UWB-based high-precision positioning method

Abstract

The invention provides a UWB-based high-precision positioning method, and belongs to the technical field of wireless positioning. The method comprises an anchor node communication module and a label node positioning module, adopts an AltDS-TWR protocol to promote communication between a master node and slave nodes, and records timestamps when all nodes send and receive different types of data packets; calculating a time interval according to the timestamp; calculating the time difference tau (m, i) between the signal from the master node to the tag and the signal from the slave node to the tag; estimating a label coordinate p by using a least square method according to tau (m, i) and a preset anchor node coordinate; and obtaining the standard deviation of the noise value and the energy ratio of the first path to all other paths from the channel pulse response of the UWB signal, adjusting the Kalman gain, and updating the original position to the calibrated position based on the Kalman filter. According to the invention, infinite quantity of targets can be supported, each label node only needs to monitor data packet exchange between anchor points without sending any data packet, and clock synchronization is also not needed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wireless positioning technology, and in particular to a high-precision positioning method and system based on channel impulse response quality weighting and passive monitoring multi-target localization (PMTPL) time difference construction of ultra-wide band (UWB). BACKGROUND

[0002] Location-based services are essential in modern life and industry. To ensure accuracy and reliability, effective positioning technology is crucial. The global positioning system (GPS) is currently the most widely used positioning system. By receiving and interpreting GPS satellite signals, positioning functions can be achieved. However, in indoor environments or areas with many obstacles such as urban canyons, GPS cannot effectively perform positioning. In recent years, there has been a strong interest in high-precision positioning technology based on ultra-wide band (UWB) technology. UWB has a wide frequency spectrum and low power density characteristics. With its excellent penetration ability and strong anti-interference ability, UWB technology is expected to solve the limitations of GPS and other positioning technologies in challenging scenarios.

[0003] Existing UWB positioning technology mainly relies on two principles: time of flight (ToF) and time difference of arrival (TDoA). Methods based on ToF (such as AltDS-TWR) have the highest accuracy. However, AltDS-TWR relies on frequent data packet exchange between nodes. As the number of tag nodes increases, the probability of signal collision also increases, ultimately reducing the reliability of the positioning results. At the same time, positioning multiple targets is usually more inclined to TDoA-based methods. These methods can efficiently handle multiple targets, so they are preferred. However, even if the devices maintain close clock synchronization, the accuracy of TDoA calculation will be affected by the clock drift of anchor nodes and tag nodes, so the accuracy of TDoA methods is greatly reduced. Considering this, some people have also proposed a TDoA positioning method that can avoid the influence of clock synchronization. However, this method is still affected by the combined influence of multiple device clock drifts. In addition, due to the complex calculation strategy, small hardware errors will be amplified, so the performance of this method in practical applications is not stable enough. Therefore, a TDoA positioning scheme is needed that can eliminate or reduce clock drift, take into account the quality of channel impulse response (CIR), and adaptively process observation noise in the filter, to achieve more stable high-precision multi-target positioning. SUMMARY

[0004] In view of the above problems in the prior art, the present application provides a high-precision positioning method based on CIR quality weighting and passive monitoring multi-target localization (PMTPL) time difference construction, which solves the problems of the prior art that the positioning method based on UWB has the following problems in achieving high-precision positioning: 1) time difference offset and additional error caused by the clock asynchronization between the transmitting end and the receiving end; 2) time stamp measurement offset or large error caused by multipath effect, non-line-of-sight (NLOS) environment and signal quality fluctuation; 3) insufficient robustness of commonly used positioning operators (such as least squares) in the face of severe noise and abnormal measurement; and 4) the need for tracking of dynamic targets requires that the estimation method take into account real-time performance and accuracy.

[0005] To achieve the above object, the technical scheme adopted by the present application is as follows: a high-precision positioning method and system based on CIR quality weighting and PMTPL time difference construction of UWB signals, comprising the following steps:

[0006] S1, first, a dedicated UWB signal coverage environment is established, including the establishment of an anchor node communication module and a tag node positioning module, and the AltDS-TWR protocol is used to facilitate communication between the master node and the slave node, and the time stamps of all nodes sending and receiving different types of data packets in the communication process are recorded;

[0007] S2, for any slave node a i ∈A, the present application calculates the time interval and time interval T mi according to the distance between the master node a m and the slave node a i ;

[0008] S3, according to the method of calculating the time difference, the time difference τ (m,i) between the signal from the master node to the tag and the signal from the slave node to the tag is obtained;

[0009] S4, according to τ (m,i) and the preset anchor node coordinates, the least squares method is used to estimate the tag coordinates

[0010] S5, the standard deviation of the noise value (std_noise) and the ratio of the first path to all other path energies (fp_power) are obtained from the channel impulse response (CIR) of the UWB signal, and the Kalman gain is adjusted, wherein the Kalman gain isu Adjusting, based on Kalman filter, updates the original position to the calibrated position.

[0011] The present application has the beneficial effects that: the present application proposes a novel high-precision passive monitoring-based positioning method, which has no limit to the number of tag nodes and can locate any tag node with UWB signal receiving capability. The method does not require the tag node to actively participate in the communication process, on the contrary, it can passively monitor the existing UWB signals in the environment, and the tag node can only receive UWB signals and will not actively participate in data packet exchange, thereby eliminating the competition for limited channel resources and ensuring the unlimited scalability of the tag node, thereby independently estimating the coordinates. At the same time, the proposed time difference calculation method is deeply analyzed in theory, and it is found that the positioning error is only affected by the clock drift of a single anchor node. The Kalman filter-based method is used for error calibration, the original position is updated to the calibrated position, and the positioning accuracy is improved.

[0012] Further, the timestamps of the different types of data packets are expressed as follows:

[0013]

[0014] Among them, T T T T T T T T T

[0015] Further, the expression of the time interval is as follows:

[0016]

[0017] Among them, T mi represents the transmission time of the data packet from the master node a m to the slave node a s , d mi represents the distance between the master node and the slave node, and c is the speed of light. represents the transmission time of the data packet from the master node am From sending the Poll request to receiving a request from slave node a i The time interval between Resp responses, Indicates the timestamp of the Resp response data packet sent by the slave node. Indicates the timestamp of the master node sending the Poll data packet; This indicates that the master node receives data from slave node a. i The time interval between responding to the request and sending the final confirmation. Indicates the timestamp of the master node sending the Final data packet. This indicates the timestamp at which the master node received the Resp response data packet; This indicates that the tag node is derived from the main node a. m Receive Poll to label node from slave node a i The time interval for receiving Resp. This indicates the timestamp at which the tag received the Resp data packet. This indicates the timestamp when the label received the Poll data packet; This indicates that the label node is related to the subordinate node a. i Receive response Resp to receiving master node a m The time interval between sending the final status "Final". This indicates the timestamp of the Final data packet received by the tag.

[0018] Furthermore, the time difference τ between the signal from the master node to the tag and the signal from the slave node to the tag (m,i) The expression is as follows:

[0019]

[0020] This expression is based on the aforementioned time interval. T mi , and The expression is derived and obtained, and further, since and They are equal, therefore the time difference τ (m,i) The expression can also be written as:

[0021]

[0022] The beneficial effect of the above-mentioned further scheme is that it proposes a method for calculating the time difference between the signal arriving at the tag from the master node and the signal arriving at the tag from the slave node. This method eliminates the requirement for clock synchronization and ensures that the final result is only affected by the clock drift of a single device.

[0023] Furthermore, the expression for the least squares method is as follows:

[0024]

[0025] This is a nonlinear hyperbolic equation set proposed by the application according to the known anchor node coordinates, wherein c is the speed of light. For the estimated tag coordinates The expression is:

[0026]

[0027] Wherein, τ (m,i) represents the actual value, represents the measured value.

[0028] Further, the CIR of the UWB signal received by the tag can be expressed as a time-domain response function of time t:

[0029]

[0030] Wherein, N is the number of multipaths in the signal, α n represents the step coefficient of the nth multipath, t n represents the delay of the nth multipath, δ(t) represents the unit impulse function, η(t) represents the shape of the pulse, and g(t) is the additive white Gaussian noise (AWGN) with a mean of zero.

[0031] Further, the expression of the adjusted Kalman gain w u is: w u = w n · w p , wherein w n and w p represent the contributions of std_noise and fp_power respectively, std_noise represents the standard deviation of the CIR noise value, and fp_power represents the ratio of the energy of the first path to the energy of all paths.

[0032] The beneficial effects of the above further scheme are: the application marks the estimated coordinates as However, in the NLoS scenario, the signal is affected by the multipath effect, and the accuracy of the estimated coordinates may not be enough. We use a Kalman filter-based method to calibrate the estimated coordinates. Using the method of adjusting the Kalman gain according to the CIR of the UWB signal can generate a relatively accurate tag trajectory prediction value, even if there are incomplete measurement values or noise interference.

[0033] Further, the expression of w n is:

[0034]

[0035] where n e is an empirical threshold, n represents the std_noise of the signal. w p is expressed as:

[0036]

[0037] where p e is an empirical threshold, p represents the fp_power of the signal.

[0038] The above further scheme has the advantage that the value of w u is determined by the interaction of two key factors to evaluate the quality of the current signal, w n and w p are combined to form a total weight w u . This weight determines the degree of trust of the system in the current measurement: the smaller the weight, the less trust in the current measurement; the weight of 1 means full trust. For w n , when the noise n is less than or equal to the empirical threshold n e , the noise is considered acceptable, w n = 1, and the weight is not weakened. When the noise n exceeds the threshold n e , the weight is scaled down by the ratio n e / n; for w p , when the first path energy p is greater than or equal to the threshold p e , the first path is strong enough, w p = 1, and the weight is not weakened, and when p is less than the threshold p e , it means that the first path is weak, possibly NLoS or severe multipath, and the weight is scaled down by p / p e . The above scheme turns the two quality indicators of the CIR into a 0-1 weight w u , reduces the trust in the poor quality measurement by thresholding and scaling, and thus reduces the impact of bad measurements on position estimation in the positioning filter. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a flowchart of the method of the present application.

[0040] Figure 2 is an application scenario diagram of the localization system.

[0041] Figure 3 is a timeline diagram of the data packet exchange process. DETAILED DESCRIPTION

[0042] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0043] Example

[0044] like Figure 1 As shown, this invention provides a high-precision multi-target localization method based on CIR mass weighting and PMTPL time difference, the implementation method of which is as follows:

[0045] S1. First, establish a dedicated UWB signal coverage environment, including the establishment of anchor node communication module and tag node positioning module. Use AltDS-TWR protocol to facilitate communication between master node and slave node, and record the timestamps of different types of data packets sent and received by all nodes during the communication process.

[0046] S2, for any subordinate node a i ∈A, calculate the time interval based on the timestamps collected during the communication process. and Time interval T mi According to the master node a m With subordinate node a i The distance between them is calculated;

[0047] S3. Calculate the time difference τ between the signal from the master node to the tag and the signal from the slave node to the tag using the time difference calculation method. (m,i) ;

[0048] S4, according to τ (m,i) Using the preset anchor node coordinates, the label coordinates are estimated using the least squares method.

[0049] S5. Obtain the standard deviation (std_noise) and the power ratio (fp_power) of the first path to all other paths from the channel impulse response (CIR) of the UWB signal, and adjust the Kalman gain, where the Kalman gain is determined by w u The adjustment, based on a Kalman filter, updates the original position to the calibrated position.

[0050] In this embodiment, as Figure 2As shown, the system consists of two types of UWB signal transceivers: anchor nodes and tag nodes. Anchor nodes are further divided into one master node and multiple slave nodes. Anchor nodes are the infrastructure in the positioning system, whose positions are known and fixed. They provide the reference signals and measurement data required for positioning tag nodes through the communication between master and slave nodes. Tag nodes do not need to communicate with anchor nodes. Tag nodes only need to have the ability to receive and decode UWB signals to estimate their own position coordinates independently. The system can locate any tag node with UWB signal receiving capability. Without the active participation of the tag node in the communication process, it can passively monitor the existing UWB signals in the environment to estimate the coordinates independently.

[0051] System communication model:

[0052] UWB signal transceivers are mainly divided into two types: anchor nodes and tag nodes. According to this classification, the positioning system is divided into two main modules. The description of each module is as follows:

[0053] (1) Anchor node communication module:

[0054] Installing anchor nodes is the initial deployment task of the localization system in the application environment. As shown in Figure 2 , these nodes are strategically placed at predetermined locations as reference points. The system has n+1 anchor nodes. One of them serves as the master node, denoted as a m . The coordinates of the master node are represented by q m (x m ,y m ,z m ). The remaining anchor nodes operate as slave nodes. Use A = {a1, a2,..., a n} to represent the set of n slave nodes, where a i denotes the i-th slave node. In addition, the coordinates of each slave node a i ∈A are denoted as q i (x i ,y i ,z i ).

[0055] In order to establish a specific signal environment that is crucial for accurate positioning and accurate results, the AltDS-TWR protocol is used to facilitate communication between the master node and the slave nodes. This protocol uses a Poll-Resp-Final structure. For a given a i ∈A, there are six timestamps recorded from the master node and the slave nodes: In each communication round, the timestamp of the master node broadcasting the Poll packet is recorded as The slave node records the timestamp as The timestamp record of the time when the slave node sends the Resp response packet is The timestamp record of the time when the master node receives the Resp response packet is The timestamp record of the time when the master node sends the Final final packet is The timestamp record of the time when the slave node receives the Final final packet is For the convenience of understanding the symbol representation in the communication process, the timeline of the data packet exchange process is listed in Figure 3

[0056] (2) Tag node positioning module:

[0057] The purpose of the system is to accurately locate and track the tag node and provide its coordinates. In the positioning process, the tag node does not need to communicate with the anchor node. On the contrary, as long as the tag node has the ability to receive and decode UWB signals, it can independently estimate its own position coordinates. The tag node can be attached to various objects, pedestrians and unmanned vehicles.

[0058] The focus is on analyzing the positioning process of a single tag node, because the position estimation of each tag node is independent of other nodes. Let p(x, y, z) represent the coordinates of the tag. In the existing signal environment, the tag node entering the system can estimate its own position by monitoring the UWB signals transmitted by the anchor node. The timestamps of the Poll, Resp and Final data packets received by the tag node are defined as and

[0059] Let T m represent the time of flight of the Poll data packet or the Final data packet from the master node to the tag. This value is obtained by:

[0060]

[0061] Similarly, T i represents the time of flight of the Resp response data packet from the slave node to the tag, which can be represented as:

[0062]

[0063] Therefore, the time difference between the signal from the master node a m to the tag and the signal from the slave node a i to the tag can be represented as

[0064] τ (m,i) = T m -T i

[0065] According to the known anchor node coordinates, the present application proposes a nonlinear hyperbolic equation set, represented as​

[0066]

[0067] where c is the speed of light.

[0068] To obtain the best estimate of the tag node coordinates, the present invention utilizes the least square method. The measured time difference is denoted as The coordinates of the tag should be calculated by minimizing the difference between the true value and the measured value, which is expressed as

[0069]

[0070] In the TDoA positioning algorithm, accurate time difference is crucial for effective estimation of coordinate accuracy. The measured time difference is significantly affected by clock errors, including bias and drift. Let b m , b s and b t denote the time bias of the master node, slave node and tag node, respectively. In addition, e m , e s and e t denote their respective time drift factors. Therefore, the clock bias model of the timestamps and can be expressed as:

[0071]

[0072] where denote the timestamps recorded by the master node, slave node and tag node, respectively.

[0073] Based on the above formula analysis, the error between the measured value and the actual value τ (m,i) can be obtained by the following formula:

[0074]

[0075] where, and are the measured time intervals.

[0076] According to the above formula, the error is obviously affected by the cumulative clock bias and drift of multiple devices. Clock synchronization can be achieved through labor-intensive tasks to eliminate the effects of clock bias. However, the clock bias between multiple devices is still one of the main factors affecting TDoA accuracy. Therefore, the present invention faces two major challenges. The first is to eliminate the need for clock synchronization between devices. The second is to mitigate the impact of multi-device clock drift on TDoA calculation. The method proposed by the present invention aims to minimize the impact of clock drift and calculate the most accurate time difference.

[0077] PMTPL algorithm overview:

[0078] The present invention proposes a novel high-precision positioning algorithm that can locate any tag node with UWB signal receiving capability. The algorithm does not require the tag node to actively participate in the communication process. Instead, it passively monitors the UWB signals present in the environment, thereby independently estimating the coordinates. The algorithm of the present invention consists of the following four steps. The pseudo code of these steps is described in Table 1.

[0079] In the first step, for any slave node a i ∈A, the present invention calculates the time interval and the time interval T mi from the timestamps collected during the communication process; m i ;

[0080] In the second step, the time difference τ (m,i) between the signal from the master node to the tag and the signal from the slave node to the tag is obtained according to the method of calculating the time difference.

[0081] In the third step, the tag coordinates are estimated using the least squares method based on τ (m,i) and the pre-set anchor node coordinates.

[0082] In the fourth step, std_noise and fp_power are obtained from the CIR of the UWB signal, the Kalman gain is adjusted, and the original position is updated to the calibrated position based on the Kalman filter. u

[0083] Table 1 PMTPL algorithm

[0084]

[0085] Time difference derivation for tag nodes:

[0086] Next, the new method for calculating the time difference between the signal from the master node to the tag and the signal from the slave node to the tag in the present invention is introduced. The method of the present invention eliminates the requirement for clock synchronization and ensures that the final result is only affected by the clock drift of a single device. The derivation process can be explained as follows:

[0087] Let T mi represent the transmission time of the data packet from the master node a m to the slave node a s . Since the positions of the master node and the slave node are fixed and predetermined, the present invention knows the distance between them in advance, denoted by d​​mi T can be expressed as: mi T can be expressed as:

[0088]

[0089] where c is the speed of light. In addition, the present application can also describe the relationship between the time stamp recorded in the communication process as: mi T can be expressed as:

[0090]

[0091] The present application will T is defined as the time interval between the master node a m sending a Poll request and receiving a Resp response from the slave node a i T can be expressed as:

[0092]

[0093] Let T denote the time interval from the tag node receiving a Poll from the master node a m to the tag node receiving a Resp from the slave node a i According to the following formula derivation, we can get:

[0094]

[0095] Similarly, T denotes the time interval between the master node receiving a response from the slave node a i and sending a Final final confirmation packet, which can be expressed as:

[0096]

[0097] Let T denote the time interval from the tag node receiving a response Resp from the slave node a i to receiving the final state Final sent by the master node a m According to the following formula derivation, we can get:

[0098]

[0099] According to the time interval calculation formula of and T can be calculated as: (m,i)

[0100]

[0101] Here is equal to , so the above formula can also be written as:​

[0102]

[0103] The CIR calibration coordinates are used:

[0104] In this example, the present application uses a calibration least square method to estimate the coordinates. The estimated coordinates are denoted as However, in NLoS scenarios, the signal is affected by multipath effects, and the estimated coordinates may not be accurate enough. To solve this problem, the present application adopts a KF-based algorithm to calibrate the estimated coordinates. The KF-based algorithm includes two main steps: a prediction step and an update step. The former is to obtain the predicted value of the system state. The latter is to correct the predicted value using the measurement value. KF can generate a relatively accurate label trajectory prediction value even if there are incomplete measurement values or noise interference.

[0105] In the following, the present application will introduce the KF-based coordinate calibration method, which realizes calibration by determining the weight assigned to the measurement value in the update step.

[0106] Under the influence of multipath effects, the UWB signal generated by the transmitter will superimpose on the receiver after experiencing different delays. Therefore, the CIR of the UWB signal received by the tag can be represented by the time-domain response function of time t, i.e.:

[0107]

[0108] where N is the number of multipaths in the signal, α n represents the step coefficient of the nth multipath, t n represents the delay of the nth multipath, δ(t) represents the unit impulse function, η(t) represents the shape of the pulse, and g(t) is the additive white Gaussian noise (AWGN) with a mean of zero. According to the above formula, we can obtain two key factors for evaluating the signal quality, i.e. std_noise and fp_power. Among them, std_noise represents the standard deviation of the CIR noise value, and fp_power represents the ratio of the energy of the first path to the energy of all paths. Let w u represent the weight of the communication channel quality. The value of w u is determined by the interaction of the two key factors, denoted as w n and w p represent the contributions of std_noise and fp_power, respectively. The assignment relationship of w n and w p can be described as:

[0109]

[0110] where n e and p e are empirical thresholds, n and p represent std_noise and fp_power of the signal respectively.

[0111] Error theory analysis:

[0112] In this example, the error of PMTPL method in calculating the time difference of arrival of signals is analyzed. It is noted that and represent the time interval recorded by the same device, which means that they are not affected by clock offset, but only by clock drift. The measured time interval is modeled as:

[0113]

[0114] where and represent the measurement time period. According to τ (m,i) the calculation formula and the above four formulas, the calculation result is represented as:

[0115]

[0116] Therefore, the error between the calculated value and the actual value is:

[0117]

[0118] Based on the theoretical error value e m of the time difference of arrival of UWB signals, (m,i) the PMTPL method not only eliminates the need for strict clock synchronization between multiple devices, but also completely limits the error within the clock drift of a single device. This error reaches the minimum achievable value in time-of-flight (ToF) and time difference of arrival (TDoA) positioning algorithms, which has a significant advantage over methods that are susceptible to clock drift of multiple devices (VULoc).

[0119] In summary, compared with the prior art, the present application has the following advantages:

[0120] By using the architecture of "active exchange between anchors + passive listening of tags" in PMTPL, the present application can theoretically realize unlimited number of tags based on passive monitoring. Specifically, the present application not only eliminates the uplink interaction of each tag, avoids message conflicts from the source, constructs reference signals through local exchange between anchors, and simplifies the timing management, but also reduces the error sources without increasing the complexity of the tags, so that the exchange of data packets required for positioning only occurs between a fixed number of anchor nodes, and the tags only passively receive, thereby fundamentally decoupling the system message overhead and the number of tags.

[0121] By performing a combination of the time intervals of multiple messages between anchors, the application constructs a time difference estimator, so that most of the drift terms in the tag observations are automatically canceled out, and the final residual error is only proportional to the clock drift of a certain master node: error = e m τ (m,i) Not only is the error source greatly reduced, only the drift of a single master node needs to be concerned, the upper limit of the positioning error is significantly lower than methods affected by multi-device drift, the calibration and maintenance overhead is small, system maintenance is mainly for the master anchor, robustness is enhanced, long-term operation and stability in complex scenarios are better. The method also combines calibration methods such as CIR-based weighting and Kalman filtering to further suppress noise and multipath effects, making the obtained tag coordinates more accurate.

[0122] The application solves the scalability bottleneck and the precision upper limit problem caused by multi-device drift at the same time, so that the system can support geometric scale tag concurrency while still maintaining more accurate positioning accuracy, which has significant theoretical and engineering value.

Claims

1. A high-precision positioning method based on UWB, characterized in that, The method comprises the following steps: S1, first establish a dedicated UWB signal coverage environment, including the establishment of anchor node communication module and tag node positioning module, using AltDS-TWR protocol to promote the communication between master node and slave node, record the time stamp of all nodes sending and receiving different types of data packets in the communication process; S2, for any slave node a i ∈ A, the time interval is calculated herein from the timestamps collected during the communication and the time interval T mi from the master node a m and the slave node a i ; S3. Derive the time difference τ between the signal from the master node to the tag and the signal from the slave node to the tag according to the method of calculating the time difference (m,i) ; S4. The method of S3, (m,i) and the preset anchor node coordinates, estimating the tag coordinates using a least squares method S5. Obtain the standard deviation (std_noise) and the power ratio (fp_power) of the first path to all other paths from the Channel Impulse Response (CIR) of the UWB signal, and adjust the Kalman gain, where the Kalman gain is determined by w u The adjustment, based on a Kalman filter, updates the original position to the calibrated position.

2. The method for high-precision positioning based on UWB according to claim 1, characterized in that, The time stamp of the different types of data packets is as follows: wherein, Tstamp Poll, Tag represents the timestamp of the Poll packet received by the tag node, Tstamp Resp, Tag represents the timestamp of the Resp packet received by the tag node, Tstamp Final, Tag represents the timestamp of the Final packet received by the tag node Tstamp Poll, Master represents the timestamp of the Poll packet sent by the master node, Tstamp Final, Master represents the timestamp of the Final packet sent by the master node, Tstamp Resp, Master represents the timestamp of the Resp packet received by the master node, Tstamp Poll, Slave represents the timestamp of the Poll packet received by the slave node, Tstamp Final, Slave represents the timestamp of the Final packet received by the slave node, Tstamp Resp, Slave represents the timestamp of the Resp packet sent by the slave node.

3. The method for UWB-based high-precision positioning according to claim 1, characterized in that, The expression of the time interval is as follows: where T mi represents the transmission time of the data packet from the master node a m to the slave node a s , d mi represents the distance between the master node and the slave node, and c is the speed of light. represents the time interval between the master node a m sending the Poll request and receiving the Resp response from the slave node a i , represents the time stamp of the slave node sending the Resp response data packet, represents the time stamp of the master node sending the Poll data packet; represents the time interval between the master node receiving the response from the slave node a i and sending the final confirmation, represents the time stamp of the master node sending the Final data packet, represents the time stamp of the master node receiving the Resp response data packet; represents the time interval between the tag node receiving the Poll from the master node a m and the tag node receiving the Resp from the slave node a i , represents the time stamp of the tag receiving the Resp data packet, represents the time stamp of the tag receiving the Poll data packet; represents the time interval between the tag node receiving the response Resp from the slave node a i and receiving the final state Final sent by the master node a m , represents the time stamp of the Final data packet received by the tag. 4.The UWB-based high-precision positioning method of claim 1, wherein, The time difference expression between the signal from the master node to the tag and the signal from the slave node to the tag is as follows: The expression is derived from the expression of claim 3 for the time interval and Further, since is equal to the expression for the time difference τ (m,i) can also be written as:

5. The method for UWB-based high-precision positioning according to claim 1, characterized in that, The least squares method estimates the coordinates The expression is as follows: where the first equation is a nonlinear hyperbolic system of equations derived from the known anchor node coordinates, where c is the speed of light, p represents the estimated coordinates, q m represents the master node coordinates, q1-q n represents the slave node coordinates, and for the estimated tag coordinates the expression where, τ (m,i) represents the actual value, represents the measured value. 6.The UWB-based high-precision positioning method according to claim 1, characterized in that, The weight w of the adjustment Kalman gain u The expression for calibrating the original position update is as follows: where h(t) denotes a Channel Impulse Response (CIR) of a UWB signal received by a tag expressed by a time-domain response function of time t, N denotes a number of multipaths in the signal, a n denotes a step coefficient of the nth multipath, t n denotes a delay of the nth multipath, δ(t) denotes a unit impulse function, η(t) denotes a shape of a pulse, g(t) is an Additive White Gaussian Noise (AWGN) with a mean of zero, w n denotes a contribution of a standard deviation (std_noise) of a CIR noise value, w p denotes a contribution of a ratio (fp_power) of an energy of a first path to an energy of all paths, n e denotes an empirical threshold of std_noise, p e denotes an empirical threshold of fp_power, n denotes std_noise of a signal, p denotes fp_power of a signal. 7.The UWB-based high-precision positioning method of claim 1, wherein, The derivation of the high-precision low-error is as follows: The method according to claim 3, and The time intervals represent time intervals recorded by the same device, which means that they are not affected by clock drift, only by clock drift. The measured time intervals are modeled as: wherein and denotes a measurement period, e t and e m denotes a drift factor of the device, τ (m,i) The calculation formula is calculated The calculation expression of The error between the calculated value and the actual value is:

Images