative positioning of low-altitude aircraft. Summary of the Invention

[0005] The purpose of this invention is to provide a cooperative positioning method for low-altitude aircraft based on Bluetooth self-organizing networks, so as to solve the problem that drone swarms cannot perform low-cost and effective positioning in environments without GPS signals.

[0006] To address the problems existing in the prior art, this invention proposes a cooperative positioning method for low-altitude aircraft based on Bluetooth ad hoc networking, the method comprising:

[0007] S1. Each aircraft node in the cluster periodically broadcasts a Bluetooth signal containing its own unique ID and scans to receive broadcast signals from other nodes, building a dynamic neighbor list and forming a decentralized Bluetooth self-organizing network.

[0008] S2. After a node collects messages broadcast by its neighboring nodes, it collects RSSI samples, performs Kalman filtering on the RSSI samples, and estimates the distance between the two nodes after filtering.

[0009] S3. Construct an initial coordinate system and select anchor nodes based on the distance between the two nodes;

[0010] S4. Among the non-anchor nodes in the network, find a node K that is a neighbor of at least three anchor nodes; use the filtered distance between node K and these three anchor nodes to calculate the three-dimensional coordinates of node K using the trilateration algorithm. After successful localization, upgrade node K to a new anchor node and add it to the anchor node set.

[0011] S5. Sequentially locate the non-anchor nodes that meet the conditions in the network and upgrade them to anchor nodes until all nodes in the network have been located.

[0012] S6. The system automatically triggers the relocation process at fixed time intervals or according to the network topology change rate, and returns to step S2.

[0013] S7. Monitor the network topology in real time. When a node is detected to have fewer than 3 neighbors or when there are significant changes in the network structure, return to step S3.

[0014] The beneficial effects of this invention are:

[0015] This invention fully utilizes consumer-grade Bluetooth chips, eliminating the need for dedicated high-precision positioning hardware such as UWB, thus significantly reducing cluster deployment costs. Employing a self-organizing network architecture, it requires no ground infrastructure or central server, offering strong resilience and suitability for emergency scenarios. By effectively suppressing instantaneous RSSI fluctuations through Kalman filtering and combining it with an online calibration mechanism based on the path loss model, it dynamically adapts to environmental changes, significantly improving ranging reliability from the data source and ensuring the accuracy and stability of the overall positioning system. Specifically designed for dynamic aircraft, this invention effectively addresses network topology changes caused by rapid node movement through dynamic update and relocation mechanisms. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the cooperative positioning method for low-altitude aircraft based on Bluetooth ad hoc networks according to the present invention.

[0017] Figure 2 This is a schematic diagram illustrating the establishment of the initial coordinate system of the present invention;

[0018] Figure 3 This is a schematic diagram of the three-sided positioning of the present invention;

[0019] Figure 4 This is a schematic diagram of the iterative triaxial positioning process of the present invention;

[0020] Figure 5 This is a schematic diagram of the whole-network positioning diffusion process of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] A cooperative localization method for low-altitude aircraft based on Bluetooth ad hoc networks, such as Figure 1 As shown, the method includes:

[0023] S1. Each aircraft node in the cluster periodically broadcasts a Bluetooth signal containing its own unique ID and scans to receive broadcast signals from other nodes, building a dynamic neighbor list and forming a decentralized Bluetooth self-organizing network.

[0024] S2. After a node collects messages broadcast by its neighboring nodes, it collects RSSI samples, performs Kalman filtering on the RSSI samples, and estimates the distance between the two nodes after filtering.

[0025] S3. Construct an initial coordinate system and select anchor nodes based on the distance between the two nodes;

[0026] S4. Among the non-anchor nodes in the network, find a node K that is a neighbor of at least three anchor nodes; use the filtered distance between node K and these three anchor nodes to calculate the three-dimensional coordinates of node K using the trilateration algorithm. After successful localization, upgrade node K to a new anchor node and add it to the anchor node set.

[0027] S5. Sequentially locate the non-anchor nodes that meet the conditions in the network and upgrade them to anchor nodes until all nodes in the network have been located.

[0028] S6. The system automatically triggers the relocation process at fixed time intervals or according to the network topology change rate, and returns to step S2.

[0029] S7. Monitor the network topology in real time. When a node is detected to have fewer than 3 neighbors or when there are significant changes in the network structure, return to step S3.

[0030] A cooperative localization method for low-altitude aircraft based on Bluetooth ad hoc networks includes the following steps:

[0031] S1: Network construction and neighbor discovery.

[0032] Each aircraft node in the cluster periodically broadcasts a Bluetooth signal containing its own unique ID and scans to receive broadcast signals from other nodes, thereby establishing and maintaining a dynamic "neighbor list" and forming a decentralized Bluetooth self-organizing network.

[0033] S2: Relative distance measurement and filtering.

[0034] After receiving a broadcast signal from its neighbor node j, node i records its RSSI value. To improve ranging stability, node i continuously collects multiple RSSI samples and smooths them using a Kalman filter. Subsequently, an online-calibrated path loss model is used to convert the filtered RSSI value into an estimated distance d_ij between node i and node j.

[0035] Furthermore, the Kalman filtering step S21 mentioned in step S2 is as follows:

[0036] S211: The Kalman filter model is defined as follows:

[0037] Equations of state:

[0038]

[0039] Observation equation:

[0040]

[0041] in, Indicates at time The state vector. In this application, it is defined as a two-dimensional vector. . Indicates time Filtered RSSI estimate (unit: dBm). Indicates time The rate of change of RSSI (in dBm / s) is used to track the trend of RSSI and improve dynamic performance. A is the state transition matrix; it describes how the state changes from time 1 to 2. Evolution to time .

[0042] For the model, assuming the rate of change of RSSI is constant over a short period of time, then:

[0043]

[0044] in, This indicates the time interval between two consecutive RSSI measurements (in seconds). This is process noise; it follows a pattern with a mean of 0 and a covariance matrix of... The Gaussian distribution, i.e. . The matrix represents the uncertainty we have in predicting the state, which is usually determined through experimental debugging. Indicates at time The actual measured raw RSSI value (unit: dBm). This is the observation matrix; it maps the state vector to the observation space. Since only RSSI values ​​can be directly measured, , representing the observed value Only with the state vector Related. The observed noise; it follows a pattern with a mean of 0 and a variance of . The Gaussian distribution, i.e. . The value reflects the noise level of the RSSI measurement itself and can be obtained through statistical analysis of RSSI in static scenarios.

[0045] S212: Prediction.

[0046] Prior state estimation:

[0047]

[0048] Prior error covariance:

[0049]

[0050] S213: Updated.

[0051] Kalman gain calculation:

[0052]

[0053] Posterior state estimation:

[0054]

[0055] Update error covariance:

[0056]

[0057] After filtering, the state vector is taken. The first element in The smoothed RSSI value at the current moment is used for subsequent distance calculations.

[0058] Furthermore, step S22 of the path loss model scheme mentioned in step S2 is as follows:

[0059] S221: The path loss model formula is as follows:

[0060]

[0061] in, : Indicates distance Path loss at the location (unit: dB). It is related to the received signal strength. The relationship is ,in It is the transmit power (unit: dBm), which is usually a known or estimable fixed value in Bluetooth broadcasting. : The distance between the transmitting node and the receiving node to be determined (unit: meters). Reference distance (unit: meters): a known short distance close to the source of the emission, usually taken as 1 meter. At reference distance The average path loss at a given location (in dB). This value can be calculated theoretically (e.g., using a free-space model) or... The actual measurement was obtained. Path loss exponent. It is a key parameter characterizing the rate at which path loss increases with distance. Its value is closely related to the environment: Free space: Urban environment: Indoor spaces with obstacles: . Shadow fading (unit: dB), a Gaussian random variable with a mean of 0, i.e. It reflects the random changes in signals caused by environmental obstacles.

[0062] S222: Distance calculation formula.

[0063] To solve for the distance Ignore random variables (Because its mean is 0), and the model is rearranged:

[0064]

[0065] Further derivation yields the distance The calculation formula is as follows:

[0066]

[0067] S223: Online calibration of model parameters.

[0068] To improve the model's adaptability to different environments, this invention introduces an online calibration mechanism to dynamically estimate the most critical parameter—the path loss exponent. .

[0069] Calibration method: When there are at least two reference nodes with known precise locations in the network (their coordinates can be obtained in advance via GPS or other high-precision methods), the measured RSSI and known true distance between them can be used to inversely solve the problem. .

[0070] Let the true distance between reference nodes i and j be... .

[0071] The smoothed RSSI values ​​between them are obtained by Kalman filtering. .

[0072] Will and Substitute into the path loss model and reverse the calculation. :

[0073]

[0074] To improve robustness, data from multiple pairs of reference nodes can be collected, and multiple [data points] can be calculated. The average or median value is taken as the calibration value for the current environment. And update it to the above distance calculation formula.

[0075] S3: Initial coordinate system establishment and anchor node selection.

[0076] like Figure 2 As shown, select three mutually neighboring nodes A, B, and C from the network, and the distances between them satisfy the triangle inequality: AB + BC > AC, AB + AC > BC, AC + BC > AB.

[0077] With node A as the origin (0, 0, 0) and the direction from A to B as the positive X-axis, determine the Y-axis and Z-axis according to the right-hand rule (or through the position of node C) to establish a three-dimensional global coordinate system. Calculate the coordinates of nodes B and C accordingly, and mark nodes A, B, and C as initial anchor nodes, adding them to the anchor node set.

[0078] like Figure 4 The diagram illustrates the process of establishing the initial coordinate system and selecting anchor nodes.

[0079] Furthermore, the initial coordinate system establishment step S31 mentioned in step S3 is as follows:

[0080] S311: Search the network for three neighboring nodes A, B, and C that satisfy the triangle inequality.

[0081] S312: Let , in (AB represents the measured distance).

[0082] S313: Coordinates of node C This is obtained by solving the following system of equations:

[0083]

[0084]

[0085] (By selecting a reasonable quadrant, assuming...) Alternatively, an initial height could be used to simplify the problem into a two-dimensional solution, and then the Z-axis could be gradually corrected using information from other sensors.

[0086] S4: Iterative trilateration.

[0087] In the network's non-anchor nodes, find a node K that is a neighbor of at least three anchor nodes. Using the filtered distances between node K and these three anchor nodes, calculate the three-dimensional coordinates of node K using a trilateration algorithm (solving a set of spherical equations). Upon successful localization, upgrade node K to a new anchor node and add it to the anchor node set.

[0088] Furthermore, such as Figure 3 and Figure 4 As shown, step S41 of the trilateration algorithm mentioned in step S4 is as follows:

[0089] S411: For the node K to be located, assume its three anchor node neighbors are M1(x1,y1,z1), M2(x2,y2,z2), and M3(x3,y3,z3), with distance measurements of d1, d2, and d3 respectively. For example... Figure 5 As shown, the process of trilateration is illustrated.

[0090] S412: Establish a system of equations:

[0091]

[0092]

[0093]

[0094] S413: Solve for the coordinates (x, y, z) of node K using numerical methods such as the least squares method.

[0095] S5: Network-wide location dissemination. For example... Figure 5 As shown, step S4 is repeated to locate non-anchor nodes that meet the conditions in the network and upgrade them to anchor nodes in turn, until all nodes in the network have been located.

[0096] A non-anchor node that meets the criteria is one that can observe at least three anchor nodes that have already been located in the current neighbor list. This non-anchor node will then be promoted to an anchor node.

[0097] S6: Dynamic positioning update.

[0098] Considering the continuous movement of the aircraft, the system automatically triggers a relocation process starting from step S2 at fixed time intervals (e.g., 1 second) or according to the network topology change rate to update the position information of all nodes.

[0099] Calculate the network topology change rate:

[0100] Step 1: Define the time window: Let the current time be... The previous calculation time was ,in For the testing cycle;

[0101] Step 2, Record the neighbor list: Each node in and Maintain your own neighbor list at all times. and ;

[0102] Step 3: Detect Individual Changes: For a given node i, determine whether its neighbor relationships have changed. If... If the local topology of node i has changed, then it is considered that the local topology of node i has changed.

[0103] Step 4: Statistical Analysis of Network-wide Changes: Let the number of network nodes be M. Within the time window... Count the number of nodes whose neighbor relationships have changed, and denote them as follows: ;

[0104] Step 5: Calculate the rate of change based on the number of nodes whose neighbor relationships have changed.

[0105]

[0106] The higher the TCR value, the stronger the network dynamics, the faster the nodes move, and the more frequently they update their locations.

[0107] Step 6: Set trigger thresholds, which include high dynamic thresholds and reconstruction thresholds. The high dynamic threshold is: if TCR > 0.3 (i.e., 30% of nodes have changed their neighbor relationships in the last second), the system determines that the network is in a high dynamic state and should trigger a shorter period of re-ranging and repositioning. The reconstruction threshold is: if TCR > 0.6, it may mean that the cluster is undergoing drastic maneuvers or formation changes, at which point the S7 network reconstruction mechanism can be triggered.

[0108] S7: Topology changes and relocation mechanisms.

[0109] The system monitors the network topology in real time. When it detects that a node has fewer than 3 neighbors or that there are significant changes in the network structure, the system can automatically return to step S1 or S3 to reconstruct the network or reinitialize the coordinate system to ensure the continuity of positioning.

[0110] Significant changes in network structure include: the current network state is severely inconsistent with the "basic topology framework" upon which the current global coordinate system and subsequent iterative positioning rely. Continuing iterative positioning in the current coordinate system will lead to a sharp amplification of errors or complete positioning failure. Standard: Statistically, the proportion of nodes in the network with fewer than 3 neighbors is used. If this proportion exceeds 50%, it means that more than half of the nodes cannot perform even the most basic trilateration, and the network is on the verge of positioning failure.

[0111] The above-described embodiments further illustrate the purpose, technical solution, and advantages of the present invention. It should be understood that the above-described embodiments are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made to the present invention within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for cooperative positioning of low-altitude aerial vehicles based on Bluetooth ad hoc networks, characterized in that, The method comprises the following steps: S1, each aircraft node in the cluster periodically broadcasts a Bluetooth signal containing its own unique ID, and scans to receive the broadcast signals of other nodes, builds a dynamic neighbor list, and forms a decentralized Bluetooth ad hoc network; S2, after the node collects the messages broadcast by the neighbor nodes, RSSI samples are collected, Kalman filtering is performed on the RSSI samples, and the distances between the two nodes after filtering are estimated; S3, an initial coordinate system is constructed, and anchor nodes are selected according to the distances between the two nodes; S4, in the non-anchor nodes of the network, a node K is found which is a neighbor of at least three anchor nodes; the three-dimensional coordinates of the node K are calculated by a trilateration algorithm using the filtered distance values between the node K and the three anchor nodes, and after successful positioning, the node K is upgraded to a new anchor node and added to the anchor node set; S5, the non-anchor nodes in the network that meet the conditions are positioned in turn and upgraded to anchor nodes until all nodes in the network are positioned; S6, the system automatically triggers the repositioning process at fixed time intervals or according to the network topology change rate, and returns to step S2; S7, the network topology is monitored in real time, and when it is detected that the number of neighbors of a node is less than 3 or the network structure changes significantly, step S3 is returned. 2.The method of claim 1, wherein, The Kalman filtering of the RSSI sample comprises: defining the state equation and observation equation of the Kalman filter model; obtaining a state transition matrix; performing prior state estimation on the RSSI sample according to the state equation, the observation equation and the state transition matrix, and calculating the prior error covariance; calculating the Kalman gain according to the prior state estimation, and performing posterior state estimation and error covariance updating according to the Kalman gain, to complete the Kalman filtering of the RSSI sample. 3.The method of claim 1, wherein, The distance between the two nodes after filtering is estimated, which comprises: calculating the path loss of the filtered signal by using a path loss model; and calculating the distance between the two nodes according to the path loss.

4. The method of claim 3, wherein, The path loss model is: ； in, Indicates distance Path loss at the location, Let be the distance between the transmitting node and the receiving node to be determined. For reference distance, To be at the reference distance Average path loss at that location This is the path loss index. For the shadow to fade.

5. The method of claim 3, wherein, The distance between the two nodes is calculated as: ； ； in, For reference distance, For transmission power, To be at the reference distance Average path loss at that location To receive signal strength, This is the path loss index.

6. The method of claim 1, wherein, The anchor node is selected according to the distance between two nodes: searching three nodes A, B and C in the network which are neighbors of each other and satisfy the triangle inequality; setting , wherein ; Coordinates of the computing node C .

7. The method of claim 1, wherein, The three-dimensional coordinates of the node K are calculated by a trilateration algorithm, which comprises: for the node K to be positioned, its three anchor node neighbors are M1(x1, y1, z1), M2(x2, y2, z2), and M3(x3, y3, z3), and the distances are d1, d2, and d3, respectively; an equation group is constructed according to the anchor nodes; and the least square method or other numerical method is used to solve the equation group to obtain the coordinates (x, y, z) of the node K. 8.The method of claim 1, wherein, The non-anchor node that meets the condition is: the non-anchor node can observe at least three anchor nodes that have completed positioning in the current neighbor list, and the non-anchor node is upgraded to an anchor node.

9. The method of claim 1, wherein, The network topology change rate is calculated as: Step 1, define time window: let the current time be the last calculation time be wherein is the detection period; Step 2, record neighbor list: each node maintains its own neighbor list at and all times and ; Step 3, detecting individual change: For a node i, determine whether its neighborhood relationship has changed, if then consider that the local topology of node i has changed; Step 4, count the total network changes: let the number of network nodes be M; count the number of nodes whose neighbor relationship has changed in the time window , denoted as ; Step 5, the change rate is calculated according to the number of nodes whose neighbor relationship has changed: ； The higher the TCR value, the stronger the network dynamics, the faster the node moves, and the more frequently the positioning is updated. Step 6, setting a trigger threshold, the trigger threshold includes a high dynamic threshold and a reconstruction threshold; wherein the high dynamic threshold is: if TCR>0.3, the system determines that the network is in a high dynamic state, and a shorter period of re-ranging and positioning should be triggered; the reconstruction threshold is: if TCR>0.6, the cluster is undergoing a dramatic maneuver or formation change, at which time the network reconstruction mechanism is triggered.

10. The method of claim 1, wherein, The network structure changes significantly include: the state of the current network is inconsistent with the basic topology skeleton relied on the construction of the current global coordinate system and the subsequent iterative positioning; and the proportion of nodes with less than 3 neighbors in the network is less than 50%, otherwise the network structure is unchanged.

Images