A pseudorange smoothing method for low-cost GNSS receivers

Abstract

The present invention proposes a pseudorange smoothing method suitable for low-cost satellite navigation and positioning system (GNSS) receivers. The method combines the Hatch filter and the Doppler smoothing method, takes into account the limitations of the two methods, and the performance of low-cost GNSS receivers, and establishes a complete set of pseudorange smoothing methods, which can meet the smoothing of pseudorange observation values ​​in complex scenes such as low, medium and high speed motion and tree shade, and improve the accuracy of pseudorange single-point positioning. The present invention is more suitable for pseudorange smoothing under conditions such as low receiver hardware cost, low hardware computing power, and low observation data quality.

Description

Technical Field

[0001] The present invention belongs to the technical field of global satellite navigation and positioning, and in particular, the present invention relates to a pseudorange smoothing method for a low-cost global satellite navigation and positioning system (GNSS) receiver. Background Art

[0002] Pseudorange point positioning is a method that uses the broadcast ephemeris of the global navigation satellite system (GNSS) and the pseudorange observations of a receiver to independently determine the absolute coordinates of the receiver in the earth coordinate system. It is also called standard point positioning (SPP). It has a wide range of navigation and positioning applications, such as mobile phone navigation, logistics tracking, shared bicycles, robots, and autonomous driving. However, due to the inherent characteristics of pseudorange, pseudorange observations contain large noise and are easily affected by multipath effects, which results in pseudorange point positioning only being able to achieve meter-level accuracy.

[0003] The principle of pseudo-range measurement is based on the pseudo-random noise code signals transmitted by GPS satellites. These signals arrive at the receiver antenna after a certain time △t from the satellite. In theory, △t multiplied by the speed of electromagnetic waves in a vacuum c can give the spatial geometric distance ρ from the satellite to the receiver. However, since the actual propagation time △t includes factors such as the error of the satellite clock and the receiver clock being out of sync, the satellite ephemeris error, the receiver measurement noise, and the delay error of the ranging code propagating in the atmosphere, the distance value obtained is not the true station-satellite geometric distance, so it is called pseudo-range.

[0004] Pseudorange single-point positioning is a navigation and positioning technology that uses pseudorange observations for calculation. It measures and tracks based on the phase changes of satellite carrier signals to improve positioning accuracy. This technology receives GNSS satellite signals through the antenna, amplifies and converts them through the RF front end, demodulates and measures the phase of the received signals, parses the satellite's navigation message, and then calculates the pseudorange between each satellite and the observation station. Since the original pseudorange observation value has a large error, smoothing is required to improve positioning accuracy. Pseudorange smoothing reduces the pseudorange error and improves the pseudorange accuracy by using the carrier phase of multiple times of continuous measurement. Finally, the smoothed pseudorange is combined with the known satellite position information and the positioning algorithm is used to perform positioning and calculation to obtain the position of the observation station, thereby achieving high-precision navigation positioning.

[0005] Hatch filtering is a method that uses carrier phase observations to smooth pseudorange observations. Carrier phase observations are highly accurate and not easily affected by multipath, and can achieve millimeter-level accuracy. However, carrier phase observations can have integer jumps, which can cause errors in the smoothed pseudorange. The Doppler observation smoothing pseudorange method replaces the carrier phase with the Doppler observation to avoid integer jumps. However, this method also has defects. In the signal tracking stage, the Doppler is initialized by the pseudorange observation. When there is large noise or multipath effect in the pseudorange observation, the Doppler observation is unreliable.

[0006] In actual situations, the combination of Hatch filtering and Doppler pseudorange smoothing methods can effectively eliminate the noise and multipath effects in the pseudorange observations, thereby improving the accuracy of pseudorange single-point positioning. Summary of the invention

[0007] In order to improve the quality of pseudo-range observations of low-cost GNSS receivers and improve the accuracy of pseudo-range single-point positioning, the present invention proposes a pseudo-range smoothing method suitable for low-cost GNSS receivers. The method combines the Hatch filter and the Doppler smoothing method, while taking into account the limitations of the two methods and the performance of low-cost GNSS receivers, and establishes a complete set of pseudo-range smoothing methods, which can meet the smoothing of pseudo-range observations in complex scenes such as low, medium and high speed motion and tree shade, and improve the accuracy of pseudo-range single-point positioning. The present invention is more suitable for pseudo-range smoothing under conditions such as low receiver hardware cost, low hardware computing power, and low observation data quality.

[0008] The technical solution of the present invention is as follows:

[0009] A pseudorange smoothing method for a low-cost GNSS receiver, characterized in that:

[0010] For the pseudorange and carrier equations at time k:

[0011]

[0012] In the formula,

[0013] P and L represent the pseudorange and carrier observation values ​​in meters, respectively.

[0014] ρ represents the geometric distance between the satellite and the station,

[0015] c is the speed of light,

[0016] t r represents the receiver clock error,

[0017] t s represents the satellite clock error,

[0018] dion represents the ionospheric delay,

[0019] dtrop represents the tropospheric delay,

[0020] λ represents the wavelength,

[0021] N is the integer ambiguity of the carrier phase,

[0022] ε and δ represent random errors;

[0023] represents the carrier observation value in weeks,

[0024] s stands for smooth, and

[0025] Pseudorange and carrier equations at time k+1:

[0026]

[0027] Get the pseudorange and carrier changes between epochs:

[0028]

[0029] It can be seen that the ambiguity parameters in the carrier observation equation are eliminated, and the only two parameters between △P and △L are the ionospheric delay change and the random error change.

[0030] Since the change in ionospheric delay in a short period of time is almost 0, it can be considered that △P is equal to △L, so △L is replaced by △P to obtain the smoothed pseudorange observation value at time k+1

[0031]

[0032] When time k is not the first epoch, the smoothed pseudorange at time k is used. Replace P in the above formula k , we get a recursive smoothing formula, namely Hatch smoothing filter:

[0033]

[0034] Taking into account the influence of the cumulative change of ionospheric delay, the improved Hatch smoothing filter weights the original pseudorange and the smoothed pseudorange on this basis to obtain the final smoothed pseudorange, which is expressed as follows:

[0035]

[0036] Here, M is called the smoothing factor, which reflects the weight of the two (original pseudorange and smoothed pseudorange). The larger M is, the greater the proportion of the smoothed pseudorange obtained by the carrier phase.

[0037] At the beginning of smoothing, the DOP value of the pseudo-range single point positioning in the previous epoch is used to determine whether the receiver is in an open or blocked environment. If the receiver is in a blocked environment, the single epoch smoothing method is used, that is, formula (4);

[0038] If the receiver is in an open environment, the Hatch filter smoothing method is used, that is, formula (6).

[0039] According to the present invention, weighting is performed according to 1 / M and (M-1) / M to achieve the weighted effect, wherein M is a smoothing factor, reflecting the weights of the two. The larger M is, the greater the proportion of the smoothed pseudorange obtained by the carrier phase is, and the smoother the final pseudorange is.

[0040] According to the present invention, the final form of smoothing is a combination of three methods: Hatch smoothing filtering, Doppler observation value smoothing, and single epoch smoothing.

[0041] Figure 2 The △P before and after smoothing is shown, where the blue line is before smoothing and the red line is after smoothing. It can be seen that the smoothed △P is more stable and the random error is controlled.

[0042] According to the present invention, a pseudorange smoothing method for a low-cost GNSS receiver is characterized by:

[0043] In the data collected by the receiver using a static measurement method, the value of M is 100-150.

[0044] Due to the differences in receiver manufacturers, various test environments, and the quality of carrier phase observations, the selection of the smoothing factor M needs to be debugged based on the actual observation data. By selecting the same receiver, testing static data received in static positioning mode at the same location and on different dates (static positioning means that the receiver is placed on the ground and does not move, and the same point is positioned multiple times), it is determined that the optimal M value for static data is 100-150.

[0045] According to the present invention, a pseudorange smoothing method for a low-cost GNSS receiver is characterized by:

[0046] After testing multiple sets of dynamic data, it was found that the faster the movement speed, the worse the quality of the carrier phase due to the Doppler effect. In this regard, the smoothing factor M is set in segments according to the current horizontal speed of the receiver.

[0047] First, obtain the receiver's east (E) and north (N) speeds at the previous moment, and then calculate the total horizontal speed:

[0048]

[0049] V H Indicates the total horizontal velocity of the receiver, V E Receiver east (E) speed, V N Receiver (N) direction speed, and according to V H The size of M is obtained.

[0050] A pseudorange smoothing method for a low-cost GNSS receiver according to the present invention is characterized in that:

[0051] The M value is obtained by setting multiple sets of M values, smoothing the same set of dynamic data to obtain multiple sets of positioning results, and then selecting a set of speed ranges.

[0052] Among the multiple groups of positioning results mentioned above, find the group of positioning results with the highest plane accuracy. The M selected by this positioning result is the optimal M value corresponding to the speed (the highest plane accuracy reflects the best smoothing effect).

[0053] It is worth noting that only the plane speed and accuracy are considered in this process, because the vertical accuracy is greatly affected by atmospheric delay and cannot reflect the actual positioning situation. In this regard, the vertical accuracy can be directly ignored, because the positioning results of the receiver are usually only displayed on the plane map, and the user does not perceive the vertical accuracy very well.

[0054] However, for situations where the atmospheric influence in the vertical direction is large, there are two ways to deal with it. The first is that the positioning observation data used to determine M must be collected at night. The reason is that the ionosphere is not active at night and has little effect on the vertical accuracy. The second is to use a dual-frequency ionospheric-free pseudorange combination, which can directly eliminate the influence of ionospheric delay, but this method requires the receiver to provide signals of two frequencies.

[0055] A pseudorange smoothing method for a low-cost GNSS receiver according to the present invention is characterized in that:

[0056] After testing multiple groups of dynamic data, it was found that by selecting 100 as the maximum value of M and dividing it according to the following intervals, a total of 10 M values ​​can be obtained. When the horizontal speed of the receiver is within a certain speed range, selecting the corresponding M value can achieve a better smoothing effect.

[0057]

[0058] According to the present invention, a pseudorange smoothing method for a low-cost GNSS receiver is characterized by:

[0059] When cycle slips and loss of lock occur when carrier phase smoothes pseudorange, there will be a large error in the carrier difference △L between epochs. If the error is greater than the carrier wavelength, Doppler observation value smoothing should be used instead.

[0060] Common cycle slip detection methods include: high-order difference, polynomial fitting, MW or GF combination method, and the third-order difference cycle slip detection method in the high-order difference method is preferred.

[0061] Cycle slips refer to the jump or interruption of the whole cycle count caused by the loss of satellite signal lock in the carrier phase measurement of the Global Navigation Satellite System (GNSS) technology. Correctly detecting and recovering cycle slips is one of the very important and must-solve problems in carrier phase measurement. After the satellite signal loses lock, there are two options: recapture and fast recovery.

[0062] According to a pseudorange smoothing method for a low-cost GNSS receiver of the present invention, the formula for calculating the carrier difference ΔL of k epochs from Doppler observations is as follows:

[0063]

[0064] Where △t represents the time difference between epoch k and epoch k-1;

[0065] D k-1 , D k denote the Doppler observations at epoch k-1 and epoch k respectively.

[0066] Unlike carrier phase, Doppler observations do not have cycle slips and loss of lock problems, so Doppler observations can be used to replace carrier phase for smoothing. It should be noted that Doppler cannot completely replace carrier phase for smoothing. First, the carrier phase accuracy is higher than Doppler. Second, when the pseudorange observation data is very poor, the numerical results of Doppler observations remain stable, but in fact the smoothed results are seriously distorted. The reason is that there is a strong correlation between pseudorange observations and Doppler observations during the generation process.

[0067] Therefore, Doppler can only be used as a supplement when the carrier phase cannot be used, and a stricter threshold limit is required. When Doppler smoothing is used, the difference between the pseudorange change and the Doppler change cannot be greater than 5 meters, and the difference between two adjacent Doppler changes cannot be greater than 1 week.

[0068] A pseudorange smoothing method for a low-cost GNSS receiver according to the present invention is characterized in that:

[0069] S5A: Determine whether to use carrier phase or Doppler observations for smoothing by the following processing steps:

[0070] S5A1: Doppler smoothing is used when the carrier phase is just initialized or there is no carrier observation value;

[0071] S5A2: If the triple difference cycle slip determination method is successful, carrier phase smoothing is used;

[0072] S5A3: If it is neither of the above two situations, it means that both carrier and Doppler observation values ​​exist, and it is necessary to further select the one with smaller error.

[0073] In this way, cycle slips as small as one cycle can be effectively detected, which is of great significance for the application of GNSS technology in various complex environments.

[0074] A pseudorange smoothing method for a low-cost GNSS receiver according to the present invention is characterized in that:

[0075] In step S5A3, if both the carrier and Doppler observation values ​​exist, it is necessary to further select the one with smaller error and enter the following S5B judgment:

[0076] S5B1: The carrier phase and Doppler are used to calculate the carrier difference △L between epochs respectively. If the difference between the two △L is less than 1m, it is considered that the carrier phase and Doppler accuracies are not much different. In this case, carrier phase smoothing is preferred.

[0077] S5B2: If the difference between the two △L mentioned above is greater than 1m, enter the following S5C judgment.

[0078] A pseudorange smoothing method for a low-cost GNSS receiver according to the present invention is characterized in that:

[0079] S5C1: Calculate the quadratic difference between epochs using carrier phase and Doppler Secondary difference Take the absolute value and select the one with the smaller absolute value to participate in smoothing.

[0080] This is because the result of the quadratic difference is usually a random error, and the smaller the random error, the higher the accuracy.

[0081] A pseudorange smoothing method for a low-cost GNSS receiver according to the present invention is characterized in that:

[0082] By judging the seconds in the current observation week, every certain time interval, for example, 2 14 1-2 16 Seconds, perform smooth pseudorange reset of high orbit (GEO) satellite, that is, convert the time of current observation data into seconds in week or computer time, and then reset the interval time, for example 2 14 1-2 16, perform the remainder operation, if the result is 0, it means that a reset is required.

[0083] Based on the basic principle of smoothing, we believe that the change in ionospheric delay is almost 0. However, if a satellite is observed continuously for a long time without any smoothing reset, the change in ionospheric delay will continue to accumulate, and there is a possibility of reducing positioning accuracy, especially in the GEO satellite data observed by the base station. It is necessary to perform a smooth pseudorange reset of the high-orbit (GEO) satellite.

[0084] Regarding the triple difference cycle slip judgment method, it is a technology that uses the high-order difference method to detect cycle slips. According to the present invention, the triple difference steps are as follows: each time pseudo-range smoothing is performed, the carrier epoch difference at the current moment is saved. After storing two consecutive carrier epoch differences, the two consecutive carrier epoch differences are subtracted to obtain the triple difference result;

[0085] At this time, the result of the third difference is analyzed to determine whether it is greater than the carrier wavelength. When there is no cycle slip, the result of the third difference is a very small random value, usually less than the carrier wavelength.

[0086] Several situations:

[0087] (1) Timing smooth reset of high-orbit (GEO) satellites

[0088] Based on the basic principle of smoothing, we believe that the change in ionospheric delay is almost zero. However, if a satellite is observed continuously for a long time without any smoothing reset, the change in ionospheric delay will continue to accumulate, which may reduce the positioning accuracy, especially in the GEO satellite data observed by the base station.

[0089] Therefore, by judging the seconds in the current observation week, every certain time interval, for example, 2 14 1-2 16 seconds, the smoothed pseudorange of the GEO satellite needs to be reset.

[0090] (2) Continuous time judgment

[0091] The time difference is obtained by calculating the current week seconds and the last saved week seconds. If the time difference is less than the threshold (the threshold is the sampling interval of the receiver plus 1 millisecond, and 1 millisecond is the computer rounding error), the epoch is considered to be continuous. Otherwise, smoothing should be rejected and the end stage should be entered to save various types of data and output pseudorange observation values.

[0092] If the time discontinuity indicates that the data has been interrupted, smoothing cannot be performed because it is impossible to determine what happened to the device during the interruption. Therefore, the pseudoranges before and after the interruption cannot be subtracted. Once an interruption occurs, pseudorange smoothing must be restarted from the interruption.

[0093] (3) Pseudorange quality judgment

[0094] When the absolute value of the pseudorange difference △P between epochs is greater than a certain threshold, such as 1000 meters, or the smoothed pseudorange at the previous moment is 0, it means that there is an abnormality in the pseudorange observation value, and the end stage is entered, various data are saved, and the pseudorange observation value is output.

[0095] (4) Carrier phase or Doppler quality judgment

[0096] When carrier phase is used for smoothing, when the difference between the pseudorange difference △P between epochs and the carrier difference △L is greater than a certain threshold, such as 100 meters, smoothing should be rejected. Because the difference between the two is mainly reflected in random errors, normal random errors will not be large. When Doppler is used for smoothing, when the difference between the pseudorange difference △P between epochs and the carrier difference △L is greater than a certain threshold, such as 5 meters or Doppler calculates the secondary difference between epochs. When the absolute value is greater than 1 meter, smoothing should be rejected, the end stage should be entered, various data should be saved, and pseudorange observation values ​​should be output.

[0097] (5) Single epoch smoothing

[0098] The Hatch smoothing method is a recursive smoothing method. When the observation data quality is good, a good smoothing effect can be obtained. However, smoothing in consecutive poor observations will accumulate a large amount of errors and affect the positioning accuracy. The quality of the observations can be judged by the DOP value. The larger the DOP value, the worse the observation. For poor observation data, on the one hand, the smoothing factor M of Hatch can be reduced to 2, and on the other hand, single epoch smoothing can be used.

[0099] Formula (4) is the basic formula for single epoch smoothing. The smoothed pseudorange of the current epoch is obtained by adding the carrier difference of the current epoch to the original pseudorange of the previous epoch. This method reduces the smoothness to avoid the error accumulation problem in recursive smoothing, so it is very important.

[0100] A pseudorange smoothing method for a low-cost GNSS receiver according to the present invention is characterized in that:

[0101] When the current smoothing mode is judged as "single epoch smoothing" by the DOP value, if Doppler smoothing is used, the threshold between the inter-epoch pseudorange difference △P and the carrier difference △L is relaxed to 20 meters.

[0102] This is because there is a correlation between Doppler and pseudorange observations. When the quality of pseudorange observations is very poor, Doppler observations appear to be very smooth numerically, but in fact the accuracy will also be reduced, so the rejection threshold should also be relaxed.

[0103] By analyzing a large amount of low-quality observation data, it is found that when the PDOP value (position positioning precision factor, obtained by the square root of the sum of squares of errors in latitude, longitude and elevation) of the previous epoch or the current epoch is greater than 1.8 or the HDOP (horizontal position precision coefficient) is greater than 1.5 or the VDOP (vertical position precision coefficient) is greater than 1.5, the single epoch smoothing method is more suitable for this smoothing.

[0104] (6) Initialization pseudorange during smooth reset

[0105] When resetting smoothing, the original pseudorange observation value of the current epoch is usually used for initialization. However, when the pseudorange observation value has significant errors, it takes a long time to smooth. Therefore, it is necessary to consider using a higher-precision pseudorange to replace the original pseudorange. The predicted pseudorange is a high-precision pseudorange that does not contain pseudorange random errors by calculating the station-satellite geometric distance, satellite clock error, atmospheric delay, and then adding the receiver clock error of the previous epoch. When resetting smoothing, the predicted pseudorange can be used to replace the original pseudorange.

[0106] However, this method will also introduce errors. Obviously, the atmospheric delay correction in the predicted pseudorange is different from the original pseudorange, because the atmospheric delay calculated by the model usually only accounts for 40% to 70% of the actual atmospheric delay, and there is also a small gap between the receiver clock error obtained by the last pseudorange single-point positioning and the original pseudorange. Therefore, although this method can eliminate random errors in the original pseudorange, it will introduce two errors, the receiver clock error and the atmospheric delay. Since the purpose of the replacement is to eliminate large errors in the initial pseudorange, it is only necessary to consider replacement when there are large errors. Through a large number of tests, it is found that it is preferred that the predicted pseudorange and the original pseudorange differ by more than ten to twenty meters under normal circumstances, so a threshold is set to judge that when the difference between the predicted pseudorange and the original pseudorange is greater than 30 meters, the replacement required for initializing the pseudorange during smoothing reset is performed. This threshold needs to be measured and debugged for specific receiver data to obtain appropriate results.

[0107] Effects of the Invention

[0108] Improve the quality of pseudorange observations and the accuracy of pseudorange single-point positioning; adopt a strategy combining Hatch filtering and single epoch smoothing method to achieve high-precision smoothing of pseudorange observations in complex scenarios; in the case of cycle slips or missing carrier observations, Doppler observations are used to supplement them, ensuring the continuity and reliability of the pseudorange smoothing process.

[0109] Positioning accuracy:

[0110] The smoothed pseudorange is used for single-point positioning. The single-point positioning accuracy of the pseudorange of static data is improved by 27%-35% in the horizontal direction and 7%-11% in the vertical direction. In dynamic data, the dynamic trajectory after smoothing is more consistent with the actual situation than the trajectory before smoothing, and the positioning error is significantly reduced.

[0111] Algorithm efficiency

[0112] The algorithm in the present invention mainly adopts the method of recursion and judgment to realize pseudorange smoothing, avoiding matrix operations and a large number of loop operations. While ensuring the positioning accuracy, the algorithm is more efficient and has low requirements on hardware equipment. It is generally applicable to various types of GNSS receivers, especially in low-cost GNSS receivers, and can also achieve good positioning effects.

[0113] cost:

[0114] The present invention improves the accuracy of pseudorange observations in low-cost GNSS receiver chips by pseudorange smoothing. In an open environment, Hatch filtering is used for pseudorange smoothing, and the Hatch filter weight factors at different speeds are obtained through a large amount of data simulation. In an occluded environment, the accuracy of the carrier phase and Doppler are compared, and smoothing is performed preferentially. For severely occluded environments, a single epoch smoothing mode is used to avoid error accumulation. Tests show that the technical solution of the present invention can improve the quality of pseudorange observations of low-cost GNSS receivers and enhance the accuracy of pseudorange single-point positioning in occluded environments and under low, medium and high speed motion. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] Figure 1 This is a detailed flow chart of the pseudorange smoothing method.

[0116] Figure 2 Schematic diagram of the inter-epoch pseudorange variation before smoothing (upper and lower broken lines) and after smoothing (middle horizontal line).

[0117] Figure 3 A schematic diagram of the process summary.

[0118] Figure 4-A to Figure 4-D A planar error plot showing static data before (peripheral area) and after smoothing (central area).

[0119] Figure 5-A to Figure 5-B The planar trajectory diagrams before and after smoothing are shown in the scene of walking around a building.

[0120] Figure 6-A to Figure 6-B The planar trajectory diagrams before and after smoothing in the scene of walking around an open space are shown respectively.

[0121] Figure 7-A Figure 7-B shows the planar trajectory diagrams before and after smoothing at the intersection of the driving scene. DETAILED DESCRIPTION

[0122] Example

[0123] according to Figure 1 The steps shown are used to describe in detail a pseudorange smoothing method suitable for a low-cost satellite navigation and positioning system (GNSS) receiver of the present invention.

[0124] Step 1: Collect the original observation data of the receiver in different scenarios. The specific operations are as follows:

[0125] (1) Static data collection. The equipment with PCB board, power supply equipment and mushroom-shaped antenna was placed on the open roof. The PCB board used ASR5301 satellite navigation receiver chip from Aojie Technology. The chip has low procurement cost but can support both GPS and Beidou systems. After the equipment was placed, the equipment was no longer moved. The observation was continued for 3 days to complete the static data collection.

[0126] (2) Walking data collection. The collector puts the device with PCB board, power supply equipment and mushroom-head antenna pole into a backpack and fixes it. The mushroom-head antenna needs to protrude from the backpack and ensure that its height is higher than the height of the collector so that there is no obstruction around the antenna. The collector completes data collection by walking a rectangular and triangular trajectory at a normal walking speed in an open environment.

[0127] (3) Driving data collection. The collector installed the mushroom-shaped antenna on the roof of the car, placed the PCB board and power supply equipment inside the car, and drove the car along Longdong Avenue and the Inner Ring Elevated Road to simulate normal vehicle driving and complete data collection.

[0128] Step 2: Select the corresponding smoothing method according to the observation conditions. The specific operation is as follows:

[0129] Analyze the DOP value in the pseudo-range single point positioning of the previous epoch. When the HDOP value is greater than 1.8 or the VDOP value is greater than 1.5, the smoothing method of this epoch is marked as the single epoch smoothing method, otherwise it is marked as Hatch filtering smoothing. When marked as Hatch filtering, the corresponding Hatch smoothing factor M needs to be matched according to the speed of the receiver in the previous epoch and the results in the table.

[0130] Step 3: Determine whether smooth reset is required at the current moment. The specific operations are as follows:

[0131] When converting the current observation data time to the computer, set the reset time interval to 2 16 seconds, then the time of the observation data is 2 16 Perform a remainder operation. If the result is 0, it means that it is divisible exactly. In this case, a reset is required and the reset parameter is marked as 1. If the result is not zero, the reset parameter is marked as 0.

[0132] Step 4: Calculate the inter-epoch pseudorange difference, time difference, carrier phase difference and carrier phase difference calculated by Doppler (hereinafter referred to as Doppler difference) of each satellite. The specific operation is as follows:

[0133] First, the wavelength of the carrier is obtained by dividing the speed of light by the signal frequency. Subsequently, the pseudorange difference is the pseudorange of the current epoch minus the original pseudorange of the previous epoch saved; the time difference is the time of the current epoch minus the time of the previous epoch saved; the carrier phase difference is the carrier phase of the current epoch minus the carrier phase of the previous epoch saved, and multiplied by the wavelength to obtain the carrier phase difference in meters; the Doppler difference is obtained by the mean smoothing method, that is, the Doppler of the current epoch and the Doppler of the previous epoch are averaged, then multiplied by the time difference, and finally multiplied by the wavelength to obtain the carrier phase difference calculated by Doppler in meters.

[0134] Step 5: Determine whether the satellite meets the conditions for pseudorange smoothing. If all conditions are met, proceed to the next step. Otherwise, set the smoothing mark to zero, skip the subsequent steps, and directly enter the end stage. The specific operations are as follows:

[0135] (1) Time continuity judgment: Compare the time difference in step 4 with the time difference saved in the pseudorange single point positioning process. If the difference between the two is less than 1 millisecond, it means that the time difference in step 4 is continuous and meets the smoothing condition. Otherwise, there is a data interruption and the smoothing condition is not met.

[0136] (2) Determine the correctness of the pseudorange difference. Determine the absolute value of the pseudorange difference in step 4. If the absolute value is less than the empirical value of 1000 meters, it means that the pseudorange difference is normal and meets the smoothing condition. Otherwise, there may be a large gross error in the pseudorange and the smoothing condition is not met.

[0137] (3) First smoothing judgment: If it is the first time to enter smoothing, the original pseudorange of the last epoch saved should be 0. If this happens, the smoothing condition is met.

[0138] (4) Smooth reset judgment of GEO satellite: If the smooth reset flag in step 3 is 1 and the current satellite is a Beidou GEO satellite, the smoothing condition is not met, otherwise the smoothing condition is met.

[0139] Step 6: Determine whether to use carrier phase difference or Doppler difference for smoothing. The specific operations are as follows:

[0140] (1) Doppler must be used for smoothing judgment.

[0141] When the LLI marker detects a cycle slip in the carrier phase, or the carrier phase observation value of the current epoch is missing, Doppler smoothing must be used, and the smoothed data used will be marked as Doppler difference.

[0142] (2) The carrier phase must be used for smoothing judgment.

[0143] When the carrier phase observation value passes the triple difference cycle slip detection method, it must be smoothed using the carrier phase, and the smoothed data used is marked as the carrier phase difference.

[0144] (3) Further judgment in non-necessary situations.

[0145] When the difference between the carrier phase difference and the Doppler difference is not large, that is, the difference is less than 1 meter, the carrier phase is usually more accurate and has priority, so the smoothed data used is marked as the carrier phase difference. Otherwise, it means that there is a significant difference in accuracy between the carrier phase difference and the Doppler difference, and the one with higher accuracy needs to be selected for smoothing.

[0146] At this time, the quadratic difference of the carrier phase and Doppler is calculated respectively. The quadratic difference can reflect the size of the random error between the two. When the quadratic difference of the carrier is smaller than the quadratic difference of the Doppler, the carrier accuracy is higher, and the smoothed data used is marked as the carrier phase difference, otherwise the smoothed data used is marked as the Doppler difference.

[0147] Step 7: Evaluate the availability of the selected carrier phase difference or Doppler difference. The specific operations are as follows:

[0148] When carrier phase difference smoothing is used, if the difference between the carrier phase difference and the pseudorange difference is greater than 100 meters, the smoothing mark is set to zero, the subsequent steps are skipped, and the end stage is directly entered.

[0149] When Doppler difference smoothing is used, if the difference between the Doppler difference and the pseudorange difference is greater than 20 meters, and the quadratic difference of the Doppler difference is greater than 1 meter, the smoothing mark is set to zero, the subsequent steps are skipped, and the end stage is entered directly.

[0150] Step 8: Calculate the smoothed pseudorange. The specific operations are as follows:

[0151] When the smoothing method is marked as Hatch filtering, according to Formula 6, the corresponding parameters are substituted into the formula to obtain the smoothed pseudorange observation value.

[0152] When the smoothing method is marked as single epoch smoothing, according to Formula 4, the corresponding parameters are substituted into the formula to obtain the smoothed pseudorange observation value.

[0153] Step 9: At the end stage, save all kinds of data and output pseudorange observation values. The specific operations are as follows:

[0154] Replace the corresponding parameters of the previous epoch with the time, smoothed pseudorange, raw pseudorange, carrier phase, Doppler, carrier phase quadratic difference in meters, carrier phase difference in meters, and Doppler difference in meters of the current epoch.

[0155] When the smoothing flag is 1, and the difference between the current smoothed pseudorange and the original pseudorange is less than 100 meters, the smoothed pseudorange will replace the original pseudorange and output for subsequent pseudorange positioning and solution. Otherwise, the original pseudorange will be directly output without any change for subsequent pseudorange positioning and solution.

[0156] Step 10: Use the output pseudo-range observations for pseudo-range positioning, and use the data without pseudo-range smoothing for pseudo-range positioning as a comparison. For the data in three scenarios, static data, walking data, and driving data, the positioning results are shown in Figures 4 to 7. It can be seen that after pseudo-range smoothing, the plane positioning error of static data is significantly reduced, and the motion trajectory in the walking and driving trajectory diagrams is smoother and closer to the actual situation.

Claims

1. A pseudorange smoothing method for low-cost GNSS receivers, characterized in that ； For the pseudorange and carrier equations at time k: In the formula, P and L represent pseudorange and carrier observation values, respectively, in m; r represents the geometric distance between the satellite and the station, in meters; c represents the speed of light, in m / s; t r Indicates the receiver clock error, in seconds; t s Represents the satellite clock error, unit is s; dion represents ionospheric delay, unit is m; dtrop represents the tropospheric delay, in meters; λ represents wavelength, unit is cm; N represents the integer ambiguity of the carrier phase, in units of weeks; ε and δ represent random errors; unit: m; Indicates the carrier observation value, unit is week; s stands for smooth, and Pseudorange and carrier equations at time k+1: Get the pseudorange and carrier changes between epochs: The ambiguity parameters in the carrier observation equation are eliminated, and the only two parameters that differ between △P and △L are the ionospheric delay change and the random error change. Since the change in ionospheric delay in a short period of time is almost zero, it can be considered that △P is equal to △L. △L is replaced by △P to obtain the smoothed pseudorange observation value at time k+1 When time k is not the first epoch, the smoothed pseudorange at time k is used. Replace P in the above formula k We get a recursive smoothing formula, namely Hatch smoothing filter: Taking into account the influence of the cumulative change of ionospheric delay, the improved Hatch smoothing filter weights the original pseudorange and the smoothed pseudorange on this basis to obtain the final smoothed pseudorange, which is expressed as follows: Where M is called the smoothing factor, which reflects the weight of the original pseudorange and the smoothed pseudorange. The larger M is, the greater the proportion of the smoothed pseudorange obtained by the carrier phase. At the beginning of smoothing, the DOP value of the pseudo-range single point positioning in the previous epoch is used to determine whether the receiver is in an open or blocked environment. If the receiver is in a blocked environment, the single epoch smoothing method is used, that is, formula (4); If the receiver is in an open environment, the Hatch filter smoothing method is used, that is, formula (6).

2. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 1, characterized in that: In the data collected by the receiver using a static measurement method, the value of M is 100-150.

3. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 1, characterized in that: According to the dynamic data of the receiver, the current horizontal speed is segmented to set the smoothing factor M; First, obtain the receiver's east (E) and north (N) speeds at the previous moment, and then calculate the total horizontal speed: V H Indicates the total horizontal velocity of the receiver, in m / s, V E Receiver east (E) speed, m / s, V N Receiver (N) direction speed, in m / s, and based on V H The size of M is obtained.

4. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 1 or 3, characterized in that: The M value is obtained by setting multiple sets of M values, smoothing the same set of dynamic data to obtain multiple sets of positioning results, and then selecting a set of speed ranges. Among the multiple groups of positioning results mentioned above, find the group of positioning results with the highest plane accuracy. The M selected by this positioning result is the optimal M value corresponding to the speed.

5. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 4, characterized in that: Select 100 as the maximum value of M, and divide it according to the following intervals to get a total of 10 M values.

6. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 1, characterized in that: If cycle slips and loss of lock occur when carrier phase is used to smooth pseudorange, there will be a large error in the carrier difference △L between epochs. If the error is greater than the carrier wavelength, carrier phase pseudorange smoothing cannot be performed and Doppler observation smoothing can be used instead.

7. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 6, characterized in that: The formula for calculating the carrier difference △L at epoch k from Doppler observations is as follows: Where △t represents the time difference between epoch k and epoch k-1, unit is s; Dk-1, Dk represent the Doppler observation values ​​between k-1 epoch and k epoch respectively, in units of weeks.

8. The pseudorange smoothing method for a low-cost GNSS receiver according to claim 1, characterized in that: The steps are as follows, S5A: Determine whether to use carrier phase or Doppler observations for smoothing by the following processing steps: S5A1: Doppler smoothing is used when the carrier phase is just initialized or there is no carrier observation value; S5A2: If the triple difference cycle slip determination method is successful, carrier phase smoothing is used; S5A3: If it is neither of the above two situations, it means that both the carrier and Doppler observation values ​​exist, but it is unclear which error is larger. Go to S5B below for judgment.

9. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 8, characterized in that: In step S5A3, if both the carrier and Doppler observation values ​​exist, but it is unclear which one has higher accuracy, the process proceeds to step S5B to determine: S5B1: The carrier phase and Doppler are used to calculate the carrier difference △L between epochs respectively. If the difference between the two △L is less than 1m, it is considered that the carrier phase and Doppler accuracies are not much different. In this case, carrier phase smoothing is preferred. S5B2: If the difference between the two △L is greater than 1m, enter S5C1 for judgment.

10. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 9, characterized in that: S5C1: Calculate the quadratic difference between epochs using carrier phase and Doppler Secondary difference Take the absolute value and select the one with the smaller absolute value to participate in smoothing.

11. The pseudorange smoothing method for a low-cost GNSS receiver according to claim 1, characterized in that: By judging the seconds in the current observation value, every certain time interval, 2 14 1-2 16 seconds, reset the smooth pseudorange of the high orbit (GEO) satellite, that is, convert the time of the current observation data into seconds in a week or computer time, and then reset the interval time, 2 14 1-2 16 , perform the remainder operation, if the result is 0, it means that a reset is required.

12. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 1 or 11, characterized in that: Continuous time judgment: The time difference is obtained by calculating the current week seconds and the last saved week seconds. If the time difference is less than the sampling interval of the receiver, the epochs are considered to be continuous.

13. The pseudorange smoothing method for a low-cost GNSS receiver according to claim 1, characterized in that: Pseudorange quality judgment: When the absolute value of the pseudorange difference △P between epochs is greater than the threshold of 1000 meters, or the smoothed pseudorange at the previous moment is 0, it means that the pseudorange observation value is abnormal, smoothing is rejected, and the end stage is entered. All kinds of data are saved and the pseudorange observation value is output.

14. The pseudorange smoothing method for a low-cost GNSS receiver according to claim 1, characterized in that: Carrier phase determination: When carrier phase is used for smoothing, when the difference between the pseudorange difference △P between epochs and the carrier difference △L is greater than the threshold of 100 meters, smoothing is rejected and the end stage is entered. Various types of data are saved and pseudorange observation values ​​are output.

15. The pseudorange smoothing method for a low-cost GNSS receiver according to claim 7, characterized in that: Doppler quality judgment: When Doppler smoothing is used, when the difference between the pseudorange difference △P between epochs and the carrier difference △L is greater than the 5-meter threshold, or the Doppler calculation of the secondary difference between epochs When the absolute value is greater than 1 meter, smoothing is rejected and the end stage is entered. Various data are saved and pseudo-range observation values ​​are output.

16. A pseudorange smoothing method for a low-cost GNSS receiver as claimed in claim 15, characterized in that: When the current smoothing mode is judged as "single epoch smoothing" by the DOP value, if Doppler smoothing is used, the threshold between the inter-epoch pseudorange difference △P and the carrier difference △L is relaxed to 20 meters.

17. The pseudorange smoothing method for a low-cost GNSS receiver according to claim 1, characterized in that: For the Hatch smoothing method, if the observation data is continuously poor, the Hatch smoothing factor can be reduced to 2, and / or single epoch smoothing can be used.

18. The pseudorange smoothing method for a low-cost GNSS receiver according to claim 1, characterized in that: Initialization pseudorange during smoothing reset, The predicted pseudorange is a high-precision pseudorange that does not contain random pseudorange errors. It is obtained by calculating the station-satellite geometric distance, satellite clock error, atmospheric delay, and then adding the receiver clock error of the previous epoch. The predicted pseudorange can be used to replace the original pseudorange during smoothing reset. When the difference between the predicted pseudorange and the original pseudorange is greater than 30 meters, the replacement required to initialize the pseudorange is performed during smoothing reset.

Images