CN115265555B - Map matching correction method and system based on hidden Markov multi-noise perception - Google Patents

Abstract

The invention provides a map matching correction method and a system based on hidden Markov multi-noise perception, which are used for obtaining a smooth track by gridding an area; inputting the smoothed track into a hidden Markov model for matching; calculating fluctuation conditions of observation points in the matching process; deleting the current observation point when the fluctuation amplitude exceeds a threshold value; deleting continuous track points in the adjacent area, predicting the positions of the deleted points, evaluating the error range of the predicted positions and the observed positions, and judging that the map error exists at the section of sampling track when the error is smaller than a threshold value; and taking the deleted points as the input of a map generator, dividing the deleted road segments into a series of sections according to the calculated inflection point positions, and calculating the line segments passing through the inflection points in the sections. The invention creatively provides a map matching framework capable of detecting map errors and deleting low-quality sampling points, which can be combined with a map matching algorithm based on a hidden Markov model to correct matching results and improve matching accuracy.

Description

Map matching correction method and system based on hidden Markov multi-noise perception

Technical Field

The invention relates to the technical field of intelligent traffic, in particular to a map matching correction method and system based on hidden Markov multi-noise perception. In particular, it relates to a map matching and correction framework based on a hidden Markov model for multi-noise perception.

Background

In recent years, with the rapid development and wide application of online taxi taking platforms, various companies have introduced their own taxi taking apps, and meanwhile, GPS devices are equipped for almost all taxis to obtain information such as the position and direction of the vehicle. Many functions such as path navigation, location and route prediction and identification, detection of abnormal behavior, etc. of taxis require accurate location and tracking.

Most of early map matching algorithm researches are based on geometric similarity, and the method cannot deal with the situation of large sampling error. Many matching algorithms based on hidden Markov models, which are proposed later, define transition probabilities by means of limitations of the road network itself, so that the situation that some sampling errors are large and wrong road sections are selected is avoided. However, these algorithms lack consideration of complexity of the road network and simultaneous occurrence of sampling errors, which results in incorrect matching results such as reverse motion. Meanwhile, the real road network is updated continuously, and the real road cannot be selected due to the lack of road sections when the maps are matched.

In addition to the hidden Markov model, there have been other improvements in map matching algorithms in recent years. With the introduction of more indexes, some algorithms based on a scoring model have good effects. For example, by classifying each GPS point by a support vector machine and rejecting low quality sampling points, etc. In general, various algorithms have sufficient matching accuracy for practical scenarios with low sampling errors or in line with road network constraints.

The Chinese patent document with publication number CN113639757A discloses a map matching method and system based on a bidirectional scoring model and a backtracking correction mechanism, and the map matching method and system comprises the following steps: selecting candidate points according to the map path information based on the acquired GPS point position information; scoring the positions, the directions and the speeds of the measured GPS points based on a bidirectional scoring model, and giving weights with different positions, directions and speeds to obtain the scores of the candidate points on the GPS points; when the score of the GPS point is lower than the threshold value, determining the GPS point as a low-quality point, and deleting the current low-quality point not to participate in matching; when the continuous GPS points are judged to be low-quality points and deleted, reversely evaluating the deleted GPS points by utilizing the first GPS point which is not deleted subsequently, and detecting whether the deleted GPS points are low-quality points again; calculating the matching probability of each candidate point and the currently reserved GPS point based on the bidirectional scoring model, and selecting the candidate point with the maximum probability value as the matching candidate point; and generating a unique map matching result based on the shortest path principle according to the matched candidate points.

For the related art, the inventor considers that the matching results of errors such as interruption or detour often occur in practical application due to the fact that map errors are ignored and sampling errors are large in the currently proposed algorithm.

Disclosure of Invention

Aiming at the defects in the prior art, the invention aims to provide a map matching correction method and system based on hidden Markov multi-noise perception.

The map matching correction method based on the hidden Markov multi-noise perception comprises the following steps:

a region windowing step: by gridding the region, the irregular drift of the track is limited, and a smooth track is obtained;

Map matching step: inputting the smoothed track into a hidden Markov model with a deleting point and a map correcting frame for matching;

Probability fluctuation detection step: calculating fluctuation conditions of the Viterbi decoding probability of each observation point in the matching process; if the fluctuation amplitude exceeds the threshold value, deleting the current observation point;

Track evaluation step: if the continuous track points are deleted in the adjacent area, predicting the deleted point positions by using a Kalman filter and the reserved track points, evaluating the error range of the predicted positions and the observed positions, and if the error is smaller than a threshold value, judging that the map error exists at the section of sampling track;

a missing road section generation step: and taking the deleted points as the input of a map generator, dividing the deleted road sections into a series of sections according to the calculated inflection point positions, and calculating the line segments passing through the inflection points by using a linear regression model in the sections.

Preferably, in the area windowing step, the geographical area is first divided into grid cells;

If the vehicle does not spend a predetermined travel time in the grid cell, the vehicle is not allowed to transition from the grid cell;

In each grid cell, a variation of the DP algorithm is used to smooth the trajectory, replacing the original trajectory with an approximate line segment;

if the replacement does not meet the specified error requirement, recursively dividing the original problem into two sub-problems by taking the selected position point as a dividing point, and judging whether the replacement meets the specified error requirement or not until the error between the approximate track and the original track is lower than a specified error threshold;

The DP algorithm uses the orthogonal Euclidean distance as an error measure to keep the direction trend in the approximate track;

by removing partition points whose distance is greater than a threshold value, speed variation in the grid cells is limited.

Preferably, in the map matching step, in the map matching stage, the smooth GPS points acquired from the region windowing step are taken as input, the smooth GPS points are mapped to the road segments on the digital map, and the road segment sequence corresponding to the time stamp sequence of the GPS raw data is determined using a hidden markov model;

The hidden markov model is a discrete time markov process with candidate states and observations;

searching candidate states using an index grid, each candidate state being a point closest to an observation point in a road segment, and generating an observation probability The observation probability/>Is defined as a GPS sampling point p _i and a candidate point/>, calculated based on a Gaussian noise modelPossibility of matching:

Wherein σ is the standard deviation and d is defined as p _i and The projection distance between the two points, p _i is the sampling point at the moment i,/>Is the j-th candidate point of p _i;

Wherein, Is p _i th candidate road segment, c is edge/>At any point above, dist (p _i, c) is the Euclidean distance between p _i and c;

the hidden markov model allows a transition from a previous candidate state to a current candidate state, and the transition is controlled by a transition probability;

empirically, the hidden Markov model assumes that the shortest path between successive candidate states has transition probabilities

Thus, transition probabilityThe calculation is as follows:

Wherein, Is a candidate point of current observation; /(I)Is a candidate point for the previous observation; d is from/>To/>An absolute value of a difference between the shortest path length and the euclidean distance between observations;

and then, using an online Viterbi decoding algorithm to find the maximum likelihood sequence of the road section corresponding to the variable sliding window of the input sequence.

Preferably, in the probability fluctuation detection step, a hierarchical graph is gradually constructed by defining an observation probability and a transition probability and adding new nodes and weighted edges after receiving GPS sampling points;

If the matching accuracy is reduced when a new GPS sample is added in the layered graph, the measurement error or map error is considered to exist;

Given the candidate point set, the hidden Markov model will calculate the maximum likelihood path using the Viterbi decoding algorithm;

When the candidate set contains the observed real position, the hidden Markov model deduces the correct path;

when there is a map error or a measurement error, the shortest path between two consecutive candidates is not a real path, and may detour;

taking the acquisition points of which the candidate set does not contain the real position as low-quality points, and deleting the detected low-quality points by a detection model;

The calculation process of the Viterbi algorithm is to maximize the product of the observation probability and the transition probability when a new GPS sampling point is added;

the sampling error or map error causes amplitude fluctuation of the optimal decoding probability;

taking the logarithm of the original value to make the fluctuation more obvious;

deleting the newly added point when the fluctuation of the matching probability caused by the newly added point sampling exceeds a preset threshold value;

Limiting the number of consecutive removal points when there is a map error;

When a series of consecutive sampling points for a region is removed beyond a limit number, the algorithm evaluator will evaluate the unmatched samples, detect if there is a map error, and if so, use a map generation model to fit the missing road.

Preferably, in the track evaluation step, a series of GPS points continuously removed from the detection model are taken as input, and the quality of the entire sequence is evaluated to determine the reason for the removal;

evaluating the series of motion patterns under position and velocity constraints that remove GPS samples;

if the detected abnormality is caused by a motion pattern that does not conform to the actual driving situation, then the change in the matching probability is considered to be caused by a measurement error, and the corresponding point is deleted; the rest points are added into the matching sequence again to continue matching;

Restoring the removed GPS sampling points and generating a missing road by using points which cannot be matched if the removed points are allowed to form a driving mode in the evaluation model;

Predicting the movement direction and position of the vehicle by using Kalman filtering, and then judging whether an abnormality exists by taking the estimated points as new matching points;

Taking the speed and direction information of the road section adjacent to the sampling point as the input of a motion trail physical model, calculating the ideal position of the vehicle at the next moment, and taking the sampled trail direction and speed information as the input of the motion model to calculate the position at the next moment as a measured value;

based on the error between the weighted estimated value and the true value of the system, the Kalman gain is calculated, and the position estimated value at the next moment is obtained as follows:

x_t|t＝K_tz_t+(I-K_tH)x_t|t-1

Wherein x _t|t is a system estimation value, x _t|t-1 is a system ideal state value, z _t is a measurement value, K _t is Kalman gain, I is an identity matrix, and H is a state space conversion matrix;

Comparing the position predicted value with the position sampling value, calculating the average distance between the position predicted value and the position sampling value, and if the distance is smaller than the threshold value delta d, considering the sampling to be caused by map errors.

Preferably, in the missing road section generating step, the digital map is composed of an attributed undirected graph; the intersection points are represented as vertices with a degree greater than two, and the road segments are represented as a series of edges; treating the bi-directional link as a single path; defining high-order attributes to supplement the graph; the geometric features are displayed as part of the map description; providing local geometry information about the road;

automatically deducing road and traffic characteristics by using a map generation model based on a clustering technology;

Fitting the GPS sampling points which cannot be matched by the map generation model, wherein the generated topological structure is close to an actual missing road;

dividing the GPS points which cannot be matched into different sections according to the positions of the inflection points;

The distance of the inflection point is measured by a euclidean distance variable of the vertical euclidean distance;

at each segment, fitting the missing road using a linear regression model;

In order to determine the position of the turning point, parameters θ _G and d _G are set;

connecting a first point and a last point in the sequence, and then projecting all other points to the connecting line;

If the specified projection distance is greater than the threshold d _G, the point with the specified distance is a corner point of the real road segment;

connecting points with specified projection distances with a first point and a last point respectively, and continuously calculating angles between the connecting points; if the angle is greater than the threshold θ _G, the sequence is split into two parts and iterated.

The invention provides a map matching correction system based on hidden Markov multi-noise perception, which comprises the following modules:

region windowing module: by gridding the region, the irregular drift of the track is limited, and a smooth track is obtained;

And a map matching module: inputting the smoothed track into a hidden Markov model with a deleting point and a map correcting frame for matching;

Probability fluctuation detection module: calculating fluctuation conditions of the Viterbi decoding probability of each observation point in the matching process; if the fluctuation amplitude exceeds the threshold value, deleting the current observation point;

Track evaluation module: if the continuous track points are deleted in the adjacent area, predicting the positions of the deleted points by using a Kalman filter and the reserved track points, evaluating the error range of the predicted positions and the observed positions, and if the error is smaller than a threshold value, judging that the positions have map errors;

The missing road section generation module: and taking the deleted points as the input of a map generator, dividing the deleted road sections into a series of sections according to the calculated inflection point positions, and calculating the line segments passing through the inflection points by using a linear regression model in the sections.

Preferably, in the area windowing module, the geographic area is firstly divided into grid cells;

If the vehicle does not spend a predetermined travel time in the grid cell, the vehicle is not allowed to transition from the grid cell;

In each grid cell, a variation of the DP algorithm is used to smooth the trajectory, replacing the original trajectory with an approximate line segment;

if the replacement does not meet the specified error requirement, recursively dividing the original problem into two sub-problems by taking the selected position point as a dividing point, and judging whether the replacement meets the specified error requirement or not until the error between the approximate track and the original track is lower than a specified error threshold;

The DP algorithm uses the orthogonal Euclidean distance as an error measure to keep the direction trend in the approximate track;

by removing partition points whose distance is greater than a threshold value, speed variation in the grid cells is limited.

Preferably, in the map matching module, in a map matching stage, smooth GPS points acquired from the regional windowing step are taken as input, the smooth GPS points are mapped to road segments on a digital map, and a hidden markov model is used for determining a road segment sequence corresponding to a time stamp sequence of GPS raw data;

The hidden markov model is a discrete time markov process with candidate states and observations;

searching candidate states using an index grid, each candidate state being a point closest to an observation point in a road segment, and generating an observation probability The observation probability/>Is defined as a GPS sampling point p _i and a candidate point/>, calculated based on a Gaussian noise modelPossibility of matching:

Wherein σ is the standard deviation and d is defined as p _i and The projection distance between the two points, p _i is the sampling point at the moment i,/>Is the j-th candidate point of p _i;

Wherein, Is p _i th candidate road segment, c is edge/>At any point above, dist (p _i, c) is the Euclidean distance between p _i and c;

the hidden markov model allows a transition from a previous candidate state to a current candidate state, and the transition is controlled by a transition probability;

empirically, the hidden Markov model assumes that the shortest path between successive candidate states has transition probabilities

Thus, transition probabilityThe calculation is as follows:

Wherein, Is a candidate point of current observation; /(I)Is a candidate point for the previous observation; d is from/>To/>An absolute value of a difference between the shortest path length and the euclidean distance between observations;

and then, using an online Viterbi decoding algorithm to find the maximum likelihood sequence of the road section corresponding to the variable sliding window of the input sequence.

Preferably, in the probability fluctuation detection module, a layering chart is gradually constructed by defining an observation probability and a transition probability and adding new nodes and weighted edges after a GPS sampling point is received;

If the matching accuracy is reduced when a new GPS sample is added in the layered graph, the measurement error or map error is considered to exist;

Given the candidate point set, the hidden Markov model will calculate the maximum likelihood path using the Viterbi decoding algorithm;

When the candidate set contains the observed real position, the hidden Markov model deduces the correct path;

when there is a map error or a measurement error, the shortest path between two consecutive candidates is not a real path, and may detour;

taking the acquisition points of which the candidate set does not contain the real position as low-quality points, and deleting the detected low-quality points by a detection model;

The calculation process of the Viterbi algorithm is to maximize the product of the observation probability and the transition probability when a new GPS sampling point is added;

the sampling error or map error causes amplitude fluctuation of the optimal decoding probability;

taking the logarithm of the original value to make the fluctuation more obvious;

deleting the newly added point when the fluctuation of the matching probability caused by the newly added point sampling exceeds a preset threshold value;

Limiting the number of consecutive removal points when there is a map error;

When a series of consecutive sampling points for a region is removed beyond a limit number, the algorithm evaluator will evaluate the unmatched samples, detect if there is a map error, and if so, use a map generation model to fit the missing road.

Compared with the prior art, the invention has the following beneficial effects:

1. The invention creatively provides a map matching framework capable of detecting map errors and deleting low-quality sampling points, which can be combined with a map matching algorithm based on a hidden Markov model to correct matching results and improve matching accuracy;

2. because the prior map matching algorithm does not consider the factors of map errors, the problems of matching interruption and the like can be caused in the actual application process, the invention can detect and update the map database in the matching process by utilizing the sampled tracks, and avoids the consumption of a great deal of manpower in the prior map updating process;

3. The invention can predict the traffic condition around the specific area based on the matched position and recommend the upcoming interest point for the user.

Drawings

Other features, objects and advantages of the present invention will become more apparent upon reading of the detailed description of non-limiting embodiments, given with reference to the accompanying drawings in which:

FIG. 1 is a diagram of an application architecture of the present invention;

FIG. 2 is a diagram of a matching process of the present invention;

FIG. 3 is a graph of trace error after window smoothing;

FIG. 4 is a schematic diagram of candidate point selection;

fig. 5 is a schematic diagram of decoding probability fluctuations within a track.

Reference numerals: moving VEHICLES WITH GPS sensors denotes a Moving vehicle with a sensor; the Wireless network represents a Wireless network; GPS records represent sample records; DATA CENTER denotes a data center; route planning represents path recommendation; traffic estimation denotes traffic flow estimation; point of interest denotes points of interest; END DEVICES running LBS means a terminal device running a location-based service; map database represents a Map database; windowing represents a window technique; MAP MATCHING denotes map matching; the Detection model represents a Detection model; the Evaluation model represents an Evaluation model Generation model represents a generative model; width window means that window technology is used; withoutwindowing denotes unused window technology; CDF represents a position error probability cumulative distribution; radius r represents the candidate Radius; sequencenumberofpoints denotes the sampling point sequence number; variationofprobability denotes probability fluctuations.

Detailed Description

The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the present invention, but are not intended to limit the invention in any way. It should be noted that variations and modifications could be made by those skilled in the art without departing from the inventive concept. These are all within the scope of the present invention.

The embodiment of the invention discloses a map matching correction method based on hidden Markov multi-noise perception, which is shown in fig. 1 and 2 and comprises the following steps:

A region windowing step: by gridding the region, irregular drift of the track is limited, and a smooth track is obtained. Firstly dividing a geographic area into grid cells;

If the vehicle does not spend a predetermined travel time in the grid cell, the vehicle is not allowed to transition from the grid cell;

In each grid cell, a variation of the DP algorithm is used to smooth the trajectory, replacing the original trajectory with an approximate line segment. If the replacement does not meet the specified error requirement, recursively dividing the original problem into two sub-problems by using the selected position point as a dividing point, and judging whether the replacement meets the specified error requirement or not until the error between the approximate track and the original track is lower than a specified error threshold. The DP algorithm uses orthogonal euclidean distances as an error metric to maintain directional trends in the approximate trajectory. By removing partition points whose distance is greater than a threshold value, speed variation in the grid cells is limited. FIG. 3 shows a Cumulative Distribution Function (CDF) distribution of collected GPS data position errors. DP English is fully called Douglas-Peucker, and Chinese translation is the Dataglas-Pracker algorithm.

Map matching step: and inputting the smoothed track into a hidden Markov model with the deleted points and the map correction frame for matching. In the map matching stage, the smooth GPS points are obtained from the regional windowing step as input, mapped to road segments on the digital map, and a hidden Markov model is used for determining a road segment sequence corresponding to the time stamp sequence of the GPS original data. The hidden markov model is a discrete time markov process with candidate states and observations. Searching candidate states using an index grid, as shown in FIG. 4, each candidate state being a point closest to an observation point in a road segment, and generating an observation probabilityThe observation probability/>Is defined as a GPS sampling point p _i and a candidate point/>, calculated based on a Gaussian noise modelPossibility of matching:

Wherein σ is the standard deviation and d is defined as p _i and The projection distance between the two points, p _i is the sampling point at the moment i,/>Is the j-th candidate point of p _i.

Wherein,Is p _i th candidate road segment, c is edge/>At any point above, dist (p _i, c) is the Euclidean distance between p _i and c.

The hidden markov model allows transitions from a previous candidate state to a current candidate state, and the transitions are controlled by transition probabilities. Empirically, the hidden Markov model assumes that the shortest path between successive candidate states has transition probabilities

Thus, transition probabilityThe calculation is as follows:

Wherein, Is a candidate point of current observation; /(I)Is a candidate point for the previous observation; d is from/>To/>The absolute value of the difference between the shortest path length and the euclidean distance between observations.

And then, using an online Viterbi decoding algorithm to find the maximum likelihood sequence of the road section corresponding to the variable sliding window of the input sequence.

Probability fluctuation detection step: calculating fluctuation conditions of the Viterbi decoding probability of each observation point in the matching process; and if the fluctuation amplitude exceeds the threshold value, deleting the current observation point.

By defining the observation probability and the transition probability, a hierarchical graph is gradually constructed by adding new nodes and weighted edges after the GPS sampling points are received. If the matching accuracy decreases when a new GPS sample is added to the hierarchy chart, then a measurement error or map error is considered to exist. Given the candidate point set, the hidden Markov model will calculate the maximum likelihood path using a Viterbi decoding algorithm. When the candidate set contains the true position of the observation, the hidden markov model deduces the correct path. When there is a map error or a measurement error, the shortest path between two consecutive candidates is not a real path, and may detour. The acquisition points of which the candidate set does not contain a real position are regarded as low-quality points, and the detection model deletes the detected low-quality points. The viterbi algorithm is calculated by maximizing the product of the observation probability and the transition probability when adding new GPS sampling points. Sampling errors or map errors cause amplitude fluctuations in the optimal decoding probability. Taking the logarithm of the original value makes the fluctuations more pronounced. And deleting the newly added point when the fluctuation of the matching probability caused by the newly added point sampling exceeds a preset threshold value. Since map errors cause the decoding probability of all subsequent points to fluctuate beyond a threshold, the number of consecutive removed points needs to be limited. When a series of consecutive sampling points for a region is removed beyond a limit number, the algorithm evaluator will evaluate the unmatched samples, detect if there is a map error, and if so, use a map generation model to fit the missing road.

Track evaluation step: if the continuous track points are deleted in the adjacent area, the positions of the deleted points are predicted by using a Kalman filter and the reserved track points, the error range of the predicted positions and the observed positions is estimated, and if the error is smaller than a threshold value, the map error at the section of sampling track is judged.

And taking a series of GPS points continuously removed from the detection model as input, evaluating the quality of the whole sequence, and judging the reason of the removal. The series of motion patterns that remove GPS samples under position and velocity constraints are evaluated. If the detected abnormality is caused by a motion pattern that does not conform to the actual driving situation, then the change in the matching probability is considered to be caused by a measurement error, and the corresponding point is deleted; the rest points will be added again to the matching sequence to continue matching. If the removed points are allowed to form a driving pattern in the evaluation model, the removed GPS sampling points are recovered and the points that cannot be matched are used to generate a missing road. The motion direction and position of the vehicle are predicted using kalman filtering, and then the estimated points are used as new matching points to determine whether there is an abnormality. And taking the speed and direction information of the road section adjacent to the sampling point as the input of a motion trail physical model, calculating the ideal position of the vehicle at the next moment, and taking the sampled trail direction and speed information as the input of the motion model to calculate the position at the next moment as a measured value. Based on the error between the weighted estimated value and the true value of the system, the Kalman gain is calculated, and the position estimated value at the next moment is obtained as follows:

x_t|t＝K_tz_t+(I-K_tH)x_t|t-1

Where x _t|t is the system estimate, x _t|t-1 is the system ideal state value, z _t is the measurement, K _t is the kalman gain, I is the identity matrix, and H is the state space transition matrix.

Comparing the position predicted value with the position sampling value, calculating the average distance between the position predicted value and the position sampling value, and if the distance is smaller than the threshold value delta d, considering the sampling to be caused by map errors.

A missing road section generation step: and taking the deleted points as the input of a map generator, dividing the deleted road sections into a series of sections according to the calculated inflection point positions, and calculating the line segments passing through the inflection points by using a linear regression model in the sections.

The digital map consists of an attributed undirected graph; the intersection points are represented as vertices with a degree greater than two, and the road segments are represented as a series of edges; treating the bi-directional link as a single path; defining high-order attributes to supplement the graph; the geometric features are displayed as part of the map description; providing road-related local geometry information. The map generation model based on the clustering technology is used for automatically deducing the road and traffic characteristics. And fitting the GPS sampling points which cannot be matched by the map generation model, wherein the generated topological structure is close to an actual missing road. According to the position of the inflection point, the GPS points which cannot be matched are divided into different sections. The distance of the inflection point is measured by a euclidean distance variable of the vertical euclidean distance. At each segment, a linear regression model is used to fit the missing road. To determine the location of the turning point, parameters θ _G and d _G are set. The first and last points in the sequence are connected and then all other points are projected onto the connection line. If the specified projection distance is greater than the threshold d _G, the point having the specified distance is a corner point of the real road segment. Connecting points with specified projection distances with a first point and a last point respectively, and continuously calculating angles between the connecting points; if the angle is greater than the threshold Δ _G, the sequence is split into two parts and iterated through.

The embodiment of the invention discloses a map matching and correcting framework based on a hidden Markov model and with multiple noise perception, and as shown in fig. 1, a high-level system model of a typical position-based service is shown.

1. Frame application: in most pervasive computing applications, the measured location of the mobile vehicle is reported to a server via wireless communication. The sampling interval varies in the same trajectory depending on the driving environment. If there are too many outliers in the GPS trajectory, filtering pre-processing such as median filtering, kalman filtering, and particle filtering is required. The server runs a map matching algorithm that uses noisy GPS points to find the actual road segments that the user is driving. Once the driving route is determined, other LBS applications may query the results to meet their requirements. For example, based on the location of the match, traffic conditions around a particular area may be predicted and upcoming points of interest may be recommended to the user. LBS english is known as Location Based Services and chinese translation is a location-based service.

2. Matching process:

As shown in fig. 2, the process flow of the inventive framework is described in detail. To conserve energy while driving, low energy devices are typically used to acquire and transmit GPS records. These acquisition devices may have problems switching back and forth. Therefore, we first smooth the trajectory data using a window technique. After smoothing, track data which is not converted back and forth is output and can be used as input to be sent to the map matching module. The present patent regards map matching and map updating as one collaborative optimization process.

In particular, when the accuracy of the matching result is low, the framework proposed by the present invention evaluates whether it is caused by GPS measurement errors or map errors. If the cause is the first factor, outliers in the GPS trajectory can simply be deleted or calibrated. Otherwise, the missing road segments are generated using a generator, and then the matching results are corrected and the map database is updated with these road segments. The process improves map matching precision and map quality by dynamically detecting map errors and automatically updating the map database, and prevents roundabout and matching interruption in the matching process.

The system comprises the following modules:

(1) Region windowing module: by gridding the region, irregular drift of the track is limited, and a smooth track is obtained.

The first module is a regional windowed template, which, to simplify the smoothing process, first divides the geographic region into uniform square grid cells G of fixed size. If the vehicle does not spend sufficient travel time in the grid cell, the vehicle is not allowed to transition from the grid cell. In each grid cell, the present patent uses a variation of the Douglas-Peucker (DP) algorithm to smooth the trajectory, which is typically used to simplify the trajectory. The basic idea is to replace the original trajectory with an approximate line segment. If the substitution does not meet the specified error requirement, it will recursively divide the original problem into two sub-problems by selecting the most erroneous location point as the dividing point. This process will continue until the error between the approximated and original tracks is below the specified error threshold. The goal of the DP algorithm is to use orthogonal euclidean distances as an error metric to maintain directional trends in the approximate trajectory. The implementation of the present invention differs from the previous method in that the present invention removes partition points that are greater than a threshold distance to limit speed variation in one grid cell. FIG. 3 shows a Cumulative Distribution Function (CDF) distribution of collected GPS data position errors. DP English is fully called Douglas-Peucker, and Chinese translation is the Dataglas-Pracker algorithm.

(2) And a map matching module: and inputting the smoothed track into a hidden Markov model with the deleted points and the map correction frame for matching.

In the map matching phase, a series of smoothed GPS points from the regional windowing module are taken as input and mapped to road segments on a digital map. The framework proposed by the present invention uses a Hidden Markov Model (HMM) to determine the sequence of road segments corresponding to the sequence of time stamps of the GPS raw data. HMM is a discrete-time markov process with a set of candidate states and observations. The present invention uses an index grid to search for candidate states as shown in fig. 4. Each state is the point closest to the observation point in the road section, and generates an observation probability defined as the GPS sampling point p _i and the candidate point calculated based on Gaussian noise modelPossibility of matching:

Wherein σ is the standard deviation and d is defined as p _i and The projection distance between the two points, p _i is the sampling point at the moment i,/>Is the j-th candidate point of p _i.

Wherein,Is p _i th candidate road segment, c is edge/>At any point above, dist (p _i, c) is the Euclidean distance between p _i and c.

The HMM allows a transition from a previous candidate state to a current candidate state, and this transition is controlled by the transition probability. Empirically, HMM assumes that the shortest path between two consecutive candidate states has a large transition probability. Thus, the transition probabilities are typically calculated as follows:

Wherein, And/>The candidate points for the previous observation and the current observation, respectively, D is the slave/>To/>The absolute value of the difference between the shortest path length and the euclidean distance between observations. The maximum likelihood sequence of the road segment corresponding to the variable sliding window of the input sequence can then be found using an online viterbi decoding algorithm.

(3) Probability fluctuation detection module: calculating fluctuation conditions of the Viterbi decoding probability of each observation point in the matching process; and if the fluctuation amplitude exceeds the threshold value, deleting the current observation point.

By defining the observation probability and the transition probability, a hierarchical graph can be built step by adding new nodes and weighted edges after receiving the GPS sampling points. The model aims to improve the robustness and the matching precision of the map matching algorithm based on the HMM when facing a real complex scene. In other words, if the matching accuracy is lowered when a new GPS sample is added to the hierarchical map, it is considered that there is a measurement error or map error. Given the candidate point set, the HMM will calculate the maximum likelihood path using the viterbi decoding algorithm. However, one disadvantage of this approach is that it is not aware of the quality of the current observation. Only when the candidate set contains the actual location of the observation, the HMM can infer the correct path. When there is a map error or a measurement error, the shortest path between two consecutive candidates is never a true path and a long detour is caused.

The present invention treats acquisition points where the candidate set does not contain a true position as low quality points, which may be caused by measurement errors or map errors. The detection model will delete the detected low quality points. The viterbi algorithm is calculated by maximizing the product of the observation probability and the transition probability when adding new GPS sampling points. Since map errors or measurement errors generally result in lower observation probabilities and transition probabilities, and the maximum value of the viterbi decoding probabilities is related to the observation probabilities and transition probabilities, sampling errors or map errors generally cause large fluctuations in the optimal decoding probabilities. Fig. 5 shows the variation in the difference between the maximum matching probabilities of the current GPS point and the previous GPS point in a trace. The invention takes the logarithm of the original value to make the fluctuation more obvious. When these points result in a large change in the probability of matching, these points can be deleted, which is typically caused by map errors or measurement errors. If there is a map error, this process may delete all subsequent GPS samples. Therefore, we should limit the number of consecutive removal points. The number of consecutive removal points is limited due to map errors. When a series of consecutive samples of an area are removed, the algorithm evaluator will evaluate these unmatched samples to detect if there is a map error, and if so, a map-generated model can be used to fit the missing road.

(4) Track evaluation module: if the continuous track points are deleted in the adjacent area, the positions of the deleted points are predicted by using a Kalman filter and the reserved track points, the error range of the predicted positions and the observed positions is estimated, and if the error is smaller than a threshold value, the map error at the positions is judged.

The module takes as input a series of GPS points continuously removed from the detection model and evaluates the quality of the entire sequence to determine if they were removed due to map errors or measurement errors. This patent evaluates this series of motion patterns that remove GPS samples under position and velocity constraints. If the detected abnormality is caused by a motion pattern that does not conform to the actual driving situation, a large change in the matching probability is considered to be caused by a measurement error, and the corresponding point is deleted. The rest of the points will rejoin the matching sequence to continue matching. If the driving pattern formed by these removed points is allowed in the evaluation model, the removed GPS sampling points are recovered and these unmatched points are used to generate a missing road. The present patent uses kalman filtering to predict the direction and position of motion of the vehicle and then uses these estimated points as new matching points to determine if there is an anomaly.

The invention takes the speed and direction information of the road section adjacent to the sampling point as the input of the motion trail physical model, calculates the ideal position of the next moment of the vehicle, and takes the sampled trail direction and speed information as the input of the motion model to calculate the position of the next moment as the measured value. Based on the minimum error between the weighted estimated value and the true value of the system, the Kalman gain is calculated, so that the position estimated value at the next moment is obtained as follows:

x_t|t＝K_tz_t+(I-K_tH)x_t|t-1

Where x _t|t is the system estimate, x _t|t-1 is the system ideal state value, z _t is the measurement, K _t is the kalman gain, I is the identity matrix, and H is the state space transition matrix. The position prediction value is compared with the position sampling value, the average distance between them is calculated, and if the distance is smaller than the threshold value deltad, the segment of the sampling is considered to be caused by map errors.

(5) The missing road section generation module: and taking the deleted points as the input of a map generator, dividing the deleted road sections into a series of sections according to the calculated inflection point positions, and calculating the line segments passing through the inflection points by using a linear regression model in the sections.

Digital maps are typically composed of attributed undirected graphs. The intersection points are represented as vertices with a degree greater than 2 and the road segments are represented as a series of edges. Because the graph is undirected, the path has no direction. Thus, a bi-directional link is considered a single path. Higher order attributes may be defined to supplement the graphics. Geometric features such as width and curvature may be displayed as part of the map description. These functions are critical for road safety applications because they provide detailed information about the local geometry of the road. It is much easier to infer the structure of the road network than these higher order link information. In addition, this patent uses a clustering-based map-generation model with general applicability to automatically infer road and traffic characteristics (e.g., speed limit, road type, and lane structure). The map generation model aims at fitting the GPS sampling points which cannot be matched, and the generated topological structure is closer to the actual missing road. Map generation is challenged by measurement errors that result in the sample points not falling exactly on the road. If the GPS sampling point is directly used as the intersection point of the road sections, the generated road sections will have the jaggies. The non-matching GPS points are divided into different sequences (like several sub-sequences) according to the location of the inflection point. The maximum distance of the inflection point is measured by a euclidean distance variable called the vertical euclidean distance. At each segment, the present patent uses a linear regression model to fit the missing road. To determine the location of the turning point, two parameters θ _G and d _G are set. The first and last points in the sequence are connected and then all other points are projected onto this line. If the maximum projected distance is greater than the threshold d _G, the point with the greatest distance may be a corner point of the real road segment. The calculation of the angle between the point with the greatest projection distance and the first and last points is continued. If the angle is also greater than the threshold θ _G, the sequence is split into two parts. This process may be iteratively completed with the projection distance threshold d _G decreasing with decreasing total distance of points within the interval and the angle threshold expanding with decreasing distance.

3. The algorithm is realized:

The invention applies the algorithm to the open source map matching platform GraphHopper, and the map network information is derived from OpenStreetMap. On a simulation platform, the framework of the method is combined with three classical map matching based on the hidden Markov model, a large number of experiments are carried out on track data sets which are 79670.6km long in total in four cities, indexes such as precision and recall rate of a matching result are verified, and after the experiment result shows that the method is combined with the framework of the method, the indexes of three algorithms are improved by about 20%, and in addition, on track proportion of the precision higher than 95% and 90%, the algorithm of the method is improved by about 30% relative to the original track matching algorithm based on the hidden Markov model. The frame detects 1962 map errors on the Shanghai data set, 200 tracks are randomly extracted, the accuracy of the algorithm on map error detection is manually verified, the accuracy rate of the algorithm is found to be about 60% near residential areas of residential areas, and the accuracy rate of the algorithm reaches about 80% on roads with dense roads and networks.

The invention greatly improves the accuracy of the map matching algorithm in the complex traffic network, and the algorithm can sense map errors, automatically correct map matching results based on the sampled tracks and update a map database. These characteristics improve the robustness and accuracy of the map matching algorithm.

The method firstly limits irregular drift of the track by gridding the area. And (3) inputting the smoothed track into a hidden Markov model with a deleting point and a map correcting frame for matching, and calculating the fluctuation condition of the Viterbi decoding probability of each observation point in the matching process. And if the fluctuation amplitude exceeds the threshold value, deleting the current observation point. If the continuous track points are deleted in the adjacent area, the positions of the deleted points are predicted by using a Kalman filter and the reserved track points, the error range of the predicted positions and the observed positions is estimated, and if the error is smaller than a threshold value, the position (at the section of the sampling track) is judged to have map errors. And taking the deleted points as the input of a map generator, dividing the deleted road sections into a series of sections according to the calculated inflection point positions, and calculating the line sections with the minimum errors among the inflection points by using a linear regression model in the sections. Since the distance errors between the estimated position and the observed position are small, the sampling quality is high, and the decoding probability fluctuation is not caused by the sampling error, but caused by map errors.

The invention provides a map matching framework resistant to various noises based on a hidden Markov model adopted by most map matching algorithms at present, and the map matching framework can process the conditions of low-precision track sampling and map errors. The low accuracy here is different from the focus of the existing proposed hidden markov model, and the current algorithm is based on a maximum likelihood estimation of a collaborative optimization of the calculated observation probability and the transition probability under the road network constraint. But ignores that the candidate radius may not contain a real candidate segment due to sampling errors.

The invention deletes the sampling points with large fluctuation amplitude based on the limitation of the fluctuation amplitude of the Viterbi decoding algorithm. And estimating the motion position and direction of the deleted points based on the Kalman filter and the current reserved points for a series of adjacent deleted points, comparing the motion position and direction of the deleted points with the conditions of sampling points, and judging whether map errors exist according to the error range. When map errors are detected, the deleted points are used as input of a map generator, the deleted road sections are divided into a series of sections according to the calculated inflection point positions, and the sections are internally calculated to obtain the line segments with the minimum errors among the inflection points by using a linear regression model.

The invention relates to intelligent transportation, track processing and map matching algorithms, and discloses a map matching framework capable of resisting various noises based on a hidden Markov model, which can process a low-precision GPS sampling track and detect and correct errors existing on a current electronic map. In the big data era, we can sample a large amount of GPS position data, which is important for tracking the current position of the vehicle in real time, but because of errors in GPS track sampling, a reliable algorithm is needed to accurately correspond the GPS track sampling to the road network so as to support other traffic applications, such as predicting whether a road section is blocked or not. At present, due to the acceleration of the urban process, new roads are continuously built in various countries, so that the road network updating speed is increased every year. The traditional map updating speed is low and relies on manpower greatly, so that when the traditional map updating speed is applied to the position-based service, reliable positioning and navigation service cannot be provided due to the problem of road section missing.

Thus, sampling errors and map errors are the greatest challenges they are currently facing for map matching algorithms. Most of the current matching algorithms are based on hidden Markov models, and the problems that no real road segments exist in the candidate set due to sampling errors and the map is inaccurate are not considered. In order to solve the problems, the patent provides a map matching and map correcting framework for resisting various noises based on a hidden Markov model, and the map matching result is corrected to improve the accuracy of a map matching algorithm based on the hidden Markov model.

Those skilled in the art will appreciate that the invention provides a system and its individual devices, modules, units, etc. that can be implemented entirely by logic programming of method steps, in addition to being implemented as pure computer readable program code, in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. Therefore, the system and various devices, modules and units thereof provided by the invention can be regarded as a hardware component, and the devices, modules and units for realizing various functions included in the system can also be regarded as structures in the hardware component; means, modules, and units for implementing the various functions may also be considered as either software modules for implementing the methods or structures within hardware components.

The foregoing describes specific embodiments of the present application. It is to be understood that the application is not limited to the particular embodiments described above, and that various changes or modifications may be made by those skilled in the art within the scope of the appended claims without affecting the spirit of the application. The embodiments of the application and the features of the embodiments may be combined with each other arbitrarily without conflict.

Claims (8)

1. The map matching correction method based on the hidden Markov multi-noise perception is characterized by comprising the following steps of:

a region windowing step: by gridding the region, the irregular drift of the track is limited, and a smooth track is obtained;

Map matching step: inputting the smoothed track into a hidden Markov model with a deleting point and a map correcting frame for matching;

Probability fluctuation detection step: calculating fluctuation conditions of the Viterbi decoding probability of each observation point in the matching process; if the fluctuation amplitude exceeds the threshold value, deleting the current observation point;

Track evaluation step: if the continuous track points are deleted in the adjacent area, predicting the deleted point positions by using a Kalman filter and the reserved track points, evaluating the error range of the predicted positions and the observed positions, and if the error is smaller than a threshold value, judging that the map error exists at the section of sampling track;

A missing road section generation step: taking the deleted points as the input of a map generator, dividing the deleted road sections into a series of sections according to the calculated inflection point positions, and calculating line segments passing through the inflection points in the sections by using a linear regression model;

In the map matching step, in the map matching stage, smooth GPS points obtained from the regional windowing step are taken as input, the smooth GPS points are mapped to road segments on a digital map, and a hidden Markov model is used for determining a road segment sequence corresponding to a time stamp sequence of GPS original data;

The hidden markov model is a discrete time markov process with candidate states and observations;

searching candidate states using an index grid, each candidate state being a point closest to an observation point in a road segment, and generating an observation probability The observation probability/>Is defined as a GPS sampling point p _i and a candidate point/>, calculated based on a Gaussian noise modelPossibility of matching:

Wherein σ is the standard deviation and d is defined as p _i and The projection distance between the two points, p _i is the sampling point at the moment i,/>Is the j-th candidate point of p _i;

Wherein, Is p _i th candidate road segment, c is edge/>At any point above, dist (p _i, c) is the Euclidean distance between p _i and c;

the hidden markov model allows a transition from a previous candidate state to a current candidate state, and the transition is controlled by a transition probability;

empirically, the hidden Markov model assumes that the shortest path between successive candidate states has transition probabilities

Thus, transition probabilityThe calculation is as follows:

Wherein, Is a candidate point of current observation; /(I)Is a candidate point for the previous observation; d is from/>To/>An absolute value of a difference between the shortest path length and the euclidean distance between observations;

and then, using an online Viterbi decoding algorithm to find the maximum likelihood sequence of the road section corresponding to the variable sliding window of the input sequence.

2. The map matching correction method based on hidden markov multi-noise perception according to claim 1, wherein in the region windowing step, a geographical region is first divided into grid cells;

If the vehicle does not spend a predetermined travel time in the grid cell, the vehicle is not allowed to transition from the grid cell;

In each grid cell, a variation of the DP algorithm is used to smooth the trajectory, replacing the original trajectory with an approximate line segment;

if the replacement does not meet the specified error requirement, recursively dividing the original problem into two sub-problems by taking the selected position point as a dividing point, and judging whether the replacement meets the specified error requirement or not until the error between the approximate track and the original track is lower than a specified error threshold;

The DP algorithm uses the orthogonal Euclidean distance as an error measure to keep the direction trend in the approximate track;

by removing partition points whose distance is greater than a threshold value, speed variation in the grid cells is limited.

3. The map matching correction method based on hidden markov multi-noise perception according to claim 1, wherein in the probability fluctuation detection step, a hierarchical map is gradually constructed by adding new nodes and weighted edges after receiving GPS sampling points by defining observation probabilities and transition probabilities;

If the matching accuracy is reduced when a new GPS sample is added in the layered graph, the measurement error or map error is considered to exist;

Given the candidate point set, the hidden Markov model will calculate the maximum likelihood path using the Viterbi decoding algorithm;

When the candidate set contains the observed real position, the hidden Markov model deduces the correct path;

when there is a map error or a measurement error, the shortest path between two consecutive candidates is not a real path, and may detour;

taking the acquisition points of which the candidate set does not contain the real position as low-quality points, and deleting the detected low-quality points by a detection model;

The calculation process of the Viterbi algorithm is to maximize the product of the observation probability and the transition probability when a new GPS sampling point is added;

the sampling error or map error causes amplitude fluctuation of the optimal decoding probability;

taking the logarithm of the original value to make the fluctuation more obvious;

deleting the newly added point when the fluctuation of the matching probability caused by the newly added point sampling exceeds a preset threshold value;

Limiting the number of consecutive removal points when there is a map error;

When a series of consecutive sampling points for a region is removed beyond a limit number, the algorithm evaluator will evaluate the unmatched samples, detect if there is a map error, and if so, use a map generation model to fit the missing road.

4. The map matching correction method based on hidden markov multi-noise perception according to claim 1, wherein in the trajectory evaluation step, a series of GPS points continuously removed from the detection model are taken as input, and the quality of the entire sequence is evaluated, and the reason for the removal is judged;

evaluating the series of motion patterns under position and velocity constraints that remove GPS samples;

if the detected abnormality is caused by a motion pattern that does not conform to the actual driving situation, then the change in the matching probability is considered to be caused by a measurement error, and the corresponding point is deleted; the rest points are added into the matching sequence again to continue matching;

Restoring the removed GPS sampling points and generating a missing road by using points which cannot be matched if the removed points are allowed to form a driving mode in the evaluation model;

Predicting the movement direction and position of the vehicle by using Kalman filtering, and then judging whether an abnormality exists by taking the estimated points as new matching points;

Taking the speed and direction information of the road section adjacent to the sampling point as the input of a motion trail physical model, calculating the ideal position of the vehicle at the next moment, and taking the sampled trail direction and speed information as the input of the motion model to calculate the position at the next moment as a measured value;

based on the error between the weighted estimated value and the true value of the system, the Kalman gain is calculated, and the position estimated value at the next moment is obtained as follows:

x_t|t＝K_tz_t+(I-K_tH)x_t|t-1

Wherein x _t|t is a system estimation value, x _t|t-1 is a system ideal state value, z _t is a measurement value, K _t is Kalman gain, I is an identity matrix, and H is a state space conversion matrix;

Comparing the position predicted value with the position sampling value, calculating the average distance between the position predicted value and the position sampling value, and if the distance is smaller than the threshold value delta d, considering the sampling to be caused by map errors.

5. The hidden markov based map matching correction method according to claim 1, wherein in the missing road section generating step, a digital map is composed of an attributed undirected graph; the intersection points are represented as vertices with a degree greater than two, and the road segments are represented as a series of edges; treating the bi-directional link as a single path; defining high-order attributes to supplement the graph; the geometric features are displayed as part of the map description; providing local geometry information about the road;

automatically deducing road and traffic characteristics by using a map generation model based on a clustering technology;

Fitting the GPS sampling points which cannot be matched by the map generation model, wherein the generated topological structure is close to an actual missing road;

dividing the GPS points which cannot be matched into different sections according to the positions of the inflection points;

The distance of the inflection point is measured by a euclidean distance variable of the vertical euclidean distance;

at each segment, fitting the missing road using a linear regression model;

In order to determine the position of the turning point, parameters θ _G and d _G are set;

connecting a first point and a last point in the sequence, and then projecting all other points to the connecting line;

If the specified projection distance is greater than the threshold d _G, the point with the specified distance is a corner point of the real road segment;

connecting points with specified projection distances with a first point and a last point respectively, and continuously calculating angles between the connecting points; if the angle is greater than the threshold θ _G, the sequence is split into two parts and iterated.

6. A hidden markov based multi-noise aware map matching correction system comprising the following modules:

region windowing module: by gridding the region, the irregular drift of the track is limited, and a smooth track is obtained;

And a map matching module: inputting the smoothed track into a hidden Markov model with a deleting point and a map correcting frame for matching;

Probability fluctuation detection module: calculating fluctuation conditions of the Viterbi decoding probability of each observation point in the matching process; if the fluctuation amplitude exceeds the threshold value, deleting the current observation point;

Track evaluation module: if the continuous track points are deleted in the adjacent area, predicting the positions of the deleted points by using a Kalman filter and the reserved track points, evaluating the error range of the predicted positions and the observed positions, and if the error is smaller than a threshold value, judging that the positions have map errors;

the missing road section generation module: taking the deleted points as the input of a map generator, dividing the deleted road sections into a series of sections according to the calculated inflection point positions, and calculating line segments passing through the inflection points in the sections by using a linear regression model;

In the map matching module, in the map matching stage, smooth GPS points are obtained from the regional windowing step as input, the smooth GPS points are mapped to road segments on a digital map, and a hidden Markov model is used for determining a road segment sequence corresponding to a time stamp sequence of GPS original data;

The hidden markov model is a discrete time markov process with candidate states and observations;

searching candidate states using an index grid, each candidate state being a point closest to an observation point in a road segment, and generating an observation probability The observation probability/>Is defined as a GPS sampling point p _i and a candidate point/>, calculated based on a Gaussian noise modelPossibility of matching:

Wherein σ is the standard deviation and d is defined as p _i and The projection distance between the two points, p _i is the sampling point at the moment i,/>Is the j-th candidate point of p _i;

Wherein, Is p _i th candidate road segment, c is edge/>At any point above, dist (p _i, c) is the Euclidean distance between p _i and c;

the hidden markov model allows a transition from a previous candidate state to a current candidate state, and the transition is controlled by a transition probability;

empirically, the hidden Markov model assumes that the shortest path between successive candidate states has transition probabilities

Thus, transition probabilityThe calculation is as follows:

Wherein, Is a candidate point of current observation; /(I)Is a candidate point for the previous observation; d is from/>To/>An absolute value of a difference between the shortest path length and the euclidean distance between observations;

and then, using an online Viterbi decoding algorithm to find the maximum likelihood sequence of the road section corresponding to the variable sliding window of the input sequence.

7. The hidden markov based multi-noise aware map matching correction system of claim 6 wherein in the region windowing module, the geographic region is first divided into grid cells;

If the vehicle does not spend a predetermined travel time in the grid cell, the vehicle is not allowed to transition from the grid cell;

In each grid cell, a variation of the DP algorithm is used to smooth the trajectory, replacing the original trajectory with an approximate line segment;

if the replacement does not meet the specified error requirement, recursively dividing the original problem into two sub-problems by taking the selected position point as a dividing point, and judging whether the replacement meets the specified error requirement or not until the error between the approximate track and the original track is lower than a specified error threshold;

The DP algorithm uses the orthogonal Euclidean distance as an error measure to keep the direction trend in the approximate track;

by removing partition points whose distance is greater than a threshold value, speed variation in the grid cells is limited.

8. The map matching correction system based on hidden markov multi-noise perception according to claim 6, wherein in the probability fluctuation detection module, a hierarchical map is gradually constructed by defining an observation probability and a transition probability and adding new nodes and weighted edges after receiving GPS sampling points;

If the matching accuracy is reduced when a new GPS sample is added in the layered graph, the measurement error or map error is considered to exist;

Given the candidate point set, the hidden Markov model will calculate the maximum likelihood path using the Viterbi decoding algorithm;

When the candidate set contains the observed real position, the hidden Markov model deduces the correct path;

when there is a map error or a measurement error, the shortest path between two consecutive candidates is not a real path, and may detour;

taking the acquisition points of which the candidate set does not contain the real position as low-quality points, and deleting the detected low-quality points by a detection model;

The calculation process of the Viterbi algorithm is to maximize the product of the observation probability and the transition probability when a new GPS sampling point is added;

the sampling error or map error causes amplitude fluctuation of the optimal decoding probability;

taking the logarithm of the original value to make the fluctuation more obvious;

deleting the newly added point when the fluctuation of the matching probability caused by the newly added point sampling exceeds a preset threshold value;

Limiting the number of consecutive removal points when there is a map error;

When a series of consecutive sampling points for a region is removed beyond a limit number, the algorithm evaluator will evaluate the unmatched samples, detect if there is a map error, and if so, use a map generation model to fit the missing road.