Fast indoor positioning algorithm based on PDDASC Bluetooth phase distance measurement

Abstract

The invention provides a fast indoor positioning algorithm based on PDDASC Bluetooth phase ranging, and relates to the technical field of indoor ranging, a plurality of Bluetooth 6.0 devices are used as base stations for deployment, a tag of a Bluetooth 6.0 chip or an intelligent device is used as a positioning beacon, a phase ranging (PBR) function of a Bluetooth 6.0 protocol is used, roles of an initiating device and a reflecting device are clarified, and the positioning accuracy is improved. Channel detection and bidirectional data exchange are carried out to calculate signal phase delay and distance, an IQ value interpolation and spatial smoothing method is adopted to supplement an IQ value of a missing frequency band so as to generate equally-spaced IQ signals, distance estimation errors are reduced, spatial power spectrum calculation is carried out through a direct propagation operator method, and signal arrival time is determined. A space power spectrum offset compensation mechanism is introduced, and a main lobe offset error caused by noise and multipath interference is judged and compensated by searching a maximum value point and a minimum value point of a space spectrum.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of indoor distance measurement, and in particular to a fast indoor positioning algorithm based on PDDASC Bluetooth phase ranging. BACKGROUND

[0002] Due to the fact that satellite navigation cannot be used in an indoor environment, with the development of Internet of Things technology, the application demand for indoor precise navigation and asset tracking is increasing. Embedded devices have the disadvantages of limited computing resources and insufficient memory space, making it difficult to deploy large-scale programs. Therefore, a fast and low-complexity indoor distance calculation method is needed for fast deployment on embedded devices. At present, there is still a lack of a solution for indoor positioning that is fast to deploy, low in price and accurate in positioning.

[0003] With the popularity of smart phones and low-power smart devices (watches, bracelets), Bluetooth is widely used in these smart devices. Therefore, it is a very promising solution to use Bluetooth to realize indoor positioning. There are currently three ways to realize indoor positioning based on Bluetooth technology: Bluetooth RSSI positioning, Bluetooth AOA lateral positioning and Bluetooth channel sounding positioning. Bluetooth RSSI positioning first establishes a signal propagation strength model to obtain the relationship between signal strength and distance. Then, through trilateration (least squares method), the position is located. Although various methods are used to improve the accuracy of RSSI positioning, the positioning accuracy cannot reach the decimeter level. Bluetooth AOA lateral positioning uses the IQ data collected by an array antenna to achieve high-precision angle of arrival estimation through MUSIC or other spatial spectrum algorithms, and then through multi-base station cooperation and least squares triangulation to realize indoor position determination of the target. Or the covariance matrix of the incident signal is obtained through the collected IQ data, and then input into the machine learning model as a feature to achieve high-precision angle of arrival estimation. Although the positioning performed by these Bluetooth AOA lateral methods has good positioning accuracy, it relies on high-precision antenna arrays. Therefore, the device has many limitations in terms of cost, size and deployment. Currently, the latest Bluetooth 6.0 protocol includes Bluetooth channel sounding technology, which includes phase-based ranging and time-of-arrival ranging functions. Among them, phase-based ranging has high measurement accuracy. It uses the frequency hopping mechanism of Bluetooth BLE channels to exchange channels within the entire 2.4 GHz frequency band. To achieve multi-channel channel switching. The phase difference on multiple channels is used to obtain accurate distance measurement. This method does not rely on complex antenna devices and has high measurement distance accuracy. Based on this, we propose a low-complexity fast phase ranging algorithm that takes into account the advantages of low cost and high accuracy. SUMMARY

[0004] To address the shortcomings of existing technologies, this invention proposes a fast indoor positioning algorithm based on Bluetooth phase ranging using PDDASC.

[0005] A Bluetooth phase ranging algorithm based on Propagator Direct Data Acquisition and Spatial Spectral Compensation (PDDASC) is implemented to achieve fast distance measurement. Measurement accuracy is improved through spatial spectrum compensation (SC). This algorithm significantly reduces computational complexity compared to classic music or OMP algorithms during ranging. To address interference caused by foot traffic or object movement in indoor environments, a DBSCON clustering-based filtering algorithm is proposed to filter out interference, thereby achieving accurate positioning in indoor environments.

[0006] Compared with the prior art, the present invention has the following advantages: 1. This invention introduces a direct propagation operator method for distance estimation in Bluetooth phase ranging schemes for the first time. This method is a fast spatial spectrum estimation method that not only guarantees distance estimation accuracy but also significantly reduces computational load. It is suitable for rapid deployment in embedded devices.

[0007] 2. This invention proposes for the first time a spatial spectrum compensation algorithm to compensate for spatial spectrum shift. This method can significantly improve the accuracy of distance estimation. Comparison with several common Bluetooth phase ranging methods shows that this method exhibits superior performance.

[0008] 3. To address indoor signal interference, especially non-line-of-sight interference caused by people walking or objects moving, DBSCON clustering is used for filtering. Outliers generated during interference filtering further improve positioning accuracy. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the arrangement of Bluetooth phase ranging base stations.

[0010] Figure 2 This is a schematic diagram illustrating the principle of Bluetooth phase ranging.

[0011] Figure 3 This is a spatial smoothing diagram of a multi-carrier phase signal.

[0012] Figure 4 This is a flowchart of the spatial power spectrum compensation process.

[0013] Figure 5 This is a diagram showing the effect of distance interference filtering.

[0014] Figure 6 This is a flowchart of the overall implementation plan. Detailed Implementation

[0015] The following are specific embodiments of the present invention, which are described in conjunction with the accompanying drawings. However, the present invention is not limited to these embodiments.

[0016] 1. Scene setup This invention employs multiple Bluetooth 6.0 devices as Bluetooth base stations for deployment. The positioning devices are tags embedded with Bluetooth 6.0 chips or low-power smart devices (watches, bracelets) with Bluetooth 6.0 functionality. The deployment method is as follows: Figure 1 As shown. These devices need to support the Bluetooth 6.0 protocol. This solution can be deployed in a variety of environments. If deployed in a large shopping mall or hospital for location navigation, four or more base stations are required. If deployed in factories and warehouses for asset tracking, Bluetooth 6.0 tags need to be embedded in the asset devices, and three or more base stations are required. If deployed in cars as digital car keys, only one base station needs to be deployed in the car, with low-power smart devices (watches, bracelets) acting as tags.

[0017] 2. Bluetooth channel detection signal representation Bluetooth 6.0's Phase-Based Ranging (PBR) feature defines two device roles: the initiator and the reflector. The initiator initiates the ranging process, while the reflector responds to it. Channel detection is achieved through one or more channel detection sub-events, each containing two or more channel detection steps. In each step, bidirectional data exchange occurs between the devices, with the initiator always sending a signal first, followed by one or more response transmissions from the reflector. (See below.) Figure 2 As shown. Signals experience phase delay as they propagate through the air. Using... Let represent the time offset, d represent the distance between the initiator and the reflector, and C represent the speed of light. This is expressed as the signal arrival delay. At this point, in the Kth channel, the phase of the signal received by the reflector... Phase of the signal received by the initiator It can be expressed by the following formula: in This indicates the frequency of the Kth channel used by the initiator. The frequency of the Kth channel used by the reflector is shown. and These represent the phase detected by the initiator and the reflector in the Kth channel, respectively. Since the same channel is used for both initiation and reflection, therefore... equal The signals received by the reflector and the initiator at this time can be represented as: in and Let represent the signal noise received by the transmitter and receiver, respectively. We can cancel out the effect of the local oscillator using the following calculation:

[0018] in and This represents the signal amplitude value of the i-th channel. and Let represent the signal phase value of the i-th channel. These two parameters can be directly represented by the acquired IQ values. Therefore, the sum of the two phases can be expressed as:

[0019] Since Bluetooth channel probing requires switching across the entire Bluetooth channel, assuming there are K channels, the entire channel probing process can be represented as:

[0020] Where:

[0021]

[0022]

[0023] 3. IQ value interpolation and spatial smoothing.

[0024] In Bluetooth Low Energy 6.0 systems, only 72 radio frequency (RF) channels are used during phase detection interactions. The center frequency of these RF channels is calculated using the formula "2402 + k MHz", where k is an integer ranging from 2 to 22 and 26 to 76. Since signals with consecutive equally spaced frequencies are not used, the acquired IQ values ​​are incomplete. A mean interpolation method is used to fill in the missing IQ values, resulting in equally spaced IQ signals and reducing distance estimation errors. After interpolation, the signal can be represented as a steering vector of a one-dimensional linear matrix:

[0025] In real-world environments, signals propagate through walls, people, furniture, and other objects, resulting in reflections and multipath effects. This causes signals from the same source to arrive at the receiver via different paths, sharing the same carrier frequency and exhibiting a stable phase relationship, thus creating interfering coherent signals. To address this issue, spatial smoothing techniques can be incorporated, such as... Figure 3 As shown. Spatial smoothing technique divides a uniform linear array into several overlapping subarrays with identical structures. Then, by "smoothing" and averaging the covariance matrices of all subarrays, the rank of the received data covariance matrix, which has been corrupted by coherence, is restored. Therefore, we smooth a 1×K array of a one-dimensional incident signal into an L×M array, where M = K - L + 1. The smoothed matrix representation is as follows: 4. Spatial power spectrum calculation Suppose that the data matrix X contains M data points acquired by L sensors. Then X is an L×M matrix. The direct propagation operator method divides the X matrix into two submatrices, G and g. The former is a 1×M vector containing the first row of r(t), and the latter is an L-1 ×M matrix.

[0026]

[0027] Where g represents the first row of matrix X, which is a 1×M matrix; G represents the remaining part of matrix X, which is an (L-1)×M matrix. A vector P is defined to represent the correlation between different sensors, and this vector P can be expressed by the following formula:

[0028] Let P represent the cross-correlation between the first row of data and other rows of data, which is a 1×(L-1) matrix. Adding a unit 1 to represent the correlation between the first row of data and itself, and merging it with P, forms an L×1 matrix, which can be represented as:

[0029] The spatial power spectrum can then be expressed by the following formula:

[0030] in , represents the signal subspace steering vector. Let τ represent the signal arrival delay, and c be the speed of light. When the spatial power spectrum reaches its maximum value, the arrival time τ can be used to calculate the maximum possible distance d.

[0031] 5. Spatial power spectrum shift compensation: The maximum unambiguous distance for Bluetooth phase detection is: The spatial spectrum depends on the propagation vector P, which represents the correlation between the time series of each sensor element and the first time series. The generated e is an L×1 dimensional vector. The theoretical width distance of the main lobe of the spatial power spectrum is:

[0032] Super-resolution direct propagation operator methods maximize the main lobe as much as possible, with the point of maximum value in the spatial spectrum corresponding to the most probable phase detection distance. However, in real-world environments, various noises and multipath interferences can cause the main lobe to shift, resulting in measurement errors. To reduce the errors caused by main lobe shift, the following compensation is implemented.

[0033] First, the point of maximum value is found by searching the spatial spectrum of the direct propagation operator. This point represents the most likely distance between the base station and the beacon. Then, the search continues at the first minimum point before the peak, i.e., to the left of the peak. If a minimum is found, the distance represented by the minimum point is assumed to be... The distance represented by the peak value is .if: in Fine-tuning is possible within the range [1, 1.5]. This indicates a possible rightward shift in the peak position, necessitating calibration compensation. The compensation value should be... for: Compensated distance:

[0034] if: Therefore, it is assumed that no compensation is needed, and the detection range is at its maximum value.

[0035] If there is no minimum value on the left, the predicted distance is also... .

[0036] The entire compensation algorithm process is as follows: Figure 4 As shown.

[0037] 6. Distance measurement interference filtering In indoor scenarios, whether in a lobby or warehouse, short-term non-line-of-sight occlusion is a major factor interfering with positioning stability. For example, personnel walking in the lab can obstruct the base station; similarly, moving warehouse beacons can cause shelves to obstruct them. These brief occlusions cause fluctuations in the collected IQ data, leading to outliers in the estimated distance. We employ the Density-Based Spatial Clustering of Applications with Noise (DBSCON) method to filter outliers and reduce signal interference caused by such brief occlusions. Setting the DBSCAN parameters to a neighborhood radius ε = 0.02 and a minimum sample size MinPts = 5, its combination with the least squares method effectively eliminates outlier measurements, achieving the best filtering effect. The filtering effect is shown below. Figure 5 As shown in the figure, after setting the DBSCAN filtering parameters, when the measurement signal experiences severe jitter, these jitter points can be effectively removed, thus helping to obtain more stable positioning results.

[0038] 7. Real-time positioning After obtaining the distances to the beacon using multiple base stations, the coordinates of the indoor beacon are obtained using the least squares method, as shown in the following formula:

[0039] Where (x, y) represents the coordinates of the beacon. Let represent the coordinates of n base stations. Further, we can obtain the following formula:

[0040] Therefore, the coordinates of the beacon can be represented as:

[0041] in:

[0042] The real-time coordinates of the beacon can be calculated using the formula above.

[0043] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A fast indoor positioning algorithm based on PDDASC Bluetooth phase ranging, characterized in that, Includes the following steps: Multiple Bluetooth 6.0 devices are deployed as base stations, and the positioning devices are tags with embedded Bluetooth 6.0 chips or low-power smart devices with Bluetooth 6.0 functionality; By utilizing the phase ranging function of the Bluetooth 6.0 protocol, the roles of the initiating device and the reflecting device are identified, channel detection is performed, and the signal phase delay is calculated through bidirectional data exchange to obtain the distance between devices. By applying IQ value interpolation and spatial smoothing techniques, missing IQ values ​​in frequency bands are supplemented to form equally spaced IQ signals, reducing distance estimation errors and solving the coherent signal problem caused by multipath effects. The spatial power spectrum is calculated using the direct propagation operator method to determine the signal arrival time and then to find the maximum possible distance. Spatial power spectrum offset compensation reduces measurement errors caused by noise and multipath interference; The DBSCAN method is used to filter out interfering outliers in distance measurements, improving positioning stability. After obtaining the distance to the beacon using multiple base stations, the real-time coordinates of the indoor beacon are obtained using the least squares method.

2. The fast indoor positioning algorithm based on Bluetooth phase ranging using PDDASC according to claim 1, characterized in that, The channel detection process includes one or more channel detection sub-events, each sub-event contains two or more channel detection steps, bidirectional data exchange between devices, and the initiating device sends a signal first.

3. The fast indoor positioning algorithm based on Bluetooth phase ranging using PDDASC according to claim 1, characterized in that, The IQ value interpolation and spatial smoothing technique includes using a mean interpolation method to fill in the missing IQ values ​​of the frequency bands and smoothing the one-dimensional incident signal into an L×M array, where M=K-L+1 and K is the number of channels.

4. The fast indoor positioning algorithm based on Bluetooth phase ranging using PDDASC according to claim 1, characterized in that, The spatial power spectrum calculation includes dividing the data matrix into two sub-matrices, constructing a comprehensive correlation matrix by calculating the correlation between different sensors, and calculating the spatial power spectrum accordingly.

5. The fast indoor positioning algorithm based on Bluetooth phase ranging using PDDASC according to claim 1, characterized in that, The spatial power spectrum shift compensation includes searching for the maximum value point of the spatial spectrum and searching for the minimum value point before the peak value, and determining whether calibration compensation is needed based on the distance between the two.

6. The fast indoor positioning algorithm based on Bluetooth phase ranging using PDDASC according to claim 1, characterized in that, The DBSCAN method filters out interference anomalies by setting neighborhood radius and minimum sample number parameters, eliminating abnormal measurements, and improving positioning stability.

Images