Estimation and compensation method for differential age error in real-time motion positioning
The method addresses RTK positioning errors by modeling and compensating for age differences through a carrier phase observation equation and sequential Kalman filtering, enhancing accuracy and reliability in RTK positioning systems.
Patent Information
- Application Number
- CN202411070393.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-06
- Publication Date
- 2024-12-03
AI Technical Summary
Existing RTK positioning technologies face challenges in maintaining high continuity, reliability, and precision under large age differences due to asynchronous time delays between base station and rover observations, leading to increased errors and positioning failures.
A method that models and compensates for age difference errors in real-time motion positioning by constructing a carrier phase observation equation, using a double difference algorithm, and applying Taylor expansion and sequential Kalman filtering to estimate and simplify the error, followed by equivalent transformation and sequential filtering to reduce computational complexity.
This approach effectively reduces RTK positioning errors and failures by accurately estimating and compensating for age differences, ensuring high accuracy and reliability even under large time delays.
Smart Images

Figure CN119064964A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of wireless communications, and in particular to a method for estimating and compensating age errors in real-time motion positioning. Background Art
[0002] Real-time kinematic (RTK) technology based on the Global Satellite Navigation System (GNSS) is widely used in autonomous driving, drones, precision agriculture, intelligent robot navigation and other fields due to its unique advantages such as high accuracy, fast initialization speed and good adaptability. Ideally, the observations of the base station and the rover at the same epoch should be obtained to completely eliminate various error terms. Otherwise, the positioning performance may deteriorate. In most practical real-time applications, base station data is transmitted through wireless communication technology. For example, in the RTK (NRTK) network widely used in autonomous driving applications, the observation results are broadcast through cellular mobile communication networks. However, due to the inherent uncertainty of the radio propagation environment, there will be uncertainty delays or interruptions when transmitting base station observations to the rover. Therefore, the asynchronous time between the base station and the rover is inevitable, which is called the difference age. Its actual value may range from less than one second to several seconds, and sometimes even hundreds of seconds. In this case, it is quite challenging to ensure that the RTK solution meets the high continuity, reliability and precise positioning requirements of high-dynamic applications. The solutions to this problem mainly include asynchronous RTK (ARTK) and prediction methods.
[0003] The ARTK algorithm directly uses the historical observation data of the base station and the latest observation data of the mobile station for RTK processing. In other words, the errors caused by the differential age are ignored in these methods. Therefore, as the differential time increases, errors such as satellite ephemeris error, clock error, and atmospheric error will increase sharply, causing the positioning accuracy of ARTK to drop rapidly.
[0004] Prediction methods mainly include state-space representation (SSR) prediction methods and observation-space representation (OSR) prediction methods. Real-time SSR precision corrections, such as satellite clock and orbit corrections and ionospheric delay, are essential for PPP-RTK applications. Due to the low frequency of providing corrections and the delays caused by network processing, users must predict these corrections in a timely manner. Several prediction methods have been proposed, such as Kalman filtering, polynomial fitting, and neural network optimization methods, which have high computational costs or only predict GPS orbit and clock corrections. On the other hand, in the case of base station age differences, OSR prediction methods are also used for RTK processing. Patent US5903236 solves the characteristics of small age differences in aircraft precision landing applications. By using the carrier phase observation data of the base station in the past few epochs, a least squares-based polynomial fitting method for base station carrier phase observation time is proposed. The accuracy of the predicted observation value of this method is related to the number of historical epochs of the base station carrier phase observation and the extrapolation time. The disadvantage of this method is that the accuracy of its predicted observation value decreases rapidly with the extension of the extrapolation time, so it is only suitable for small base station age differences.
[0005] Therefore, improving the continuity and reliability of RTK positioning solutions under large age differences is a valuable technical challenge that needs to be urgently solved in this field. Summary of the invention
[0006] According to one aspect of the present invention, a method for estimating and compensating for aging errors in real-time motion positioning is disclosed, including: constructing a carrier phase observation equation of a satellite at time t, processing the carrier phase observation equations sent by satellite i and satellite j to a base station and a mobile station using a double difference algorithm, and extracting the error caused by aging; modeling the error caused by aging in the carrier phase observation equation and providing an estimated expression; simplifying the estimated expression of the error caused by aging in the carrier phase observation equation using equivalent transformation and sequential Kalman filtering; and compensating for the aging error in the double difference operation using the simplified estimated amount of the error caused by aging.
[0007] The age difference error estimation and compensation algorithm disclosed in the present application starts from the carrier phase observation equation, models the overall age difference error through Taylor expansion, and establishes an observation and state transition model based on time difference carrier phase (TDCP). The estimator of the model is constructed based on Kalman filtering. In addition, the equivalent transformation and sequential filtering techniques are used to significantly reduce the computational complexity of the estimator. Since the error of each phase observation is modeled separately, it is easy to handle when the reference satellite changes frequently. It solves the technical problem that in many RTK applications, there is inevitably a time delay between the observation data of the base station and the observation data of the meteor station, which leads to a rapid increase in RTK positioning errors or positioning failures. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Figure 1 A schematic diagram of a positioning system 100 according to an embodiment of the present invention is provided;
[0009] Figure 2 A flow chart of a method 200 for estimating and compensating for age error in real-time motion positioning according to an embodiment of the present invention is given. DETAILED DESCRIPTION
[0010] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not have to be used to implement the present invention. In other examples, in order to avoid confusing the present invention, known circuits, materials or methods are not specifically described.
[0011] Throughout the specification, reference to "one embodiment", "embodiment", "one example" or "example" means that a particular feature, structure or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "in one embodiment", "in an embodiment", "one example" or "example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. In addition, specific features, structures or characteristics can be combined in one or more embodiments or examples in any appropriate combination and / or sub-combination. In addition, it should be understood by those of ordinary skill in the art that the diagrams provided herein are for illustrative purposes, and the diagrams are not necessarily drawn to scale. It should be understood that when an "element" is said to be "connected to" or "connected" to another element, it can be directly connected or connected to another element or there can be an intermediate element. On the contrary, when an element is said to be "directly connected to" or "directly connected to" another element, there is no intermediate element. The same reference numerals indicate the same element. The term "and / or" used herein includes any and all combinations of one or more related listed items.
[0012] Figure 1 A schematic diagram of a positioning system 100 according to an embodiment of the present invention is given, which exemplarily includes a base station, a mobile station and multiple satellites, wherein the satellite sends observation values to the base station and the mobile station, which can be expressed using a carrier phase observation equation. Formula (1) gives the carrier phase observation equation of satellite i at time t.
[0013]
[0014] Among them, the subscript b represents a base station, and the subscript r represents a mobile station.
[0015] φ b i (t) is the carrier phase observation value of satellite i at time t, in meters.
[0016] ρ b i Indicates the distance between the satellite antenna phase center and the GNSS receiver antenna phase center in meters.
[0017] I i (t) and T i (t) represents the ionospheric delay and tropospheric delay at time t, respectively, in meters.
[0018] dt b (t) and dt i (t) represents the receiver clock error and satellite clock error at time t, in seconds.
[0019] δ b and δ i are the receiver and satellite carrier phase hardware delays in meters, respectively.
[0020] o i (t) is the error of the satellite orbit at time t, in meters.
[0021] φ b and φ i represents the initial phase between the receiver and satellite i, in units of periods.
[0022] λ is the carrier phase wavelength in meters per cycle.
[0023] N i is the carrier integer ambiguity in weeks.
[0024] σ b j is the observation error.
[0025] c is the propagation speed of electromagnetic waves in space.
[0026] When performing RTK calculations, double difference (DD) operations are first used to eliminate various uncertain errors. As shown in formula (2):
[0027]
[0028] The superscript (.)ij represents the differential operation between satellites i and j.
[0029] represents the DD observation at time t.
[0030] Δt represents the asynchronous time between the base station observation and the mobile station observation, also known as the difference age.
[0031] Formula (2) can be further expressed as follows:
[0032]
[0033] Among them, ρ rb ij 、N rb ij , ε rb ij represent the distance between the satellite antenna phase center and the GNSS receiver antenna phase center of satellite i and satellite j relative to the base station and mobile station, the carrier integer ambiguity, and the observation covariance matrix, respectively.
[0034] e age ij (t; Δt) represents the error caused by age difference.
[0035] Considering the short baseline case, ignoring the atmospheric parameter differences introduced by the spatial factor, substituting (1) into (2) yields:
[0036]
[0037] Define the following function:
[0038] f i (t) = [T i (t)-I i (t)-cdt i (t)+o i (t)], (4)
[0039] Then age ij (t; Δt) can be expressed as follows (5):
[0040]
[0041] Then equation (3) can be expressed as equation (6):
[0042]
[0043] In the prior art, the short baseline double difference DD observation model widely used only considers the influence of the first three items. The present invention introduces the difference age error term, that is, e age ij (t; Δt) represents the double difference DD error at time t caused by the age difference Δt. The present invention models, estimates and compensates for the error e caused by the age difference age ij (t; Δt) to improve the application performance of RTK.
[0044] The present invention firstly models and estimates the error caused by the age difference.
[0045] The elements contained in equation (4), namely atmospheric delay, satellite clock error and orbit error, change slowly over time, and it is a smooth function mathematically. Expanding it using Taylor series at time t gives the following result:
[0046]
[0047] In the formula, (.)′ and (.)″ represent the first-order derivative and the second-order derivative. In order to balance the stability and accuracy of the model, only the first term of the Taylor series is retained and the higher-order terms are ignored. For convenience, define α i (t) is f i (t) is the first-order derivative at time t. Substituting (7) into (5) yields:
[0048]
[0049] Next, we need to estimate item α from the base station observations i (t)-α j (t).
[0050] The precise position of the base station in the geocentric earth fixed (ECEF) coordinate system is known. Using the broadcast ephemeris, the observation equation of the satellite carrier phase can be expressed as:
[0051]
[0052] Among them, v φ i =φ b i (t)-ρ b i The observed value minus the calculated value (OMC) of the undifferentiated carrier phase. b iThe variance of can be calculated using the elevation angle sine function model:
[0053]
[0054] Where El is the satellite altitude. a and b are constants. For geodetic GNSS receivers, the values of a and b are usually 4.3 mm and 3 mm.
[0055] The following time difference carrier phase (TDCP) ratio is further defined:
[0056]
[0057] Combining (9) and (7), we can obtain:
[0058]
[0059] in:
[0060]
[0061] is the time-varying observation noise.
[0062]
[0063] Since the first-order derivative α i (t) should be a value fluctuating around zero, so it is reasonable to assume that it is a first-order Gauss-Markov process. Assume that the receiver clock fluctuates η r (t) is a random walk process. In fact, in further processing, η r Item (t) will be eliminated.
[0064]
[0065] Where T i and n α i is the correlation time of α and the process noise. And n α i It follows a zero-mean normal distribution, n α i ~N(0,qi). n ηr is the process noise of ηr, which follows a zero-mean normal distribution, n ηr ~N(0,qr).
[0066] From the above model and assumptions, we can build the state transformation and observation function to estimate the differential error. Define the following state vector:
[0067] x=[α 1 α 2 ...αN ηr] T . (16)
[0068] By the assumption of formula (15), we can get:
[0069]
[0070] Where F and n t are the state transformation matrix and process noise vector respectively, which are expressed as follows:
[0071]
[0072] n t It follows a zero-mean normal distribution, n t ~N(0,Q), where Q is the covariance matrix.
[0073]
[0074] All the time difference carrier phase ratio measurements of each satellite are combined together to define the following observation vector:
[0075]
[0076] According to the state vector definition in equations (12) and (16), the following observation function can be obtained:
[0077] z=Hx+∈ (22)
[0078] Among them, the observation matrix H is:
[0079]
[0080] Assume that the observations are independent of each other and the observation noise ∈ is a Gaussian distribution vector with diagonal covariance, ε ~ N(0,R), where R = diag(e σ1 e σ2 …e σN ).
[0081] Next, the present invention simplifies the estimator. In one embodiment, the present invention uses a simplified estimator with equivalent transformation and sequential Kalman filtering.
[0082] The observation matrix H in formula (23) can be expressed by formula (24):
[0083] H=[B1 B2] (24)
[0084] Where B1 is the identity matrix I, B2 = [1 1...1] T .
[0085] According to the equivalent transformation principle, the observation function can be transformed into:
[0086]
[0087] Among them, ∈′ is the first-order derivative of ∈, has the same static properties as ∈, and obeys the normal distribution with zero mean, ∈′~N(0,R).
[0088] x′ is the first-order derivative of the state vector x, expressed as follows:
[0089] x′=[α 1 α 2 ... α N ] T , (26)
[0090] M=I-B2(B2 T RB2) -1 B2 T R. (27)
[0091] Among them, B is a vector of all 1s and R is a diagonal matrix. Therefore, M can be simplified to:
[0092]
[0093] Therefore, we can get the complete Kalman filter expression:
[0094]
[0095] in,
[0096] A=exp(FΔt)≈I+ΔtF (30)
[0097] Since it is reasonable to assume that the observations are independent, a sequential Kalman filtering technique is used here to avoid the high computational complexity of the matrix inversion operation.
[0098] Next, the above derivation results are used to compensate for the error caused by the difference in the double differential DD operation.
[0099] When estimating α for each observed satellite i When , the differential error in the double difference DD operation can be obtained.
[0100] The differential matrix D that defines the differential operation between the norm satellite and the reference satellite.
[0101]
[0102] The estimated age error is:
[0103]
[0104] where D and x' are the differential matrices at the current epoch t and the estimated α. The final covariance of the double-differenced DD observation with forecast error uncertainty is:
[0105]
[0106] Where P is the prediction covariance matrix of the current epoch, Ω dd is the original double difference DD observation covariance matrix, that is, ε rb ij ~N(0,Ω dd ).
[0107] Figure 2 A flowchart of a method 200 for estimating and compensating for age difference error in real-time motion positioning according to an embodiment of the present invention is provided. The method 200 for asynchronous time compensation in real-time motion positioning comprises the following steps:
[0108] Step 201: construct the carrier phase observation equation of the satellite at time t, use the double difference algorithm to process the carrier phase observation equation sent by satellite i and satellite j to the base station and the mobile station, and extract the error caused by the age difference;
[0109] Step 202: Model the error caused by the age difference in the carrier phase observation equation and give an estimated value expression;
[0110] Step 203: using equivalent transformation and sequential Kalman filtering to simplify the estimated expression of the error caused by the age difference in the carrier phase observation equation;
[0111] Step 204: Using the simplified estimate of the error caused by the difference age, compensate for the difference age error in the double difference DD operation.
[0112] The age difference error estimation and compensation algorithm disclosed in the present application starts from the carrier phase observation equation, models the overall age difference error through Taylor expansion, and establishes an observation and state transition model based on time difference carrier phase (TDCP). The estimator of the model is constructed based on Kalman filtering. In addition, the equivalent transformation and sequential filtering techniques are used to significantly reduce the computational complexity of the estimator. Since the error of each phase observation is modeled separately, it is easy to handle when the reference satellite changes frequently. It solves the technical problem that in many RTK applications, there is inevitably a time delay between the observation data of the base station and the observation data of the meteor station, which leads to a rapid increase in RTK positioning errors or positioning failures.
[0113] As mentioned above, although the preferred embodiments of the present invention have been illustrated and described, many changes may be made without departing from the spirit and scope of the present invention. Thus, the scope of the present invention is not limited by the disclosure of the preferred embodiments. Instead, the present invention should be determined entirely with reference to the claims that follow.
Claims
1. A method for estimating and compensating age error in real-time motion positioning, comprising: Construct the satellite carrier phase observation equation at time t, use the double difference algorithm to process the carrier phase observation equation sent by satellite i and satellite j to the base station and mobile station, and extract the error caused by the age difference; The error caused by the age difference in the carrier phase observation equation is modeled and the estimated value expression is given; The equivalent transformation and sequential Kalman filtering are used to simplify the estimation of the error caused by the age difference in the carrier phase observation equation. And the simplified estimate of the error caused by the difference age is used to compensate for the difference age error in the double difference operation.
2. The method for estimating and compensating age error in real-time motion positioning according to claim 1, wherein: The carrier phase observation equation of satellite i at time t is: Among them, the subscript b represents a base station, and the subscript r represents a mobile station; φ b i (t) is the carrier phase observation value of satellite i at time t, in meters; ρ b i Indicates the distance between the satellite antenna phase center and the GNSS receiver antenna phase center, in meters; I i (t) and T i (t) represents the ionospheric delay and tropospheric delay at time t, respectively, in meters; dt b (t) and dt i (t) represents the receiver clock error and satellite clock error at time t, in seconds; δ b and δ i represent the receiver and satellite carrier phase hardware delays in meters, respectively; o i (t) is the error of the satellite orbit at time t, in meters; φ b and φ i represents the initial phase between the receiver and satellite i, in units of cycles; λ is the carrier phase wavelength in meters per cycle; N i is the carrier integer ambiguity, in weeks; σ b j is the observation error; c is the propagation speed of electromagnetic waves in space.
3. The method for estimating and compensating age error in real-time motion positioning according to claim 2, further comprising: using a double difference algorithm to process the carrier phase observation equations sent by satellite i and satellite j to the base station and the mobile station to obtain double difference observation values: in, The superscript ij represents the differential operation between satellites i and j, Δt represents the asynchronous time between the base station observation and the mobile station observation, and ρ rb ij 、N rb ij , ε rb ij are the distances between the satellite antenna phase centers and the GNSS receiver antenna phase centers of satellite i and satellite j relative to the base station and mobile station, the carrier integer ambiguity, and the observation covariance matrix, respectively. age ij (t; Δt) represents the error caused by age difference.
4. The method for estimating and compensating the age error in real-time motion positioning according to claim 3 further comprises: defining a function f i (t) = [T i (t)-I i (t)-cat i (t)+o i (t)], and we get e age ij The expression of (t; Δt) is as follows: Use Taylor series to expand the function f at time t i (t+Δt), and get its estimated value: in, α i (t) is f i (t) The first-order derivative at time t, retaining the first term of the Taylor series and ignoring the higher-order terms, obtains e age ij Estimated value of (t; Δt):
5. The method for estimating and compensating the time difference error in real-time motion positioning according to claim 4 further comprises defining the time difference carrier phase ratio in, The time difference carrier phase ratio is further expressed as: in, is the time-varying observation noise 6. The method for estimating and compensating the age error in real-time motion positioning according to claim 5, further comprising obtaining α by approximation i (t) and η r Approximate value of (t): in, T i and n α i is the correlation time of α and process noise, n α i It follows a zero-mean normal distribution, n α i ~N(0,qi),n ηr is the process noise of ηr, which follows a zero-mean normal distribution, n ηr ~N(0,qr).
7. The method for estimating and compensating for the age error in real-time motion positioning according to claim 6, further comprising defining the following state vector: x = [α 1 α 2 ... α N η r ] T ., get the approximate value of the state vector in, Where F and n t are the state transformation matrix and process noise vector respectively, n t It follows a zero-mean normal distribution, n t ~N(0,Q), Q is the covariance matrix, 8. The method for estimating and compensating for the age error in real-time motion positioning according to claim 7 further comprises defining an observation vector The observation function z=Hx+∈ is obtained, where the observation matrix H is The observation noise ∈ is a Gaussian distribution vector with diagonal covariance, ε ~ N(0,R), where R = diag(e σ1 e σ2 …e σN ).
9. The method for estimating and compensating the age error in real-time motion positioning according to claim 8 further comprises, according to the equivalent transformation principle, transforming the observation function into in, ∈′ is the first-order derivative of ∈, has the same static properties as ∈, and obeys the normal distribution with zero mean, ∈′~N(0,R), x′ is the first-order derivative of the state vector x, 10. The method for estimating and compensating age error in real-time motion positioning according to claim 9, further comprising: defining a differential matrix D of differential operations between the norm satellite and the reference satellite, The estimated differential age error is: Where D and x' are the differential matrices at the current epoch t and the estimated α, the final covariance of the double-differenced observation with forecast error uncertainty is: Where P is the prediction covariance matrix of the current epoch, Ω dd is the original double difference DD observation covariance matrix, that is, εrb ij ~N(0,Ω dd )。
