JP2004354396A - Method for coding shape vector of digital map, position information transmission method, and apparatus for executing them - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for coding the position information of a digital map with a small amount of data by utilizing a compression coding technique. <P>SOLUTION: In the coding method for coding data for indicating the shape vector on the digital map, node rows for indicating the shape vector and position information for indicating shapes are expressed by data having statistical biasing by executing arithmetic processing, and data having statistical biasing are coded to reduce the amount of data. Accordingly, the amount of transmission data when transmitting the vector shape of the digital map can be decreased greatly. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method for transmitting position information of a digital map, an encoding method for compressing and encoding the amount of data to be transmitted, and an apparatus therefor, and in particular, reduces the amount of data by using a compression encoding technique. Things.

  Conventionally, when providing traffic information to in-vehicle navigation equipment equipped with a digital map database, the location on the digital map is accurate even when the sender and the receiver hold digital maps of different origins. , A road is specified by a link number, and a node such as an intersection existing on the road is specified by a node number, and a point on the road is expressed in terms of how many meters from the node. However, the node numbers and link numbers defined in the road network need to be replaced with new numbers when new or changed roads are created, and the digital map data of the producers must be updated accordingly. Therefore, in the method using the node number or the link number, a large social cost is required for the maintenance.

  In order to improve these points, the inventors of the present invention have proposed the following digital map position information transmission method in Japanese Patent Application Nos. 11-214068 and 11-242166. In this method, when the information provider informs the road position where an event such as traffic congestion or accident has occurred, the information providing side determines the road shape of a road section of a predetermined length including the event position on a node and an interpolation point ( A vertex of a polygonal line approximating a curve of a road (in this specification, unless otherwise specified, a node including interpolation points is referred to as a "node"); The event location data representing the event location is transmitted to the receiving side based on the relative position in the road section represented by the symbol, and the receiving side performs map matching using the road shape data, and Of the road section on the digital map, and the event occurrence position in this road section is specified using the event position data. FIG. 36A illustrates “road shape data”, and FIG. 36B illustrates “event position data”.

  However, in the method of transmitting the position information of the digital map using the "road shape data" and the "event position data", when the data amount of the road shape data for specifying the road shape increases, and the data transmission amount increases. There is a problem.

  As a method of reducing the data amount of the road shape data, the inventors of the present invention have proposed a method of approximating a road shape with a spline function in Japanese Patent Application No. 2001-12127. In order to achieve this, it is necessary to further reduce the amount of data.

  The present invention addresses such a problem, a position information transmission method of transmitting position information of a digital map with a small amount of data using a compression encoding technology, an encoding method for reducing the amount of data, It is intended to provide an apparatus for performing the method.

  Therefore, according to the present invention, in an encoding method for encoding data representing a shape vector on a digital map, arithmetic processing is performed on position information of each node of a node sequence representing the shape vector, and the position information is statistically processed. The data is converted to data having a bias, and the data is encoded to reduce the data amount.

  Further, the transmitting side transmits shape data representing a shape vector on the digital map, and the receiving side performs map matching based on the received shape data to specify the shape vector on its own digital map. In the position information transmission method, the transmitting side transmits shape vector data encoded by the encoding method, and the receiving side decodes the received data to reproduce the shape, and the shape vector corresponding to the reproduced shape. Is configured to be specified by map matching.

  Further, a transmitting device that transmits shape data representing the shape vector on the digital map to the receiving side, performs arithmetic processing on the position information of each node in the node sequence representing the shape vector on the digital map, and obtains the position information. Code table calculating means for converting the data into data having a statistical bias and generating a code table used for encoding the data based on the distribution of occurrence of the data, and the position of each node of the shape vector transmitted to the receiving side Position information converting means for encoding information using the code table and generating shape data to be sent to the receiving side is provided.

  Also, a receiving device that receives encoded data representing a shape vector on a digital map from a transmitting side decodes the encoded received data and reproduces the shape data represented by position information on the digital map. Decoding means and map matching means for performing map matching using the reproduced shape data and specifying the shape vector on its own digital map are provided.

  Therefore, the data amount of the shape vector in the digital map can be efficiently compressed, and the transmission data amount in transmitting the shape vector of the digital map can be significantly reduced. On the receiving side, the transmitted shape vector can be accurately specified by restoring the shape data from the received data and performing map matching.

(1st Embodiment)
In the first embodiment, a method for compressing data by variable length coding will be described.

  In the method for transmitting position information of a digital map according to the present invention, first, a road shape is represented by shape data having a statistical bias. This is to increase the compression ratio when the shape data is compression-coded.

When the road shape is represented by coordinate points arrayed on the road, as shown in FIG. 34, the position of each coordinate point (P _J ) is determined by the distance and angle from the adjacent coordinate point (P _J-1 ). It can be uniquely specified by the two dimensions. In FIG. 34, an angle Θ _j based on “absolute azimuth” indicating a magnitude in the range of 0 to 360 degrees clockwise is shown in FIG. Expressing the coordinate points using the distance and the absolute azimuth in this way is called a full curvature function expression.

The xy coordinates of the coordinate points P _J-1 , P _J , and P _{J + 1} are respectively (x _j-1 , y _j-1 ), (x _j , y _j ), (x _{j + 1} , y _{j + 1} ) when the distance L _j (coordinate point P _J, P _{J + 1} between the distance) and absolute angle theta _j (absolute azimuth of a straight line extending from the coordinate point P _J in the coordinate point P _{J + 1)} is the following formula Can be calculated.
L _j = {(x _{j + 1} −x _j ) ² + (y _{j + 1} −y _j ) ² }
^{_{Θ j = tan -1 {(x}} j + 1 -x j) / (y j + 1 -y j)}

By the way, when a new coordinate point is set (resampled) on the road shape such that the distance from the adjacent coordinate point is constant (= L), information on the common L is added. In addition, only by transmitting information of the angle Θ _j (that is, information of one dimension) with respect to each of these coordinate points, it becomes possible to specify the position of the coordinate point on the receiving side, thereby reducing the amount of transmission data. Can be planned.

FIG. 35A shows the absolute azimuth Θ _j at each coordinate point (P _J ) when the coordinate point is resampled to a position where the distance from the adjacent coordinate point on the road is constant (= L). I have. When each coordinate point is represented by the absolute azimuth Θ _j , as shown in FIG. 35 (a ′), no statistical bias appears in the frequency of occurrence of the angle information Θ _j indicating each coordinate point.

However, the angle of each coordinate point, as shown in FIG. 35 (b), displacement difference of the absolute azimuth, i.e., may also be represented by a "deflection angle" theta _j. This declination θ _j is
θ _j = Θ _j −Θ _j-1
Is calculated as When each coordinate point is represented by the declination θ _j , in an area where there are many linear roads, as shown in FIG. 35 (b ′), the occurrence frequency of the angle information θ _j indicating each coordinate point is θ = A maximum appears at 0 °.

Further, the angle of each coordinate point can also be represented by a difference Δθ _j between the argument θ _j and the argument statistical predicted value S _j (predicted value represented by the argument) as shown in FIG. 35 (c). . The argument statistical predicted value S _j is a value obtained by estimating the argument θ _j of the coordinate point P _J of interest using the argument of the coordinate point up to P _J-1 before that. For example, the argument statistical prediction value S _j is:
S _j = θ _j-1
Or
S _j = (θ _j-1 + θ _j-2 ) / 2
Can be defined as Further, a weighted average of the argument of the past n coordinate points may be defined as S _j . The declination predicted value difference Δθ _j is
Δθ _j = θ _j −S _j
Is calculated as

Since most road shapes are straight lines or curves that bend gently, when the distance L between coordinate points is set to be constant, the predicted value difference Δθ _j of declination is concentrated near 0 °, and FIG. As shown in FIG. 1B), the frequency of occurrence of the angle information Δθ _j indicating each coordinate point shows a strong bias around θ = 0 °.

Therefore, here, in order to obtain shape data having a statistical bias, as shown in FIG. 1, road shapes (original shapes) are sampled at regular intervals with a resample section length L having a fixed distance, The position data of the sampling point (node) P _J is represented by a predicted value difference Δθ _j (= θ _j −S _j ) of the argument θ _j . Note that the distance mentioned here may be an actual distance when developed in the real world, or a length expressed in a unit of predetermined normalized coordinates.

Now, the deviation statistical prediction value S _j is given by
S _j = (θ _j-1 + θ _j-2 ) / 2
Let's define Since the road shape bends in most cases,
θ _j ≒ (θ _j-1 + θ _j-2 ) / 2 = S _j
It is considered that Δθ _j is distributed in an extremely narrow range centering on 0.

This Δθ _j can theoretically take a value of −360 ° to + 360 °. Therefore, in order to express Δθ _j with a resolution of 1 °, 10 bits obtained by adding 1 bit representing positive and negative and 9 bits representing a numerical value of 360 are required. However, an angle near ± 0 ° is encoded with a value smaller than 10 bits. By allocating a value larger than 10 bits to an angle away from ± 0 °, the average number of bits used for encoding Δθ _j can be made smaller than 10 bits, and the shape data is expressed as a short total data amount. It becomes possible.

  FIG. 2 shows an example of a code table in which an encoding code is assigned to Δθ. If Δθ = 0, it is coded to 0. In the case of Δθ = + 1, an additional bit 0 representing + is added to the code 100 and encoded as 1000. In the case of Δθ = −1, an additional bit 1 representing − is added to the code 100 and encoded as 1001.

The variable length coding will be described with reference to FIG. When the number of nodes is 6 (= starting point + 5 nodes), ordinary coding requires a fixed-length data amount of 5 × 10 bits = 50 bits in addition to the initial angle (10 bits). On the other hand, in the case of encoding using the code table of FIG. 2, if the value of Δθ _j is 0 three times and within ± 2 ° twice, then 3 in addition to the initial value angle (10 bits) × 1 bit + 2 × 4 bits = 11 bits. If this data is "0, 0, +1, -2, 0", it is represented as "00100010110" by encoding.

The receiving side can obtain each value of Δθ _j by referring to the code table sent together with the shape data (or held in advance) and applying the values of Δθ in order. Then, by sequentially accumulating the values from the initial value, the value of the argument θ _j at each coordinate point can be uniquely determined.

This code table is created by calculating the angle of Δθ _j at each coordinate point P _J , examining the occurrence frequency of the angle, and constructing the code table using a well-known Huffman tree or the like according to the occurrence frequency.

  As described above, after performing arithmetic processing on the shape data so as to have a statistical bias, and then performing variable-length encoding, the data amount of the shape data can be reduced.

Here, the resampled node position is represented by the distance and the argument of the adjacent node, but the node positions sampled at equal intervals with the resample section length L are represented by relative latitude and longitude coordinates (Δx _j , Δy _j ) can also be expressed. In this case, the statistical value S _j is defined as, for example, S _jx = Δx _j−1 , S _jy = Δy _j−1 ,
Δx _j = S _jx + δx _j = Δx _j-1 + δx _j
Δy _j = S _jy + δy _j = Δy _j-1 + δy _j
Δx _j and δy _j are variable-length coded and transmitted as shape data.

(Second embodiment)
In the second embodiment, a method for compressing data using a run-length method will be described.

In the example of the first embodiment, when Δθ _j is encoded to represent shape data, “0” is continuous on a straight road or a road turning at the same curvature. In such a case, the data compression ratio is higher when expressed as "0 times 20" than when expressed as "00000 $". Here, such run-length encoding is performed to compress data.

FIG. 3 shows a run-length code table. For example, when the same number continues five times (when the run-length is 5), it is defined to be displayed as “101”. FIG. 4 is the same code table of Δθ as in FIG. The arrangement of data is determined, for example, as follows: run length → Δθ → run length → Δθ → ‥. Δθ is “0,0,0,0,0, -2, -2,0, + 3,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0, -1 ‥ ”
If you continue with the run-length method,
“101 ・ 0_0 ・ 1011_0 ・ 1011_0 ・ 0_0 ・ 11000_1101 ・ 0_0 ・ 1001 ‥”
→ "10100101101011000110001101001001 @" (32 bits)
Is expressed as
On the other hand, if you do not use run-length expressions,
“000001011101101100000000000000000001001 ‥” (38 bits).

Also, the code tables in FIGS. 3 and 4 can be narrowed down to only those that are particularly effective, and can be combined into one code table as shown in FIG. In FIG. 5, the run length is defined only when Δθ = 0. Using the code table of FIG.
“0,0,0,0,0, -2, -2,0, + 3,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0, -1 ‥ ”
Is
“100 ・ 11101 ・ 11101 ・ 0 ・ 111100 ・ 1100 ・ 11011 ‥”
→ "10011101111010111100110011011 ‥" (29 bits)
Can be expressed as

  By the way, as compared with the fixed-length representation method of 10 bits per node, 10 bits × 25 nodes = 250 bits, and in this example, it can be seen that the compression is 29 ÷ 250 = 12%.

(Third embodiment)
In the third embodiment, an apparatus that performs the position information transmission method of the present invention will be described.

  FIG. 6 shows a position information transmitting / receiving device for exchanging event occurrence information on a road with another device 30 as an example of this device. The apparatus includes an offline processing unit 20 that generates a code table used for compression encoding of road shape data offline, and an online processing unit 10 that transmits traffic information using the code table data generated by the offline processing unit 20. The offline processing unit 20 includes a digital map database 22, a storage unit 21 for storing past traffic information, a code table calculation unit 23 for generating code table data used for compression encoding, and a generated code table data. And a code table database 24 for storing

  On the other hand, the online processing unit 10 includes a position information receiving unit 17 that receives the compression-encoded “road shape data” and “event position data” from the position information transmission unit 16 of the other device 30; A code data decompression unit 18 that decompresses (decodes) existing data, a digital map database 13 that stores digital map data, and performs map matching using decompressed road shape data and event position data. A map matching unit 14 for specifying an event position on a digital map, a digital map display unit 12 for superimposing and displaying the event position on the map, an event information input unit 11 for inputting event information that has occurred, and an event position Is determined, and “event position data” representing the event position by the relative position of the target road section is generated, and the shape data of the target road section is compressed using code table data 24. And a position information transmitting unit that transmits the generated “road shape data” and “event position data” to a position information receiving unit 17 of another device 30. 16 and are provided.

The flowchart of FIG. 7 shows the operation procedure of this device. In the off-line processing unit 20, the code table calculating unit 23, as shown in FIG.
Step 1: Refer to the past traffic information 21,
Step 2: Select a target road section for traffic information.
Step 3: The shape data of the target road section is resampled with a fixed length L, and nodes are set.

Step 4: Convert the position data of the node into a full curvature function expression,
Step 5: Calculate Δθ of each section / each node according to the statistical value calculation formula.
Step 6: Next, the appearance distribution of Δθ is calculated.
Step 7: Next, a continuous distribution of the same value is calculated.
Step 8: Create a code table based on the appearance distribution of Δθ and the continuous distribution of the same value,
Step 9: The completed code table is stored in the code table database 24.

  This processing procedure is defined by a program for causing the computer of the offline processing unit 20 to function as the code table calculation unit 23.

Further, in the online processing unit 10, the position information conversion unit 15 performs the following operations as shown in FIG.
Step 10: When traffic information is received from the event information input unit 11,
Step 11: Select a target road section including a position where a traffic event has occurred.
Step 12: Resample the shape data of the target road section with a fixed length L and set a node.

Step 13: Convert node position data to full curvature function expression,
Step 14: Calculate Δθ of each section / each node according to the statistical value calculation formula.
Step 15: Refer to the code table data 24 of the code table created for the target road section (or a code table created for a road whose shape is similar to the target road section) and code the shape data. Convert to representation.
Step 16: The encoded shape data of the target road section is transmitted together with the data of the event position represented by the relative information of the target road section.

  This processing procedure is defined by a program for causing the computer of the online processing unit 10 to function as the position information conversion unit 15.

  FIG. 8 shows transmitted road shape data (FIG. 8A) and event position data (FIG. 8B). The road shape data includes code table data, resampled section length L data, and compression-encoded shape data.

FIG. 9 shows a processing procedure on the receiving side that has received this data.
Step 20: When the position information receiving unit 17 receives the position information,
Step 21: The code data decompression unit 18 restores the code-represented data with reference to the code table included in the received data, and converts the shape data into a full curvature function.
Step 22: Next, shape data represented by latitude and longitude coordinates is reproduced.
Step 23: The map matching unit 14 specifies a target road section by executing a map matching between the reproduced shape and the road shape of its own digital map, and further, based on the event position data, a traffic in the target road section. Identify the event occurrence location.
Step 24: The digital map display section 12 superimposes and displays traffic information on the map.

  This processing procedure is defined by a program for causing the computer of the online processing unit 10 to function as the code data decompression unit 18 and the map matching unit 14.

  In this case, the code table used for the compression encoding is included in the transmission data and transmitted. However, since the transmitting and receiving sides have the same code table in advance, it becomes unnecessary to include the code table in the transmission data. .

  Also, here, an example is described in which the online processing unit 10 uses the code table data 24 created by the offline processing unit 20 in order to obtain the compression-encoded shape data. Each road shape is compression-encoded, and the shape data of each road section represented by the code is held in advance, and the online processing unit 10 acquires the traffic event occurrence information, and the shape data held in the offline processing unit 20 when the traffic event occurrence information is acquired. From among the selected, the encoded road shape data of the road section including the traffic event occurrence position is selected, and the traffic information indicating the traffic event occurrence position is represented by the relative position of the road section is generated. The generated road shape data and the generated traffic information may be transmitted to the receiving side.

  In this case, the off-line processing unit 20 resamples the shape data of the road section to be encoded with a fixed length L, calculates Δθ at each node, and calculates the Δθ Create a code table based on. Next, using the created code table, Δθ of each resampled coordinate point is converted into a code representation, and compression-encoded shape data is created and stored in a database. By repeatedly performing this process for each road section in the target area, it is possible to retain the shape data of each road section included in the target area that has been compression-encoded.

  As described above, the result of the resampling with the fixed length L for the road shape performed offline can be used in the online processing.

(Fourth embodiment)
In the fourth embodiment, a method for introducing high-compression road shape data by introducing an irreversible compression method will be described.

  In the transmission of audio data and image data, in order to improve the compression ratio, reduce the number of sampling points in the sense (visual / auditory) that is not bothersome, reduce the number of quantization digits of the measurement information, or compress the data. A process of falsifying the measurement information so as to increase the rate is performed. When these processes are performed, the original data cannot be completely restored on the receiving side. However, if there is no problem even if the data is slightly changed, the lossy compression process can be implemented by introducing such lossy compression processing. Can be greatly compressed.

  In the position information transmission method of the present invention, since the receiving side specifies the road shape by performing map matching, it is necessary to accurately convey the shape at the start point and the end point of the target road section and at places where erroneous matching is likely to occur. However, in other places, even if the shape to be transmitted is somewhat ambiguous, the receiving side can specify the original position. Therefore, also in the position information transmission method of the present invention, it is possible to increase the compression ratio of transmission data by introducing a method of irreversible compression.

Therefore, in the compression method of this embodiment, the data compression ratio is increased by the following method.
(1) Sampling points are reduced to the extent that erroneous matching does not occur. On a road with a large curvature and a sharp curve, the matching point may be off the road and erroneous matching may occur. Therefore, as shown in FIG. 10, the sampling section length L is set based on the magnitude of the curvature.
(2) The number of quantization digits expressing Δθ is reduced within a range where erroneous matching does not occur. For example, the minimum resolution is set to 2 °, and Δθ is quantized in this unit. In this case, the reproduced node position shifts right and left around the true value, so that the reproduced shape naturally becomes distorted. Therefore, the receiving side performs an interpolation process for smoothing the reproduction shape.
(3) The road shape is approximated by arcs and straight lines. When compression encoding is performed by giving a bias to a value in a differential expression based on the argument statistical prediction value _Sj , the bias is reduced to 0 in a road section having the same curvature represented by an arc or a straight line. Because of the concentration, the statistical bias is further increased, and the compression efficiency is dramatically increased. Therefore, by approximating the road shape with arcs and straight lines, the compression ratio is dramatically improved. In addition, the effect of the run-length encoding is improved.

  The methods (1), (2), and (3) may be performed alone or in combination. Here, a specific example in which the compression encoding is performed by applying the methods (1), (2), and (3) will be described.

The arc and straight line approximation of the road shape can be performed by linearly approximating the road shape expressed by the total curvature function. As in FIG. 1, a road shape points P _J on the road expressed in full curvature function, as shown in FIG. 11, the vertical axis θ (= Σθ _j), the horizontal axis L (= ΣL _i (where , Where _{Li is} constant)), and is displayed as a solid curve in a coordinate system. Approximating the road shape with an arc and a straight line means approximating this curve with a straight line indicated by a dotted line (θ = aL + b). In this coordinate system, a straight line having a slope of 0 (θ = b) represents a straight road shape, and a straight line having a slope of ≠ 0 (θ = aL + b) represents an arc road shape.

At the time of this approximation, the allowable error is determined along the road section by a method separately proposed by the present inventors (Japanese Patent Application Nos. 2001-129665 and 2001-132611). In this method, as the permissible errors, the permissible amount (permissible distance error) of an error relating to distance (distance error) and the permissible amount of permissible error (azimuth error) (permissible direction error) are included in each node included in the road shape. Or, it is set in the unit of a link so as to satisfy the following conditions.
(1) The allowable distance error is set small near the start point and the end point of the target road section.
(2) When parallel running roads are adjacent, the allowable distance error is set small.
(3) The allowable distance error is set to be small around an intersection where there is a connection road with a small intersection angle, such as an interchange entrance / exit road.
(4) The allowable azimuth error is set smaller as the distance from the surrounding road is shorter.
(5) Since there is a high possibility that the divergence of the azimuth error is large in a road shape portion having a large curvature, the allowable azimuth error is set small.

  The magnitude of the permissible error at each node is set separately for the left and right sides of the target road section. In the above proposal, a calculation method for quantitatively calculating the permissible error for each node is specifically shown.

  When the permissible error is determined along the road section, the road shape is approximated by an arc and a straight line so as to fall within the range of the permissible error, and as shown in FIG. To divide.

Next, the length of each resample section in each section is determined. The resample section length is determined for each section by the following equation according to the curvature a _j of each section j.
L _j = K × 1 / | a _j |
(K is a predetermined constant)

The value of L _j may be quantized. Assuming that a value that L _j can take by quantization is any one of eight values of, for example, 40/80/160/320/640/1280/2560/5120 meters, the value of L _j is encoded into 3 bits and transmitted. be able to.

At this time, the compression effect can be enhanced when the resample section length _Lj does not fluctuate between adjacent sections. FIG. 13 shows a determination procedure for determining the section length so that the section lengths are continuous when the curvature does not change much in order to suppress fluctuation in the resample section length. In order from the No. 1 segment (step 30), the curvature of a _j in each section j obtains the resample section length calculated value D _j (step 31), and resample section length calculated value D _j-1 of the adjacent section rate of change _{H j (= | D j -D} j-1 | / D j) the calculated (step 32), the ratio between the resample section length L _j-1 of the adjacent section I _j (= D _j /
L _j-1 ) is obtained (step 33). Then, by comparing the predetermined constant H _a and the change rate H _j resample section length calculated value, also, the ratio I _j of resample section length L _j-1 of the adjacent section, a predetermined value I _a1, compared with the I _a2, rate H _j is equal to or less than H _a, and, when I _j is between I _a1 and I _a2 is resampled segment adjacent sections resample section length L _j The length is set equal to the length L _j-1 (step 35). Here, resample to the resample section length calculated value D _j of is compared with resample section length L _j-1 of the adjacent section, the rate of change H _j resample section length calculated value D _j is less by continuing to set the section length to the same value, in order to prevent the divergence between the resample section length calculated value D _j and resample section length L _j occurs.

In step 34, when it is not, on the basis of the table of Figure 13 in which the relationship between the range and the section length of D _j, to determine the section length L _j from the value of D _j (step 36). This is performed for all sections (steps 37 and 38).
H _a A generally about 0.2 value, also, I _a1 is about 0.7, I _a2 is set to a value of about 2.0.

Next, as shown in FIG. 14, each section n is sampled at equal intervals with a resample section length L _n to obtain a node P _J, and the angle θ _{j of} P _J and the angle statistical prediction value S _j are calculated. The quantization value of the predicted value difference Δθ _j (= θ _j −S _j ) is calculated. Here, the argument θ _j−1 of the preceding node is used as the argument statistical predicted value S _j (S _j = θ _j−1 ).

The quantization value of [Delta] [theta] _j as ° the minimum resolution of [Delta] [theta] _j [delta] (minimum resolution = [delta]), obtaining the value.

At this time, since Δθ _j is set in units of δ, the node P _{J + 1} reproduced based on the distance L _n and the angle information Δθ _j from the preceding node P _J is the original road shape (or an approximated shape). It is not always located above. As shown in FIG. 15, when determining the next node P _{J + 1} from P _J, the candidate point of the node P _{J + 1} appears several by-taking quantized values [Delta] [theta] _j. From the candidate points, the next node P _{J + 1} is selected so that the value of Δθ becomes 0 as continuously as possible within the range of the allowable error. If the error between the selected node position and the true value (point on the original road shape) increases to near the limit of the allowable error by continuing such node selection, the node is reduced in a direction to reduce the error. Nodes need to be selected. In this case, nodes are selected so that Δθ is continuously zero.

FIG. 16 shows a procedure for selecting one candidate point from a plurality of candidate points P _{J + 1} (i) relating to one node P _{J + 1} .
Step 40: setting the candidate point P _{J + 1} (i) from P _J distance L _n, the position of Δθ = δ · i. Here, i is a quantized value of Δθ, and is 2m + 1 positive / negative integers composed of −m, ‥, −1, 0, 1, ‥, m and centered on 0.
Step 41: Distance D _i from each candidate point P _{J + 1} (i) to the closest point of the original road shape, and the intercept direction of the closest point and the intercept direction of candidate point P _{J + 1} (i) The error Δ 誤差_i is calculated.

Step 42: The evaluation value ε _i for each candidate point P _{J + 1} (i) is calculated by the following equation.
ε _i = α · (δ · | i |) + β · D _i + γ · | ΔΘ _i | + Ψ
α, β, γ: Predetermined coefficients
Ψ: Penalty value imposed when exceeding the allowable error range Step 43: The candidate point P _{J + 1} (i) having the smallest ε _i is adopted as the node P _{J + 1} .
The evaluation value epsilon _i until D _i and .DELTA..theta _i enlarged penalty value Ψ is added is minimized in the case of i = 0. Therefore, candidate points are adopted so that Δθ becomes zero.

Also, the fractional interval length D _n of the interval n, and processed as follows.
When L _n <L _{n + 1} : Section n is resampled at distance L _n , and if the remainder (fraction) of section n is shorter than L _n , the fraction and a part of section n + 1 are distance combined is resampled at L _n until within the time interval n + 1 so that L _n, is resampled from this point on the interval n + 1 at L _{n + 1.}
When L _n > L _{n + 1} : Section n is resampled at distance L _n , and if the fraction of section _n is shorter than L _n , L _{n +} from this point of section n over section n + 1 Resample with ₁ . As described above, the accuracy is not reduced for resampling with a short section length.

  If the minimum resolution δ ° of Δθ is increased, the number of digits of the angle is reduced, but the followability of the shape of the arc is deteriorated, the probability of Δθ = 0 is reduced, and the encoding compression effect is reduced. Conversely, when δ ° is reduced, the number of digits of the angle is increased, but the shape following of the arc is improved, the probability that Δθ = 0 is increased, and the encoding compression effect is also increased. Also, the run length compression effect is increased. In consideration of these points, it is necessary to determine the minimum resolution δ ° of Δθ actually used.

  Next, encoding of data in this case will be described. The predicted value difference Δθ of the corresponding node is encoded such that the data length becomes shorter around Δθ = 0. As for the run length, since most of continuous data is data of Δθ = 0, the run length of Δθ = 0 is encoded.

  Also, a section length change code indicating a change point of the resample section length is set. A special code is assigned to this section length change code, and the section length is defined by fixed bits (about 3 bits) immediately after the special code.

  Also, a reference point setting code indicating the identification code of the reference point node in each section is set. A special code is assigned to the reference point setting code, a fixed bit (e.g., 6 bits) immediately after the special code is used as a reference node number, and coordinates appearing after the reference node number are defined as a reference node (without additional bits). , A node number initial value may be determined in advance, and a node number system of +1 may be used each time this code is found.) Further, a special code is assigned as an EOD (End of Data) code indicating the end of data. This code is used as the end of the shape data string expression. FIG. 17 illustrates a code table used for this encoding.

FIG. 18A shows a procedure up to creating this code table off-line, and FIG. 18B shows a procedure up to transmitting traffic information online using the code table. I have. In FIG. 18A,
Step 50: Refer to the past traffic information,
Step 51: Select a target road section for traffic information.
Step 52: Calculate an allowable error range along the target road section.

Step 53: Convert the nodes of the target road section into a full curvature function expression,
Step 54: The shape vector of the target road section is approximated to an arc and a straight line.
Step 55: determining the resample length L _n of each section n approximating an arc or a straight line.
Step 56: The shape data of the target road section is quantized and resampled by Ln to set a node.
Step 57: Calculate Δθ of each section / each node according to the statistical value calculation formula.
Step 58: The appearance distribution of Δθ is calculated.
Step 59: Calculate a continuous distribution of the same value.
Step 60: Create a code table based on the appearance distribution of Δθ and the continuous distribution of the same value,
Step 61: Store the completed code table in the code table database 24.

The online processing in FIG. 18B is as follows.
Step 62: When traffic information is received from the event information input unit 11,
Step 63: Select a target road section including the position where the traffic event has occurred.
Step 64: Calculate an allowable error range along the target road section.
Step 65: Convert the nodes of the target road section into a full curvature function expression,
Step 66: The shape vector of the target road section is approximated to an arc and a straight line.
Step 67: determining the resample length L _n of each section n approximating an arc or a straight line.

Step 68: Quantize and resample the shape data of the target road section with Ln to set nodes.
Step 69: Calculate Δθ of each section / each node according to the statistical value calculation formula.
Step 70: Referring to the code table, the shape data is converted into a code representation.
Step 71: Transmit the encoded shape data of the target road section together with the traffic information.

  Here, in the online processing, an example in which only the code table data created in the offline processing is used has been described. However, as described in the third embodiment, each road shape in the target area is encoded by the offline processing. In the online processing, when traffic event occurrence information is input in the online processing, the shape data of the road section including the traffic event occurrence position is extracted from the shape data generated in the offline processing. Selected encoded road shape data, generates traffic information indicating the position of the traffic event by the relative position of the road section, and receives the selected encoded road shape data and the generated traffic information. It may be transmitted to the side. As described above, the result of resampling at a fixed length L for a road shape performed offline can be used in online processing.

  FIG. 19 shows the transmitted road shape data. The data includes code table data and encoded shape data, and includes encoded shape data such as Δθ, a reference node of each section, and a sample section length.

  FIG. 20 schematically shows data exchanged between transmission and reception. On the transmitting side, as shown in FIG. 20 (a), a node position after quantization and resampling is calculated to represent a road shape, and as shown in FIG. 20 (b), data representing this node position is received. Sent to the side. On the receiving side, as shown in FIG. 20C, the received data is smoothed to reproduce the shape. In this case, interpolation using a B-spline (an interpolation curve such as a Bezier spline / Bezier curve is also possible) or smoothing by a smoothing function is performed. In addition, the intercepts of the generated interpolation points are also distributed on average.

FIG. 21 shows the procedure on the receiving side.
Step 80: Upon receiving the location information,
Step 81: Referring to the code table, the shape data of the code expression is converted into a total curvature function.
Step 82: Next, the data is converted into latitude / longitude coordinates, smoothing / interpolation processing is performed, and the shape data is reproduced.
Step 83: Obtain the reference node position,
Step 84: Perform map matching, identify the target road section,
Step 85: Reproduce traffic information.

  As described above, the shape data is highly compressed using the irreversible compression method described in this embodiment, so that the amount of transmission data can be significantly reduced.

  The arc / straight line approximation of the shape data expressed by the full curvature function can be performed simultaneously with the quantization resampling, instead of approximating the shape in advance as described herein.

  The logic for determining the resample section length and the procedure for determining the quantized resample described above can also be applied to a case where the shape data is not approximated by an arc.

(Fifth embodiment)
In the fifth embodiment, a method of encoding road shape data without resampling coordinate points will be described.

As described above with reference to FIG. 34, the coordinate points (P _J ) arranged on the road are uniquely specified by two dimensions of the distance and the angle from the adjacent coordinate point (P _J-1 ). be able to. In the first to fourth embodiments, the coordinate point position is resampled so that the distance is constant, and only the angle is encoded to reduce the amount of transmission data. However, in this case, resampling processing is required.

  On the other hand, when encoding the road shape data using the nodes and interpolation points included in the road shape of the digital map as the coordinate points, the resampling process is not required. However, in this case, since the distance between the nodes and the interpolation points is not constant, it is necessary to encode the angle and the distance.

FIG. 22 illustrates a method for encoding both angles and distances. For the coding of the angle is the same as the first embodiment, which is a difference angle information of each node (including the interpolation point) P _J, the deflection angle theta _j and deviation angle statistical predicted value S _j prediction expressed by the value difference [Delta] [theta] _j, the [Delta] [theta] _j for example, 1 ° units quantized by (such as 2 ° units, may also be other resolutions), create a code table of [Delta] [theta] based on the occurrence frequency of the magnitude of the quantized [Delta] [theta] _j I do. At this time, the argument statistical predicted value S _j is defined as, for example, S _j = θ _j-1 or S _j = (θ _j-1 + θ _j-2 ) / 2.

FIG. 23B shows an example of the code table of Δθ created in this way. This table is the same as the code table of the first embodiment (FIG. 2). The angle information (Δθ _j ) of each node is variable-length encoded using the code table of Δθ.

On the other hand, distance encoding is performed as follows. First, distance information of each node P _J is represented by a predicted value difference ΔL _j (= L _j −T _j ) which is a difference between a distance L _j to an adjacent node P _{J + 1} and a distance statistical predicted value T _j , ΔL _j is quantized in, for example, 10 m units (50 m units, 100 m units, or other resolutions may be used). At this time, the distance statistical prediction value _Tj is defined as, for example, _Tj = _Lj-1 or _Tj = ( _Lj-1 + _Lj-2 ) / 2.

Next, a code table of ΔL is created based on the frequency of occurrence of the quantized ΔL _j . FIG. 23A shows an example of the code table of ΔL created in this way. The additional bits in this code table are bits added to indicate the sign of ΔL. When ΔL ≠ 0, 0 is added if ΔL is positive, and 1 is added if ΔL is negative. Therefore, if we define T _j = L _j−1 ,
L _j is longer than _{L j-1 (L j -L} j-1> 0) when 0
L _j is shorter than _{L j-1 (L j -L} j-1 <0) when 1
Is added.

Using the code table of ΔL, distance information (ΔL _j ) of each node is variable-length coded. The order of the data at the time of encoding the distance and the angle is determined in advance as ΔL _j → Δθ _j → ΔL _{j + 1} → Δθ _{j + 1} → ‥. Now, the sequence of ΔL−Δθ is “0 ・ 0_0 ・ 0_0 ・ -2_ + 2 ・ -2_0 ・ + 3_-5 ・ 0_0 ・ 0_0 ・ +6”
In this case, this data string is subjected to variable-length encoding using the code tables shown in FIGS.
“0 ・ 0_0 ・ 0_0 ・ 1011_1010 ・ 1011_0 ・ 11000_11101 ・ 0_0 ・ 0_0 ・ 111100”
→ "00000101110101011011000111010000111100" (38 bits)
If the distance component is represented by a fixed length of 8 bits and the angle component is represented by a fixed length of 10 bits, (8 bits + 10 bits) × 8 nodes = 144 bits are required, and the data amount is compressed to 26% by variable length coding. Can be.

FIG. 24A shows a processing procedure when these code tables are created off-line. First, referring to the past traffic information (step 90), a target road section of the traffic information is selected (step 91). The position data of the nodes included in the target road section is converted into a total curvature function expression (step 92), and ΔL _j and Δθ _j of each node in each section are calculated according to the statistical value calculation formula (step 93). Next, the appearance distributions of ΔL _j and Δθ _j are calculated (step 94), a code table of ΔL is created based on the appearance distribution of ΔL _j , and a code table of Δθ is created based on the appearance distribution of Δθ _j. Create (steps 95 and 96).

FIG. 24B shows a processing procedure for encoding road shape data using the created code table in order to transmit traffic information. When the traffic information is received (step 97), a target road section including the position where the traffic event has occurred is selected (step 98). The position data of the nodes included in the target road section is converted into a total curvature function expression (step 99), and ΔL _j and Δθ _j of each node in each section are calculated according to the statistical value calculation formula (step 100). Next, with reference to the code table data of the code table created for the target road section (or the code table created for the road whose shape is similar to the target road section), ΔL _{j of} each node is referred to. And Δθ _j are converted into a code representation (step 101). The encoded shape data of the target road section is transmitted together with the event position data represented by the relative information of the target road section (step 102).

FIG. 25 shows transmitted road shape data (FIG. 25A) and event position data (FIG. 25B). The road shape data includes code table data, the absolute coordinates of the start node p1 of the section (nodes p1 and p2) represented by the code, the absolute azimuth of the node p1, the distance L from the node p1 to the next node, and the node p1. To p2 (bit strings obtained by encoding ΔL _j and Δθ _j ).

  The receiving side that has received this data converts the code-represented data into a full curvature function with reference to a code table, and reproduces the road shape data, as in the processing flow of FIG. Next, the target road section is specified by executing map matching between the reproduced shape and the road shape of the own digital map, and a traffic event occurrence position in the target road section is specified from the event position data.

  As described above, in the method of this embodiment, the data of the angle and the distance specifying the coordinate point are both variable-length coded without resampling the coordinate point, and the transmission data amount of the road shape data is reduced. be able to.

(Sixth embodiment)
In the sixth embodiment, a method will be described in which the coordinate point position is resampled so that the angle component becomes constant on the road, and only the distance component is encoded.

As described above with reference to FIG. 34, the coordinate points (P _J ) arranged on the road are uniquely specified by two dimensions of the distance and the angle from the adjacent coordinate point (P _J-1 ). be able to. In the first to fourth embodiments, of the two dimensions, the coordinate point position is resampled so that the distance is constant, thereby encoding only the angle to reduce the amount of transmission data. In the sixth embodiment, conversely, the coordinate point position is resampled so that the angle becomes constant, and the amount of transmission data is reduced by encoding only the distance.

FIG. 26 shows the resampled coordinate points when the angle information is fixed (declination angle θ = constant) and the distance information is encoded. The resampling process of the shape data is performed as follows.
(1) over the road shape toward the start node P ₀ to the end node trace, set the next node P ₁ to the position reaching to the angle deviation angle is predetermined theta (or - [theta]).
(2) However, when you are tracing in (1), before the deflection angle reaches the theta (or - [theta]), when it reaches a distance L _max the distance from the starting node P ₀ is determined in advance, the setting the next node P ₁ to the position.
(3) Starting from the node P ₁ determined in the above (1) or (2), the next node P ₂ is determined by applying the rules of the above (1) and (2), and this is sequentially repeated. Determine P ₃ , ‥, P _J , ‥.

The resampled distance information at each node P _J is represented by a predicted value difference ΔL _j (= L _j −T _j ) which is a difference between the distance L _j to the adjacent node P _{J + 1} and the distance statistical predicted value T _j. , ΔL _j are quantized in, for example, 10 m units (50 m units, 100 m units, or other resolutions may be used). At this time, the distance statistical prediction value _Tj is defined as, for example, _Tj = _Lj-1 or _Tj = ( _Lj-1 + _Lj-2 ) / 2. Next, a code table of ΔL is created based on the frequency of occurrence of the quantized ΔL _j . At this time, a continuous distribution of ΔL _j may be calculated, and a code table incorporating run-length encoding may be created.

FIG. 27 shows an example of the code table of ΔL created in this way. In this code table, when ΔL = 0, one bit (0 when positive, 0 when negative, 1 when negative) of the argument θ is added to the code as an additional bit, Also, when ΔL ≠ 0, a total of two bits, one bit for indicating the sign of the declination θ and one bit for indicating the sign of ΔL (0 if ΔL is positive, 1 if ΔL is negative) Is added to the code as an additional bit. Therefore, when T _j = L _j−1 is defined,
When ΔL ≠ 0,
When L _j is longer than L _j−1 (L _j −L _j−1 > 0), 0 is added as an additional bit representing the sign of ΔL.
When L _j is shorter than L _j−1 (L _j −L _j−1 <0), 1 is added as an additional bit indicating the sign of ΔL, and
When the azimuth of P _j-1 → P _j is on the left side of the azimuth of P _j-2 → P _j-1 (left turn), 0 is added as an additional bit representing the sign of θ.
When the azimuth of P _j-1 → P _j is on the right side of the azimuth of P _j-2 → P _j-1 (turn right), 1 is added as an additional bit indicating the positive or negative of θ.

  In the fourth embodiment, an example in which the distance component (resample section length) is changed depending on the section when the coordinate point is resampled so that the distance component becomes constant has been described. Can be switched depending on the section even when the resampling is performed so that is constant. In this case, as in the fourth embodiment, the value of θ in each section can be identified on the code-converted shape data sequence, as in the fourth embodiment.

  FIG. 28 (a) shows a processing procedure when this code table is created off-line, and FIG. 28 (b) encodes road shape data using the created code table and transmits traffic information. It shows a processing procedure when performing. These procedures are different from the procedure described in the third embodiment (FIG. 7) in that the shape data of the target road section is resampled at a fixed angle θ (or −θ) instead of being resampled at a fixed length L. (Steps 112 and 121), the point where ΔL is calculated instead of calculating Δθ of each resampled node (step 114 and step 123), and the sign of Δθ based on the distribution of Δθ. The difference is that a code table of ΔL is created based on the distribution of ΔL instead of creating a table (steps 115 and 117), but the other procedures are the same.

FIG. 29 shows transmitted road shape data. This road shape data includes information on the sample angle θ instead of the sample section length L, as compared with the road shape data (FIG. 8A) described in the third embodiment, and is used as encoded data. , Δθ _j is replaced by a bit string encoding ΔL _j , but the other points are the same.

  The receiving side that has received this data converts the code-represented data into a full curvature function with reference to a code table, and reproduces the road shape data, as in the processing flow of FIG. Next, the target road section is specified by executing map matching between the reproduced shape and the road shape of the own digital map, and a traffic event occurrence position in the target road section is specified from the event position data.

  As described above, in the method of this embodiment, the coordinate point position is resampled so that the angle component is constant on the road, and only the distance component is subjected to variable-length coding to reduce the transmission data amount of the road shape data. be able to.

(Seventh embodiment)
In the encoding method according to the seventh embodiment, in order to convert a road shape into shape data having a statistical deviation, as an expression method of angle information, either an expression using a declination or an expression using a prediction value difference is selected. I can do it.

As described above with reference to FIG. 35, angle information of the coordinate points, expressed in deviation angle theta _j (FIG. 35 (b) (b ') ), and the predicted value difference [Delta] [theta] _j of the deviation angle theta _j In the case where any of the expressions (FIGS. 35 (c) and (c ′)) is adopted, the road shape data can be converted into data having a statistical bias.

The greater the statistical bias, the greater the effect of reducing the data amount by variable-length coding. Comparing the case where the angle information of the coordinate point is represented by the declination θ _j and the case where the angle information is represented by the predicted value difference Δθ _j of the declination θ _j , the latter is generally larger in statistical deviation.

However, as shown in FIG. 30, in the case of the road 40 in which a straight line continues for a while between the curves, when expressed by the predicted value difference Δθ _j of the declination,
0, ‥, 0, θ ₁ , -θ ₁ , 0, ‥, 0, θ ₂ , -θ ₂ , 0, ‥
However, when represented by the argument θ,
0, ‥, 0, θ ₁ , 0,0, ‥, 0, θ ₂ , 0,0, ‥
When the angle information of the coordinate point is represented by the argument θ _j , a statistical bias can be provided as compared with the case where the angle information is represented by the predicted value difference Δθ _j . Thus, depending on the road shape, who expresses the angle information of coordinate points resampled at a constant distance L in polarization angle theta _j is, if suitable for variable-length coding there.

  In the method of this embodiment, the data size when the road shape is represented by the argument θ and subjected to variable length coding and the data size when the road shape is represented by the prediction value difference Δθ of the argument and subjected to variable length encoding are represented by By comparison, the encoded data having the smaller data size is transmitted.

Therefore, first, a declination θ code table for expressing a road shape by a declination θ _j and performing variable length encoding, and a variable length encoding by expressing a road shape by a prediction value difference Δθ _j of the declination θ _j And a Δθ code table to be created.

  FIG. 31A shows a procedure for creating a declination θ code table, and FIG. 31B shows a procedure for creating a Δθ code table. The procedure in FIG. 31B is the same as the procedure (FIG. 7A) in the third embodiment. Further, the procedure of FIG. 31A is different only in that Δθ in the procedure of FIG. 31B is replaced by the argument θ.

FIG. 32 shows a processing procedure when road shape data is encoded using these code tables created offline and traffic information is transmitted.
Step 130: Upon receiving traffic information,
Step 131: Select the target road section including the position where the traffic event occurred,
Step 132: Resample the road shape data of the target road section with a fixed length L and set a node,
Step 133: Convert the position data of the set node into a full curvature function expression.

Step 134: Next, the code data of θ is created with reference to the code table of θ, and the data size (A) is calculated.
Step 135: Next, the code data of Δθ is created with reference to the code table of Δθ, and the data size (B) is calculated.
Step 136: Compare the data size (A) with the data size (B), adopt an angle expression having a smaller data size, and use the “angle expression identification flag” indicating the adopted angle expression in the shape data to be transmitted. And “encoded data” in the adopted angle expression are set.
Step 137: The encoded shape data of the target road section is transmitted together with the data of the event position represented by the relative information of the target road section.

  FIG. 33 shows the transmitted road shape data. The road shape data includes information of an “angle expression identification flag” (0 when the expression using the argument θ is used, 1 when the expression using the predicted value difference Δθ is used) representing the adopted angle expression, and It contains information on “encoded data” in the adopted angle expression.

  On the receiving side receiving this data, based on the information specified by the “angle expression identification flag”, from the “encoded data” that is code-represented, refer to the code table and restore θ or Δθ, and Convert the position data of the node to a full curvature function. Subsequent processing is the same as that of the third embodiment and the like. The road shape data is reproduced, and the target road section is specified by executing map matching between the reproduced shape and the road shape of the own digital map. From the event position data, a traffic event occurrence position in the target road section is specified.

  As described above, in the method of this embodiment, the transmission data amount can be further reduced by selecting either the expression based on the argument or the expression based on the difference of the predicted value as the expression method of the angle information.

  The encoding method according to the present invention can be applied to compression of the map data itself. Further, the present invention can be applied to exchange of map data on the Internet (eg, a client-server type map display system using a vector map), a map data distribution service, and the like.

  Also, when the traveling locus data is transmitted from the on-board unit of the vehicle to the center for emergency notification or floating car data (FCD), the data can be compressed using this encoding method.

  Also, when the vector shape is compressed by the spline compression method and transmitted as data of each node sequence, the encoding method of the present invention may be applied to compress the data of the node sequence using a code table. It is possible.

  The encoding method according to the present invention can also be applied to a case where shape data of an area (polygon) on a digital map is transmitted. For example, when a polygon is specified and the weather forecast of the area is transmitted, by transmitting the shape data of the boundary line of the polygon, the receiving side can specify the polygon. When transmitting the shape data of the boundary line, the amount of transmission data can be compressed by applying the encoding method of the present invention. At this time, if it is not necessary to precisely specify the polygon shape as in the application region of the weather forecast, the receiving side can omit the matching processing with the shape on the digital map.

The illustrated code table is merely an example, and is not optimal. In practice, it is necessary to investigate the distribution of variables (θ _j , Δθ _j , L _j, etc.) and create a code table using a Huffman tree or the like.

  There are various types of encoding techniques, such as a fixed character compression method, a run-length method, a Shannon Fano encoding method, a Huffman encoding method, an adaptive Huffman encoding method, an arithmetic encoding method, and a dictionary method (LHA method). It is also possible to use these encoding methods. Although the case where the code table is generated offline has been described here, online coding becomes possible by using an adaptive Huffman coding method or an arithmetic coding method.

  As is clear from the above description, the encoding method of the present invention can efficiently compress the amount of vector-shaped data in a digital map. Therefore, the position information transmission method and apparatus of the present invention can significantly reduce the amount of transmission data when transmitting the vector shape of the digital map. On the receiving side, the transmitted vector shape can be accurately specified by restoring the shape data from the received data and performing map matching.

FIG. 7 is a diagram illustrating resampled nodes when the encoding method according to the first embodiment is applied; FIG. 3 is a diagram illustrating a code table in the encoding method according to the first embodiment; FIG. 9 is a diagram illustrating a run-length code table used in the encoding method according to the second embodiment; FIG. 9 is a diagram illustrating a code table of Δθ used in the encoding method according to the second embodiment; FIG. 9 is a diagram showing a code table of Δθ in consideration of run length used in the encoding method according to the second embodiment; FIG. 9 is a block diagram illustrating a configuration of a device that implements the position information transmission method according to the third embodiment; Flow chart showing a code table creation procedure (a) and a shape data creation processing procedure (b) in the encoding method of the third embodiment. FIG. 9 is a diagram illustrating a configuration of transmission data in the position information transmission method according to the third embodiment; FIG. 14 is a flowchart showing a processing procedure on the receiving side in the position information transmission method according to the third embodiment; FIG. 14 is a diagram illustrating a relationship between a sample section length and a curvature of shape data in the encoding method according to the fourth embodiment; FIG. 9 is a view for explaining arc / straight line approximation in the encoding method according to the fourth embodiment; FIG. 14 is a diagram showing sections divided by the encoding method according to the fourth embodiment; FIG. 13 is a flowchart showing a method for determining a resample section length in the encoding method according to the fourth embodiment; FIG. 9 is a view for explaining quantization resampling in the encoding method according to the fourth embodiment; FIG. 9 is a diagram for describing a candidate point of a next node in the encoding method according to the fourth embodiment; FIG. 14 is a flowchart showing a next node determination procedure in the encoding method according to the fourth embodiment; FIG. 14 is a diagram showing a code table in the encoding method according to the fourth embodiment; Flow chart showing a code table creation procedure (a) and a shape data creation processing procedure (b) in the encoding method of the fourth embodiment. FIG. 14 is a diagram illustrating a configuration of transmission data in the position information transmission method according to the fourth embodiment; The figure which shows typically the transmission of the data by the encoding method of 4th Embodiment, A flow chart showing a processing procedure on the receiving side in the position information transmission method of the fourth embodiment, FIG. 14 is a diagram illustrating node position, distance, and angle information when the encoding method according to the fifth embodiment is applied; FIG. 14 is a diagram illustrating a code table used in the encoding method according to the fifth embodiment; Flow chart showing a code table creation procedure (a) and a shape data creation processing procedure (b) in the encoding method of the fifth embodiment. FIG. 14 is a diagram illustrating a configuration of transmission data in the position information transmission method according to the fifth embodiment; FIG. 14 is a diagram illustrating node position, distance, and angle information when the encoding method according to the sixth embodiment is applied; FIG. 14 is a diagram illustrating a code table used in the encoding method according to the sixth embodiment; A flow chart showing a code table creation procedure (a) and a shape data creation processing procedure (b) in the encoding method of the sixth embodiment. The figure which shows the structure of the transmission data in the position information transmission method of 6th Embodiment, The figure which shows the road shape suitable for applying the encoding method of 7th Embodiment, Flow chart showing a procedure (a) for creating a θ code table and a procedure (b) for creating a Δθ code table in the encoding method of the seventh embodiment. FIG. 17 is a flowchart showing a shape data creation processing procedure in the encoding method according to the seventh embodiment; FIG. 17 is a diagram illustrating a configuration of transmission data in the position information transmission method according to the seventh embodiment; A diagram illustrating a distance and an angle specifying a coordinate point, (A) (a ') Diagram showing the full curvature function expression of the shape data, (b) (b') diagram showing the declination expression of the shape data, (c) (c ') predicted value of the declination of the shape data Diagram showing a differential expression, FIG. 7 is a diagram showing a data configuration in a conventional position information transmission method.

Explanation of reference numerals

10, 30 Online processing unit
11 Event information input section
12 Digital map display
13, 22 Digital map database
14 Map matching section
15 Location information converter
16 Location information transmitter
17 Location information receiver
18 Code data decompression unit
20 Offline processing unit
21 Past traffic information
23 Code table calculator
24 Code table data
40 road

Claims (32)

In an encoding method for encoding data representing a shape vector on a digital map,
Arithmetic processing is performed on the position information of each node of the node sequence representing the shape vector, the position information is converted into statistically biased data, and the data is encoded to reduce the data amount. Encoding method of shape vector data to be used.   The position information of the node is expressed by information of a distance from an adjacent node and an angle of a straight line extending from the adjacent node, and the distance or the angle is represented by data having a statistical bias, and the data is encoded. 2. The method according to claim 1, wherein the encoding is performed.   The method according to claim 1, wherein the position information of the node is converted into difference data from a statistical prediction value, and the difference data is encoded.   4. The method for encoding shape vector data according to claim 2, wherein the angle information is represented by an argument, and the argument is encoded.   4. The method according to claim 2, wherein the angle information is represented as a difference with respect to a statistical predicted value of the argument, and the difference is encoded. 5.   4. The method according to claim 2, wherein the distance information is represented by a difference from a statistical predicted value of the distance, and the difference is encoded. 5.   The node according to any one of claims 1 to 6, wherein the node is resampled so that at least one element of the position information of the node takes a constant value within a predetermined section of the shape vector. A coding method of the described shape vector data.   8. The method according to claim 7, wherein the node is resampled to a position where the distance from an adjacent node is equal, and the position information of the node is expressed only by angle information. .   The shape vector data according to claim 7, wherein the node is resampled to a position where a declination of a straight line extending from an adjacent node takes a certain angle, and the position information of the node is expressed only by distance information. Encoding method.   The method according to any one of claims 7 to 9, wherein the shape vector is divided into a plurality of sections, and the fixed value is set for each section.   The method according to claim 10, wherein a fixed resample interval length is set for each of the intervals, and the nodes are resampled at intervals of the resample interval length.   7. The method according to claim 1, wherein arithmetic processing is performed on position information of a node or an interpolation point included in the shape vector of the digital map, and the position information is converted into data having a statistical bias. 3. A method for encoding a shape vector according to item 1.   2. The shape vector encoding method according to claim 1, wherein arithmetic processing is performed on the information of the node sequence of the spline function that approximates the shape vector to convert the information into data having a statistical bias.   14. The method for encoding shape vector data according to claim 1, wherein the data is subjected to variable length encoding to reduce the data amount.   15. The method for encoding shape vector data according to claim 1, wherein the data is run-length encoded to reduce the amount of data.   12. The method according to claim 11, wherein the resampling section length is set according to the magnitude of the curvature of the shape vector.   The encoding of the shape vector data according to claim 1, wherein an allowable error of the shape vector is calculated, and the number of quantization digits when encoding the data is reduced within a range not exceeding the allowable error. Method.   17. The method according to claim 16, wherein the shape vector is approximated by an arc and a straight line, and the resample section length is set for each section of the approximated arc or the straight line. The transmitting side transmits shape data representing the shape vector on the digital map, and the receiving side performs map matching based on the received shape data to specify the shape vector on the digital map of the digital map. In the transmission method,
The transmitting side transmits the shape vector data encoded by the encoding method according to any one of claims 1 to 18, and the receiving side decodes the received data to reproduce the shape, and corresponds to the reproduced shape. A position information transmission method characterized in that a shape vector to be specified is specified by map matching.   20. The position according to claim 19, wherein the transmitting side transmits the code table used for encoding in the data to be transmitted, and the receiving side reproduces a shape using the received code table. Information transmission method.   20. The position information transmitting method according to claim 19, wherein the transmitting side sets the resample section length within a range in which erroneous matching does not occur on the receiving side.   20. The method according to claim 19, wherein the transmitting side reduces the number of quantization digits within a range where no erroneous matching occurs on the receiving side.   20. The position information transmitting method according to claim 19, wherein the transmitting side approximates the shape vector with an arc and a straight line within a range in which erroneous matching does not occur on the receiving side.   20. The position information transmitting method according to claim 19, wherein the transmitting side prepares and holds a code table used for encoding in advance, and encodes the shape vector using the code table.   25. The position information according to claim 24, wherein the transmitting side previously holds the data of the shape vector encoded using the code table, and selects data to be transmitted from the held data. Transmission method.   20. The position information transmitting method according to claim 19, wherein a transmitting side transmits data representing a shape of a polygon on a digital map as the encoded shape vector data. In a transmission device that transmits shape data representing a shape vector on a digital map for specifying a shape by map matching to a receiving side,
Arithmetic processing is performed on the position information of each node in the node sequence representing the shape vector on the digital map, the position information is converted into statistically biased data, and the data is distributed based on the appearance distribution of the data. Code table calculating means for generating a code table used for encoding;
A transmitting apparatus comprising: position information converting means for encoding position information of each node of a shape vector to be transmitted to a receiving side using the code table and generating shape data to be transmitted to the receiving side. In a transmission device that transmits shape data representing a shape vector on a digital map for specifying a shape by map matching to a receiving side,
Position information conversion means for encoding the position information of each node of the shape vector to be transmitted to the receiving side using a code table generated by another device and generating shape data to be transmitted to the receiving side. Transmitting device. A receiving device that receives shape data representing a shape vector on a digital map from a transmitting side, performs map matching, and specifies the shape vector on its own digital map,
Code data decoding means for decoding the encoded received data to reproduce shape data represented by position information on a digital map,
A receiving device comprising: a map matching unit that performs map matching using the reproduced shape data and specifies the shape vector on its own digital map. A program of a transmission device that transmits shape data representing a shape vector on a digital map for specifying a shape by map matching to a receiving side,
On the computer,
A procedure of resampling the shape vector on the digital map with a fixed length to set nodes,
A procedure of performing arithmetic processing on the position data of the node and converting the data into data having a statistical bias,
Generating a code table used for encoding the data based on the distribution of occurrence of the statistically biased data. A program of a transmission device that transmits shape data representing a shape vector on a digital map for specifying a shape by map matching to a receiving side,
On the computer,
A step of setting a node by resampling the shape vector with a fixed length,
A procedure of performing arithmetic processing on the position data of the node and converting the data into data having a statistical bias,
Converting the statistically biased data into a code representation by referring to a code table. A program of a receiving device that receives shape data representing a shape vector on a digital map from a transmitting side, performs map matching, and specifies the shape vector on its own digital map,
On the computer,
A step of decoding the encoded received data with reference to a code table,
A procedure for reproducing the shape data represented by the position information on the digital map from the decrypted data,
Performing map matching using the reproduced shape data and specifying the shape vector on its own digital map.

Images