Algorithm for representing roads or intersections

Abstract

To provide an intersection representation algorithm for allowing a self-driving car to appropriately recognize shapes of intersections and symbols on the intersections.SOLUTION: An algorithm disclosed herein is configured to roll a 25x25x25 cm cube from a reference point (a stave at an end of a ladder road marking) of an intersection and regard a produced closed loop as an intersection shape, and to roll the cube several times on a spot to represent the presence of a symbol.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an algorithm that represents the shape of a road or intersection and the presence or absence of a symbol placed on the road or intersection, which can be applied to automatic valet parking and automatic driving. [Background technology]

[0002] By connecting intersections using ladder road markings (Patent Document 12) painted with paint containing an ultraviolet reflective agent, it has become possible for an ultraviolet line tracing car to drive automatically except at intersections. The UV-reflective lines can be made transparent, allowing them to coexist with conventional autonomous vehicle perception systems. Transparent UV reflective paint is available for sale, though the price is unknown. Intersection information can be stored in the ladder road marking Fumi-san, and the mark written on Fumi-san can be used as a reference point for the intersection shape. A conventional method of representing the shape of an intersection would be to prepare a matrix and fill every square with the number 0 or 1. 0 represents nothing, 1 represents the presence of a line, and 2 or 3 are used to represent symbols such as a pedestrian crossing. If you have a 75x75 grid, and each grid represents 2 bits of information, you would need 75x75x2=11250 bits. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent application 2022-195491 [Patent Document 2] Patent application 2023-91455 [Patent Document 3] Patent application 2023-91456 [Patent Document 4] Patent application 2023-130625 [Patent Document 5] Patent application 2023-151777 [Patent Document 6] Patent application 2023-208096 [Patent Document 7] Intention 2023-025713 [Patent Document 8] Intention 2023-025714 [Patent Document 9] Intention 2023-026595 [Patent Document 10] Patent application 2024-008905 [Patent Document 11] Patent application 2024-078245 [Patent Document 12] Patent application 2024-087548 [Patent Document 13] Patent application 2024-135157 [Patent Document 14] Patent application 2024-137933 [Non-patent literature]

[0004] [Non-Patent Document 1] National Institute for Land and Infrastructure Management 6.2 RFID tag verification experiment Summary of the Invention [Problem to be solved by the invention]

[0005] According to Non-Patent Document 1, self-driving cars that use a method of embedding RFID tags in roads can only travel at about 120km / h even with active tags. When using RFID to transmit information about intersection shapes and traffic lights to self-driving cars, it is better to transmit as little information as possible. This is because when there is only a little information to transmit, it is possible to check multiple times just to be sure, or to make it possible to read even if some information is missing. To provide an algorithm for expressing symbols such as intersection shapes and pedestrian crossings installed at intersections in a manner other than compression. [Means for solving the problem]

[0006] An algorithm that represents a line or a closed loop from a point, The algorithm is: The closed loop, such as an intersection, is a contour line, and the single line with thickness, such as a road, is a center line. From one point, it is expressed with continuous 1-bit information of rotation and forward movement, and is characterized by expressing the end or the first point or a specific symbol. The characteristic of this is that the amount of rotation is not constant when rotating continuously. The feature of expressing a symbol on the line or the closed loop is that the symbol returns to its original direction by continuously rotating; An algorithm that represents a line or closed loop seen from a single point. Use the following.

[0007] Also, The algorithm for expressing a line or a closed loop as viewed from one point according to claim 1, characterized in that the rotation amount is 90 degrees for the first rotation and 180 degrees for the second rotation. Use the following.

[0008] Also, The algorithm for expressing a line or closed loop seen from one point, as shown in claim 2, characterized in that the amount of advancement in one go is 25 cm. Use the following.

[0009] Symbols are objects on intersections that should be known for autonomous driving, such as the positions of the four corners of a crosswalk. In addition to objects, when this symbol is present, it can also be a symbol that means something like "rise 25 cm in height."

[0010] The reason why the amount of advance in one go is 25cm is because the width of the road changes in 25cm increments.

[0011] It says that "the amount of rotation is 90 degrees for the first rotation and 180 degrees for the second rotation," but by turning left or right the first time and then turning again to face the opposite direction, skipping the back, the capacity of the algorithm of the present invention is reduced. When expressing a single line, there is almost no backward movement.

[0012] The rotation in "continuous 1-bit information of rotation and forward movement" includes rotation and forward movement. Although the program becomes more complicated, even if you rotate and move forward at the same time, a small closed loop with zero area will occur, but a line can be expressed. In that case, the self-driving car that reads it needs to process the small closed loop with zero area to eliminate it. [Brief description of the drawings]

[0013] [Figure 1] Excel example of converting intersection shape using this algorithm [Diagram 2] Intersection shape example 1 [Diagram 3] Intersection shape example 2 [Figure 4] Intersection shape example 3 [Diagram 5] Overall view of the location where RFID will be installed [Figure 6] Road shape preservation example 1 [Figure 7] Intersection shape example 4 [Figure 8] Example of road shape near interchange DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] First, let me explain the general flow. This invention is an algorithm that expresses a single line or closed loop seen from a single point. Using a single point (Fumisan at the end of the ladder road marking) as the base point, the algorithm chooses between rotating or moving forward. The idea is to prepare a 25cm box and roll it, treating the place it passes through as a line. In reality, a two-dimensional array is prepared, and the intersection shape is expressed by moving forward one square at a time, with the only options being to rotate 90 degrees or move forward one square.

[0015] The shape needs to be represented using only 1-bit direction for forward and rotation. When the previous command is received, move forward one space. When one command is received in succession to the left, it means to turn left. When two commands are received in succession to the left, it means to turn right. When three commands are received in succession to the left, it means that there is a symbol such as the corner of a crosswalk or a traffic light. When four commands are received in succession to the left, it means to turn back. Five or more commands are added when increasing the variety of symbols.

[0016] For example, if the data is "forward," it will move forward one square in the direction you are currently facing. If the data is "left, forward," then move one square to the left from the direction you are currently facing. If the data is "left, left, forward," then move one square to the right from the direction you are currently facing. If the data is "left, left, left, forward," the robot will recognize that there is a symbol, such as the four corners of a crosswalk, where it is currently located, and move forward one square in the direction it is currently facing. If the data is "left, left, left, left, left, forward," it will turn backwards and move forward one square.

[0017] FIG. 1 shows an example of an Excel spreadsheet that expresses the intersection shape using the algorithm of the present invention.

[0018] In Figure 1, the XY coordinates of the starting point are stored at J4 and K4. A flag is also prepared to indicate whether one lap of the closed loop has been completed. When placing a symbol that changes the height direction, a Z coordinate can also be used to express the height direction. However, this is only possible in the Z direction, and is not very efficient.

[0019] Copy the XY coordinates of the previous point to B10 and C10 in Figure 1. This Excel flow is an algorithm that expresses intersection data divided into a matrix as shown in Figure 2 in a low-volume manner. In Figure 2, an empty space such as a line is represented as 0, a line with no symbols as 1, a symbol as 2, and the starting point as 3. Any number of different symbols can be added. For example, with normal representation methods, if the number of representations for each square in the matrix data increases (the number of symbol types doubles), the data capacity also roughly doubles. With this algorithm, the amount of rotation only increases a few times, so the data capacity does not increase in proportion to the number of symbol types.

[0020] In Figure 2, white is the starting point, light gray is the outline with a symbol, and dark gray is the outline without a symbol. The state point does not necessarily have to be on the outline. The ladder road sign Fumi is considered to be the reference point, but by setting a point offset from Fumi as the intersection start symbol, the intersection position can be offset in the XY direction.

[0021] D10 in Figure 1 shows the current direction. It can take four different values: 0 o'clock, 3 o'clock, 6 o'clock, and 9 o'clock. The initial direction can be added to the beginning of the data, or it can be decided that "the initial direction is always 0 o'clock." However, the data capacity changes slightly depending on whether the closed loop is expressed in a counterclockwise or clockwise direction. In the case of Figure 2, if the initial direction is 0 o'clock, it is 734 bits, if it is 3 o'clock, it is 722 bits, if it is 6 o'clock, it is 723 bits, and if it is 9 o'clock, it is 733 bits. The reason is that the number of times to turn right and the number of times to turn left are swapped, and the required data capacity differs depending on whether it is expressed right or left. For example, when expressing a square intersection, if it is expressed in a counterclockwise direction, it is necessary to turn right four times. However, if it is expressed in a clockwise direction, it will turn left four times. The reason for the 1-bit difference between 0 o'clock and 9 o'clock is the difference in whether or not you need to face left immediately after the start. If you face perpendicular to the outline at the start, the 1-bit data capacity will be larger. For this reason, it is best to have the initial direction facing right (3 o'clock direction) when representing an intersection. (When representing left with 1 bit and right with 2 bits)

[0022] C12, C13, C14, B13, and D13 in Figure 1 represent the data for the front, back, left, and right squares taking into account the current facing direction. 3 is the starting point, and 1 indicates that there is a contour. In the case of Figure 1, D10 is 0, so it is facing the 0 o'clock direction of the matrix, and B13 and D13 are 1, so there are contours on the left and right.

[0023] B17 in Figure 1 shows whether there is something to the left. If B13 is 0, it displays NG, if it is non-zero, it displays OK. The formula for B17 is "=IF(B13<>0,"ok","ng")".

[0024] C17 in Figure 1 is the formula for when there is a symbol at the corner of a turn to the left. If there is a symbol, you are forced to move forward one space. If there is a symbol at the corner of a turn, you need to turn backwards. The formula for C17 in Figure 1 is "=IFS(AND(B17="ok",C13=2),"left,left,left,front,left,left,left,left,front,left,left,front,",1=1,"ng")". It means that if the contour continues to the left and there is a symbol where you are now, it will output left,left,left,front,left,left,left,left,front,left,left,front. "Left, left, left, forward, left, left, left, left, forward, left, left, forward," instructs the robot to recognize the presence of a symbol and move forward one space, then turn backwards and move forward one space, then turn right and move forward one space.

[0025] Anything that is neither ok nor ng from B17 in Figure 1 to K17 in Figure 1 is output from B19 in Figure 1 to K19 in Figure 1. The leftmost one among the outputs is the output for that square, and is output while moving one square at a time. The concatenation of these outputs is the final output. The formula for B19 is "=IFS(AND(B17<>"ok",B17<>"ng"),B17,1=1,"")".

[0026] D17 in Figure 1 shows whether there is something before it. If C12 is 0, it displays NG, otherwise it displays OK. The formula is "=IF(C12<>0,"ok","ng")". E17 in Figure 1 is the formula for when there is a symbol on a line. The formula for E17 in Figure 1 is "=IFS(AND(D17="ok",C13=2),"left,left,left,front,",1=1,"ng")".

[0027] F17 in Figure 1 shows whether there is something to the right. If D13 is 0, it displays NG, otherwise it displays OK. The formula for F17 is "=IF(D13<>0,"ok","ng")".

[0028] G17 in Figure 1 is the formula for when there is a symbol at a right-hand corner. The formula for G17 in Figure 1 is "=IFS(AND(F17="ok",C13=2),"left,left,left,front,left,left,left,left,front,left,front,"1=1,"ng")". It means that if the contour continues to the right and there is a symbol where you are now, it will output left,left,left,front,left,left,left,left,front,left,front. "Left, left, left, forward, left, left, left, left, forward, left, forward," instructs the robot to recognize the presence of a symbol and move forward one space, then turn backwards and move forward one space, then turn left and move forward one space.

[0029] In Figure 1, H17 has a contour line continuing to the left, but the current location (C13) is not a symbol. The formula is "=IFS(AND(B17="ok",C17="ng"),"Left,Front,",1=1,"ng")".

[0030] In Figure 1, I17 has a contour line continuing from before, but the current location (C13) is not a symbol. The formula is "=IFS(AND(D17="ok",E17="ng"),"previous,"1=1,"ng")".

[0031] In Figure 1, J17 has a contour line continuing to the right, but the current location (C13) is not a symbol. The formula is "=IFS(AND(F17="ok",G17="ng"),"left,left,front,",1=1,"ng")".

[0032] K17 in Figure 1 is a case where the left, front, and right are all 0, such as at the end of a road, and you need to go back. The formula is "=IFS(AND(C17="ng",G17="ng",E17="ng",H17="ng",J17="ng",D17="ng"),"left,left,left,left,front,",1=1,"ng")".

[0033] Note that there is no way to prevent errors such as starting from an empty space. If you want to add a new one, add it to the right of K17. The output of that square is displayed in B17 to K17 in Figure 1.

[0034] O19 in Figure 1 simply displays the leftmost item from B19 to K19. The formula is "=IFS(B19<>"",B19,C19<>"",C19,D19<>"",D19,E19<>"",E19,F19<>"",F19,G19<>"",G19,H19<>"",H19,I19<>"",I19,J19<>"",J19,K19<>"",K19)".

[0035] P19 in Figure 1 shows the output in different cases and displays numbers to simplify later processing. The formula is "=IFS(O19="left, left, left, front, left, left, left, left, front, left, left, front,",1,O19="left, left, left, front, left, left, left, left, front, left, front,",2,O19="left, left, left, front,",3,O19="left, front,",4,O19="left, left, front,",5,O19="left, left, left, left, front,",6,O19="front,",7,1=1,"The output result is incorrect")". As can be seen from this cell, when there is one type of symbol, the final result is any one of the seven possible outcomes. B22 to D23 in Figure 1 are the coordinates of the next square and the direction it faces.

[0036] D23 in Figure 1 is the next direction. D22 corrects the direction it was facing according to the number in P19. D22 is the pre-processing for D23. The formula for D22 is "=D10+IFS(P19=1,9,P19=2,3,P19=3,0,P19=4,9,P19=5,3,P19=6,6,P19=7,0)". The formula for D23 is "=IFS(D22<12,D22,1=1,D22-12)".

[0037] B22 and C22 in Figure 1 are the next squares to move to. The algorithm of this invention is like taking one square from a two-dimensional array, processing that square, and then processing the next square. D23 in Figure 1 is the direction you are currently facing. The formula for B22 in Figure 1 is "=IFS(D23=0,B10-1,D23=3,B10,D23=6,B10+1,D23=9,B10)". The formula for C22 in Figure 1 is "=IFS(D23=0,C10,D23=3,C10+1,D23=6,C10,D23=9,C10-1)".

[0038] K22 in Figure 1 is a flag indicating whether the robot has returned to its original location. The formula for K22 in Figure 1 is "=IFS(K5=1,1,AND(B22=$J$4,C22=$K$4),1,1=1,0)". When this changes from 0 to 1, it means that it has completed one cycle.

[0039] In Figure 1, if R19 has not gone around once, O19 is output. The final result is the concatenation of all the strings in the row of R. The formula for R19 is "=IFS(K22=0,O19,K22=1,"")".

[0040] This concludes the processing for square 1 in Figure 1. Next, we will explain the processing for square 2. First, copy the range from A8 to S24 down. Next, copy B17, C17, and D17 in Figure 1 to B22, C22, and D23. All that's left to do is copy A25 to S41 down.

[0041] Finally, combine all the outputs from column R. The formula for the results display field is "=CONCAT(R:R)".

[0042] When FIG. 2 is displayed using the algorithm of the present invention, "Left, front, front, front, left, front, left, front, left, left, front, left, left, front, left, left, front, left, left, front, left, front, left, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, front, front, front, front, left, front, left, front, left, front, left, front, left, front, left, front, front, left, front, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left ,left,front,front,front,front,front,front,front,front,front,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,front,left,front,left,front,front,left,front, left, front, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, ,front,front,left,left,front ... left, front, left, front, left, left, front, left, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, ,front,left,front,left,front,left,front,left,front,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front,left,front, left, front, front, front, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front,left, left, front, left, left, front, left, left, front, left, left, front, left, left, front, left, front, left, left, front, left, front, left, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, left, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, left, front, Figure 2 is drawn in a 75x75 grid. Each grid is 25cm. There are also three pedestrian crossings in Figure 2. An intersection that fits within an area of ​​19m in length and width can be expressed with 1000 bits.

[0043] When expressing an intersection, it can be difficult to express an object in the center of a roundabout. The algorithm of the present invention can express one closed loop, so it can also express a roundabout like the one in Figure 3. The roundabout in Figure 3 (with three crosswalks) can be represented in 858 bits. The initial direction is 3 o'clock. Because the initial direction is 3 o'clock and the left side of the initial direction is the inside of the closed loop, 2 bits are added for each side, front, back, left and right, making the actual data capacity 860 bits. When representing a roundabout, it is important to decide which direction to represent. In the case of Figure 3, if turning left is given priority, the inner circle cannot be represented unless it is expressed in a clockwise direction.

[0044] The intersection in Figure 4 can be expressed in 589 bits.

[0045] This algorithm compresses data at the same time as it is represented. Since the data is represented in a Huffman-coded format at the time it is completed, compression is difficult.

[0046] In order to recognize a roundabout without gaps, the closed loop contour of the present invention is within the intersection. If the contour is an intersection up to the center, problems will occur in processing the edges, so the intersection is defined as being within the intersection up to the outer edge of the contour.

[0047] It is also possible to add the position of a traffic light (symbol) to the intersection shape as shown in Figure 7, and correct the actual self-position from the position of the traffic light. Figure 7 can be expressed in 285 bits.

[0048] By placing symbols at the four corners that eliminate the intersection shape in a specific range, it is possible to handle two or more closed loops that are not adjacent. This is merely a possibility, and the expression efficiency is not very good. For example, by connecting the four symbols in Figure 4 and setting the rectangular range outside the intersection, multiple closed loops can be expressed.

[0049] When expressing a single road using this algorithm, there is a problem in that it is not clear where the data ends because the data is variable length. Therefore, it is a good idea to create a symbol to indicate the end of the data. It is also possible to end the data by going backwards twice in a row.

[0050] Although the data volume doubles, by using this algorithm to represent the shape in two ways (clockwise and counterclockwise) in opposite directions of 180 degrees from the starting point, the inner and outer shapes can be correctly conveyed.

[0051] If you want to represent the image in Figure 2, you need to add information about where the starting point is (75+75=150 bits).

[0052] As for decoding, since the Huffman tree is basically always the same, it can be easily converted into two-dimensional shape data. The encoding procedure will now be described. 1. Split the data by the comma before "," and store it in a one-dimensional array 2. Replace the list from top to bottom with "Turn left and move forward," "Turn right and move forward," "Move forward," "Go back," "Recognize symbol 1," "Recognize symbol 2," etc. 3. Expand one-dimensional data to two-dimensional data. (Complete) Impressions

[0053] Current self-driving cars have many problems, such as being unable to drive on snowy roads. We have filed applications with the goal of creating self-driving cars that are safer, more efficient, and more romantic than ever before (running at 300km / h on roads excluding the aforementioned intersections, and 120km / h at intersections). We have filed three applications in order to realize a self-driving car with a maximum speed of 300km / h. The first is a line tracing car (ultraviolet line tracing car) that can trace lines even when snow is piled up. The second is a line for line tracing (ladder road markings) that can hold various information such as speed limits. The third is an algorithm (this algorithm) that expresses the shape of the line branching point (intersection).

[0054] The self-driving cars manufactured by this invention are level 4, but because they can be used on all roads simply by drawing lines and embedding RFID tags, they can be used on all roads at a relatively low cost compared to conventional methods. Therefore, they are essentially level 5 self-driving cars.

[0055] In the comments section, describe how your invention is linked to previous inventions you have filed.

[0056] The theoretical ODD (operational design domain) assumed when creating an autonomous vehicle based on these three inventions is shown below.

[0057] Road condition 1 is a road with a ladder road marking and an intersection connected to it, which is a road with one or more lanes in each direction or a one-way road with one or more lanes. (It does not have to be a one-way road if there is technology that allows passing each other.) Road condition 2 is an intersection where RFID is embedded at the beginning of the ladder road marking. (Road condition 2 is not necessary if you can turn left at the intersection without any information.) Geographical conditions do not consider large skies such as over the sea or lakes (bridges are fine), but roads and building walls and ceilings are considered. Small skies where ladder road signs cannot be drawn, such as between buildings, can be treated as intersections. - Environmental conditions are all weather and time (assuming snow is about 10cm thick, but for vehicles with snow removal functions in front of the vehicle, such as self-driving snow removal vehicles, there is no problem even if the snow is more than 10cm thick) The driving location is from the tag (parking spot) at the start of the ladder road marking to the tag (parking spot) at the start of the ladder road marking. If you want to set a destination somewhere along the way, you can freely set the destination by setting an intersection along the way. -The maximum driving speed is 300km / h on expressways, 300km / h on general roads, 120km / h when going straight through intersections, and slow down when turning left at intersections with pedestrian crossings. 300km / h is the maximum speed at which an appropriate distance can be maintained from the vehicle ahead. 120km / h is the maximum speed at which RFID information can be read. There will be no need for a control room to communicate with all autonomous vehicles. -Can coexist with multicopters and flying cars.

[0058] The technical requirements for creating a normal autonomous land vehicle, excluding flying autonomous vehicles, are shown below.

[0059] - Ability to maintain a distance from the vehicle in front within the speed limit (the speed limit can be read appropriately from the ladder road markings) -An ultraviolet line scan camera is installed on the underside of the self-driving car, allowing it to trace lines painted on the road with ultraviolet reflective paint. The line rate of the ultraviolet line scan camera must be 20 kHz or higher. When driving at 120km / h and passing over an RFID tag embedded in the road, the RFID tag can transmit data of several tens of kilobits to the self-driving car (no data transmission is required from the self-driving car to the RFID tag). - When driving on a one-lane road, if there is a car parked on the side of the road, the car will stop and report the car parked on the side of the road. (If you can avoid parking on the street, you can avoid it.) - Route creation to the destination that only involves turning left, taking into account traffic lanes - The ability to automatically change lanes when it is known that there is a road where you can change lanes several meters away - Ability to automatically turn left at an intersection, knowing when the traffic light will change, the location of the crosswalk, and the shape of the intersection. - Technology to correct the orientation of intersection shape 1 by comparing the feature points of intersection shape 2 derived by this algorithm with the intersection shape 1 of the actual intersection -The car must be able to reach a speed of 300km / h in a straight line. -There are inexpensive transparent paints that transmit UV rays. - Technology that automatically removes snow (either by root or leaving some) while tracing a line ------------------------------------------------------------- ------------------ (Background technology) ------------------ -------------------------------------------------------------

[0060] The intersection with the most branches in Japan is Sugawarabashi intersection, which has 11 forks. Therefore, we consider the road to have 11 branches, including the road where you make a U-turn. The longest road in the world has eight lanes. The largest intersection in Japan is Hakotsutsumi Intersection, with 7 lanes on each side for a total of 10 lanes. The fork in the road at the parking lot has the most branches. In the present invention, the path is assumed to branch into a maximum of 11 branches. The road has a lane width of about 2.5m, and the size changes in increments of 25cm.

[0061] There are documents showing that RFID tags can be read from moving vehicles at speeds of approximately 40km / h for passive and 120km / h for active, but an RFID tag embedded in the road surface cannot generally be read by an autonomous vehicle traveling at 300km / h.

[0062] As long as the paint contains an ultraviolet scattering agent, the color of the paint does not matter, so the code painted on the road signs can be anything that can be read by a line scan camera.

[0063] ------------------------------------------------------------- ------ (Challenges in creating self-driving cars) ------ ------------------------------------------------------------- This document focuses on the following points and describes the problems and solutions when creating an autonomous vehicle. The area separated by the wavy line is one block. ---------------------------------------------------------------------------- How to make autonomous vehicles drive around sharp curves on highways almost 100% safely -Specific usage of ladder road markings -Areas in which rules need to be revised Snow countermeasures - Device for preventing dirt from entering the sensor under the vehicle · Measures against ice burns -About the color of the road markings on the ladder About the width of the ladder road markings Where will the money come from to maintain ladder road markings? ·About the branching of ladder road markings About the destination About the construction - Regarding RFID installed at the start of the ladder road marking About the "Fumisan" on the ladder road markings - Direction correction of RFID showing intersection shape -About lane changes About passing each other -About manholes on roads About grating Signal Recognition About the highway entrance How to tell if your car is self-driving Driving from the highway to a nearby parking area About Multicopters -Example of the general flow of the demonstration experiment Summary ----------------------------------------------------------------------------

[0064] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ How to make autonomous vehicles drive around sharp curves on highways almost 100% safely A line will be painted in the center of the road and the self-driving car will follow that line. In this case, the line needs to be designed so that it can coexist with existing manually driven cars. The solution is to "put a ladder made of paint containing UV reflectors in the center of the road," which allows the automatic driver to trace the line without interfering with the driving of the manually driven vehicle. If the paint is colorless, it does not need to be ladder-shaped. By giving the ladder-shaped ladder information such as the speed limit, it can automatically reduce the speed before a curve.

[0065] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Specific usage of ladder road markings --About reading-- The ladder is read by a line camera attached to the underside of the vehicle. Conventional self-driving cars use normal cameras or lidars, but the method proposed in this document uses a line camera to read the ladder, so the control cycle is very short and the vehicle is capable of reading the ladder even when traveling at a speed of 300 km / h. Currently, there is technology that allows following vehicles to travel at 300km / h, but there is a risk of losing sight of the vehicle ahead on curves. Therefore, the vehicle itself should follow the line, and only care should be taken to maintain the distance between vehicles outside of curves, and by slowing down in advance on curves, the current technology theoretically allows for safe line tracing at 300km / h. --About the implementation-- A ladder road marking is written on the road, and then painted over with a transparent paint that does not reflect ultraviolet light. Then, before it dries, sand is spread over the road and the spaces between the vertical bars are painted over with a transparent paint. This makes it possible to pass over the ladder on a motorcycle without any problems. --About ladder maintenance-- There will be no problem if you paint a transparent paint that does not reflect UV rays over it before it peels off. --About the weather other than snow-- The vehicle is equipped with a sensor on the underside, so it will not lose track of the line even at sunset or in the rain. Even if you cannot see anything ahead due to a sandstorm, you can still drive within the range where you can follow the line.

[0066] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Areas in which rules need to be revised The maximum speed on national expressways will be set at 300km / h. If the ladder road marking is not colorless, the size of the diamond mark needs to be changed.

[0067] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Snow countermeasures We will operate self-driving snowplows to remove a certain amount of snow.

[0068] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Device for preventing dirt from entering the sensor under the vehicle There is an invention that solves this problem by attaching a rotating glass, etc. A rotating glass is attached around the sensor, and the snow that has adhered to the glass is removed with a brush.

[0069] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ · Measures against ice burns Drive at a slower speed if the outside temperature is below 4 degrees.

[0070] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -About the color of the road markings on the ladder If it is transparent, it will not interfere with diamond marks, etc., so colorless is preferable. In parking lots, it may be useful to have visible lines, preferably white, to indicate areas where self-driving cars may pass.

[0071] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About the width of the ladder road markings I think 1m is fine. However, I wish the width of the ladder and the distance between the steps were standardized all over the world. The tuktuk is considered to be the thinnest vehicle. If the tuktuk is given small wings of several tens of centimeters to block sunlight, it will be able to read lines up to about one meter wide. Considering that the basic width of about 1.3m is required for light vehicles to read the markings, I think that a width of about 1m is appropriate for ladder road markings. Specifically, the width of each vertical line of the ladder should be 10cm to 15cm, the width of the step should be 25cm, and the width of one color of step should be 12.5cm. I think that the width of the ladder including the vertical lines should be 1m (the width between the vertical lines is 70cm to 80cm). There is no problem with the spacing between the steps always being constant. The width of the ladder road markings can be 1m00cm because road markings are in 25cm increments. I think the width dimensional tolerance should be 1m plus 0cm minus 5cm between 90cm and 100cm.

[0072] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Where will the money come from to maintain ladder road markings? The cost is recovered through advertising costs when the ladder is driven over. Information such as the company's corporate number is stored in an RFID tag placed at the start of the ladder road marking. The longer the ladder road marking, the higher the maintenance costs and the longer the advertising time is displayed. The problem is that while it may not be a problem in urban areas, it may be difficult to raise advertising funds in rural areas. Will the only option in rural areas be to have drivers drive normally?

[0073] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ·About the branching of ladder road markings There are no forks in the ladder. All forks in the ladder road markings are intersections. The intersection also has curves so sharp that they cause problems when tracing the line.

[0074] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About the destination The RFID tag attached to the start of the ladder road markings determines whether the vehicle has reached its destination. Therefore, the system automatically drives from the start of the ladder road markings to the start of the ladder road markings. It seems possible to realize autonomous driving up to the main road and the parking spot (intersection) of the parking lot adjacent to the main road, but it is doubtful whether it can be realized in narrow alleys. The realistic compromise is to have the system manually drive to the main road and then drive autonomously to the main road near the destination. If you want to install a parking spot at a house where you can reach by self-driving, you can do so for a fee. In that case, it would be better to divide the parking spots into four categories: parking spots (intersections) that anyone can go to on the car navigation software, parking spots that only certain cars can reach, parking spots that anyone can park in, and parking spots that only certain people can park in. For example, if a user enters a password, which can be changed at will, into the navigation system, it will be able to guide the user to the parking lot of a private home, and the user can then have a friend enter the password to bring the car to the home. Basically, the starting point of the ladder road marking is a parking lot. If one point is a parking lot, only one car can be parked there. Also, you must draw a line to the middle of all parking spots. Representing a parking lot with a symbol in an intersection reduces safety, but it allows you to represent a parking lot with multiple cars at a dead end.

[0075] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About the construction The construction work can be divided into two parts: on the intersection and on the ladder road markings. When construction is taking place on a ladder road marking, if there are two or more lanes on one side, it is sufficient to change lanes, but if there is only one lane, it becomes necessary to drive in the oncoming lane. With current technology, driving in the oncoming lane is extremely dangerous, so in the first place, vehicles should not enter areas with only one lane on each side under construction. Therefore, when construction is taking place on a ladder road marking, the contents of the RFID are rewritten to change lanes or restrict turning in a specific direction. If the vehicle enters a road under construction due to an error in rewriting the RFID, it is necessary to switch to manual driving or to follow the vehicle in front at a low speed. If the vehicle enters a construction site, a Fumi-san-shaped sheet indicating that construction is taking place can be placed on the road to safely stop it. When construction work is being carried out on an intersection, a new impassable area can be created based on the shape of the intersection expressed in the present invention. Lanes are restricted by writing information about whether or not construction is occurring on the tag at the beginning of the ladder road marking. To make construction even safer, the construction sign should be a photo of the back of a car, with the words "Under Construction" superimposed so that they do not cover the car. Even if the words "Under Construction" are not visible, a function will be used to keep a safe distance from the vehicle ahead, at the very least causing the car to stop. Even better, it would be better if the brake lights in the photo of the car were red. From the program's perspective, if the words "Under Construction", a mark, or a two-dimensional code are displayed above a stopped vehicle, it can be determined that construction is underway. In cases where a single detour can make a big difference in distance, such as mountain passes, it would be better to read the construction Fumisan and inform the passengers by voice or some other means if the route is changed. For example, a message like "The destination has become 100 km farther away due to construction work. Would you like to manually drive through the construction site?" If you ignore it, you can just drive 100 km extra, but from an efficiency perspective, it would be better to manually drive only through the construction site. If you want to make one-lane road construction fully automated, even though it will cost money, you can just install a movable traffic light. The intersection should be set up on the road with oncoming traffic, and a traffic light that alternates between green and red should be installed at the entrance.

[0076] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Regarding RFID installed at the start of the ladder road marking The information in the list below enclosed in ">>" will be stored in Fumi's RFID tag at the start of the ladder road sign. To represent a road without ladder road markings, the vehicle height limit can be set to 0m. In this invention, there are three things that we want to express at an intersection: the shape of the road that an autonomous vehicle may enter at maximum speed, the space (sidewalk) where people or vehicles may enter, such as a pedestrian crossing, and the start of a ladder road marking. There is no problem with the median strip, etc., as they should not be in the space that an autonomous vehicle may enter. Road markings on ladders are always one-way. An example of where to place an RFID tag (where the branch is at the branch of the line) is shown in Figure 5. 001 in Fig. 5 is a "ladder road marking." 002 in Fig. 5 stores the "contents of the RFID tag to be placed at the start of the intersection," which will be described later. >>>>>>>>>>>>>>>>>> Information to be stored in the RFID tag at the start of the ladder road marking >>>>>>>>>>>>>>>>>> --------------- -The first bit is set to 1 to indicate that it is an early version (1 bit) 1bitx1=1bit --------------- - Is there a communication-enabled traffic light at this intersection? (Normal traffic lights that do not communicate, traffic lights that communicate, or no traffic lights) (2 bits) · Whether the entire intersection is considered a crosswalk (whether it is a scramble intersection or not) (1 bit) -Speed ​​limit when going straight through an intersection (0km / h to 310km / h in 10km / h increments) (5bit) -Speed ​​limit when turning at an intersection (0km / h to 310km / h in 10km / h increments) (5bit) --------------- - Traffic light type (whether it is a normal traffic light with a set green / red time or a push button type where the green / red time is not fixed) (1 bit) -When communication-enabled traffic lights are installed, five (two types) pieces of information for each of the 11 forks (current green or red (1 bit), offset time until arrival at the intersection (approximately 15 bits), time information until the next green or red (yellow is considered red) (9 bits), time information until the next green or red (9 bits), time information until the next green or red (9 bits)) (minimum total of 37 bits x 11 forks x 8 lanes = 3256 bits) If you enter an intersection after this time has passed, slow down as you approach the intersection and proceed by reading the traffic lights in the conventional way. The offset time until arrival at the intersection is set to a time slightly smaller than the time from reading the RFID tag at the start of the ladder road marking to arriving at the intersection while traveling at the fastest speed. The more time information there is until the next green or red light, the smoother driving will be. These two items can be omitted if there are no traffic lights or if the traffic lights can be perfectly read using image processing alone. --------------- Number of lanes (7bit) This is used to recognize how many pieces of data there are in the algorithm described below. The road with RFID A in Figure 5 has 6 lanes. The reason why it is called 6 lanes is because a normal road has 3 lanes, so there are 3 lanes to begin with. Since it branches off to the left along the way, it becomes the 4th lane if you assume that there is a road from RFID A. If you include the road after the merging, it becomes the 5th lane. If you include the lane for turning right, it becomes the 6th lane. The road with RFID A in Figure 5 has a branch, and a ladder road marking sticks out on the left, but it does not actually have to stick out (0 cm sticking out is fine). If it sticks out, paint will be applied to the area where the car passes, and the paint will disappear. For example, the 14-way intersection on the right side of Figure 5, which says "I want a traffic light," can be a 14-lane road instead of an intersection, because there is no possibility of other cars coming. A fork where no other cars will come can be either a lane or an intersection. --------------- · Algorithm mode (10bit) For example, if three rotations on the spot are represented by symbol A and four rotations are represented by symbol B, then by switching them, the data represented by this algorithm is reduced.Also, by placing a symbol that doubles the forward movement, the data capacity of the algorithm representing intersections and roads excluding intersections is reduced. --------------- -Road shape for each lane represented by this algorithm, excluding intersections (Total of about 10,000 bits x number of lanes: 8 = about 80,000 bits) Assuming the length of the road from the start of the ladder road sign to its end at Fumi-san (the intersection) is about 1 km in a straight line, this would amount to about 10,000 bits of information. In theory, autonomous driving is possible with just the road shape data excluding intersections or the ladder road markings, but to deal with unforeseen problems such as losing sight of the ladder road markings or the ladder road markings disappearing, the system recognizes road shapes in two different ways. In addition to these two, it also uses recognition by a conventional neural network, making it a triple road shape detection system. The RFID tag at the start of the road markings on the center ladder of a three-lane road stores data that resembles the gray lines in Figure 6. The gray circle indicates that you cannot change lanes into the adjacent lane. Signs and road markings can be represented by symbols. By placing symbols that double the forward movement, straight lines can be represented with less data. Approximately 80% of the total data stored in Fumi's RFID at the start of the ladder road sign is data for this item, 10% is data for the shape of the next intersection, and the remaining 10% is other data. An example of the data termination code would be two consecutive wraparounds. --------------- -Intersection shape represented by this algorithm (Total approx. 1500 x number of lanes = approx. 12000 bits) Stores information on when you enter an intersection from the end of the ladder road markings painted on each lane of the current road. An example of the data termination code would be two consecutive wraparounds. --------------- Intersection symbol number (4 bits) The branch destination to be expressed from now on will be the symbol number of the intersection expressed by this algorithm. - Height limit for vehicles when traveling on that road (0m to 4.5m in 10cm increments) (6 bits) -Width limit for vehicles when traveling on that road (0m to 6.3m in 10cm increments) (6 bits) Weight limit for the road (0.5t to 10t in 0.5t increments, 10t to 1t increments) (6 bits) -Type of lines on the road (at least four types: 2-bit, 1m-wide ladder road markings, thin lines for multicopters, underwater lines (rope roads), and magnetic tape roads) -Is the road passable for small multicopters such as those used for delivery (1 bit) It is best for RFID to have a reading distance of around 5m so that it can be read from multicopters in the air. -Is the road passable for large multicopters such as flying cars? (1 bit) - Is there a possibility of a car coming from the front on that road? (1 bit) Even if no car is coming, there is a possibility of a car parking or suddenly appearing, so check ahead in a conventional and reasonably safe way. If Fumi-san has a direction, this item is not necessary. - Angle of approach to the road (6bit) From what angle should you approach from an intersection to the RFID at the beginning of another ladder road marking? (The above data is required for each branch in the road, so if the road branches into 11, the total is 33 x 11 = 363 bits.) --------------- Symbol information for signs and road markings (10 bits) Only the position of the sign is expressed by the symbol of this algorithm, and the type of sign is recorded here. (The number of the symbol corresponds to the sign information.) Sign and road marking types (30 bits) (number of bits for all sign and road marking types) Signs with a maximum speed of 60km / h and a maximum speed of 80km / h are treated as different types of signs. A signboard initiating short-range communication may be represented here. --------------- -Can I park my car at the intersection? (1 bit) Intersection number (intersection GPS and elevation information or intersection number) (63 bits) The intersection number is used to efficiently reach a parking lot in an underground parking lot where GPS cannot reach. For example, when a symbol that turns on short-range communication is read, the system communicates with the parking lot machine and transmits information on the connection between intersection numbers to the self-driving car. 64bitx1=64bit --------------- - Corporate advertising data such as corporate numbers (64bit) In order to raise funds for the maintenance of road markings for ladders, information for displaying advertisements inside self-driving cars is stored. Specifically, two pieces of information are stored: a 12-digit corporate number and a number indicating the number of the advertisement. Online advertisements are retrieved from these two pieces of information and displayed inside the car. --------------- Parity bit A parity bit may be provided for each division. --------------- >>>>>>>>>>>>>>>>>> Contents of information instructing construction >>>>>>>>>>>>>>>>>> --------------- Where is the construction site? (2bit) There are four types: construction work on the road excluding the intersection, construction work at the intersection, construction work on the road around the turn excluding the intersection, and no construction work. --------------- -Is the construction site on an intersection or on a ladder road marking (1 bit) - Lanes without construction (1st lane from the left is the 1st bit, 2nd lane is the 2nd bit) (8 bits) -Speed ​​limit from 200m before the construction site (0km~300km in 10km increments) (5bit) How many meters away the first construction work begins (does not start at 0m) (20 bits) -How many meters the first construction work is set to (20 bits) How many meters away the second construction work begins (does not start at 0m) (20 bits) -How many meters between the second construction works are set (20 bits) The construction area at the intersection is expressed by the algorithm of the present invention. (Approximately 1500 bits) - At which intersection is the intersection where construction work is taking place (the first is the 1st bit, the second is the 2nd bit) (11 bits) It would be a pain to rewrite the RFID every time for construction work. It would be better to give up construction work and operate it manually. Or it would be better to prepare a construction-use Fumi-san equipped with RFID. --------------- The above is the data to be stored at the start of the ladder road marking. ---------------

[0077] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About the "Fumisan" on the ladder road markings Prepare at least three types of Fumi-san as follows. -Normal Fumi-san showing the direction of travel on the road -Fumi-san at the start of the RFID-embedded ladder road marking -Fumi-san at the end of the ladder road marking, which is the intersection reference point (-Fumi-san representing the speed limit) (- Fumi asks if they should stop the train and switch to manual operation because there is construction going on) (Fumi-san, who decides whether or not to enable short-distance communication with machines installed in the parking lot) Information such as construction information, short-distance communication, and speed limits can be stored in the RFID tag at the start of the ladder road marking. In simple terms, all the steps except for the steps at the start and end of the ladder road marking are the same, and all the steps can be spaced at the same intervals.

[0078] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - Direction correction of RFID showing intersection shape The intersection shape is stored in the RFID tag at the start of the ladder road marking, and the reference for the intersection shape is written in the Fumi-san at the end of the ladder road marking. An algorithm is required to grasp the intersection shape and estimate the self-position. Images from a camera mounted on the vehicle can be used to predict the range of roads and crosswalks using a neural network, and the angle from the reference can be corrected by comparing it with the map expressed by this invention. It is also possible to include the position information of traffic lights in the intersection shape and correct the intersection angle from the position of the traffic lights.

[0079] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -About lane changes Changing lanes is dangerous, so try to avoid it as much as possible. Basically, you should change lanes when turning at an intersection. You will need to change lanes at three locations: when entering from the expressway interchange, at the fork in the expressway, and at construction sites.

[0080] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About passing each other The road markings on the ladder are one-way, so there is no chance of passing each other.

[0081] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -About manholes on roads Although the reflectance of ultraviolet light is different from that of roads, it is different from the road markings on ladders, so this is not a problem.

[0082] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About grating The ladder road markings may be interrupted by gratings, etc. To deal with this, we will install line cameras on three lines: the front, rear, and center of the vehicle. If the front sensor loses sight of the ladder road markings due to gratings, we believe there will be no problem if we trace the line using the line cameras in the middle and rear. Additionally, the approximate shape of the road can be read from RFID.

[0083] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Signal Recognition There is no need to transmit traffic light color information. Only information about whether or not it is possible to proceed from each lane is transmitted from RFID to the self-driving vehicle. The advantage of not transmitting color information is that it can also handle situations where there are roads diagonally to the left and to the right. In addition, the data capacity is reduced. The disadvantages are that the data capacity becomes large for roads with many branching lanes and intersections, and that traffic light information becomes unclear over time.

[0084] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About the highway entrance One existing invention is a method to install traffic lights at the entrances to highways. If the entrances to highways are equipped with vehicle-detecting traffic lights, self-driving cars can enter the highways as safely as possible. When you are on the highway and heading to your destination, if you have to go to the right at an interchange, you will need to change lanes. If you can reach your destination by simply staying in the left lane on the highway, there will be no lane changes between the parking area and the parking area. Even if you need to change lanes, you only need to change lanes once along the long distance before you reach the interchange. By arranging the entrances and exits to the expressway as shown in Fig. 8, one can enter a facility with a parking spot simply by turning left from the expressway exit, and can get on the expressway simply by leaving the facility and turning left again.

[0085] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ How to tell if your car is self-driving As proof of self-driving cars, a method of attaching an animal's tail to the vehicle registration plate has been devised. If it were to move with a motor or move according to the state of the car interior, it would infringe on other patents, so I think it would be better to grow a simple, immovable tail on the vehicle registration plate.

[0086] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Driving from the highway to a nearby parking area The location of the parking area itself can be anywhere, but the placement of the highway entrances and exits is important. If the four entrances are lined up nicely, you can enter the parking area by simply turning left, and when you leave the parking area, you can get back on the highway by simply turning left again. If the highway exit is positioned correctly, you can reach the parking area by simply turning left from the leftmost lane without changing lanes. The closer the parking area is to the interchange, the fewer intersections will require footbridges.

[0087] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ About Multicopters It's okay if it crashes on the road, so if you can only trace a line on the road, you can fly it relatively safely. If it falls on a car, the car will protect it. However, since the line needs to be recognized while the car is on the ladder road marking, there is a possibility of some malfunction. If the RFID is compatible up to the height of the multicopter, the multicopter can read the road shape, intersection shape, and the position of the next RFID, so it may be possible to support multicopters without any special additional equipment.

[0088] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Example of the general flow of the demonstration experiment I recommend experimenting in the following order: --------------- 1. Ladder road markings will be placed on the roads of the experimental city, and the vehicle will be allowed to drive autonomously while checking how far it can trace the lines on snowy days. 2. We will recruit companies that want to invest in ladder road markings, install ladder road markings on expressways across the country, and have trucks drive automatically between parking areas near the expressways. 3. Have regular vehicles drive autonomously between parking areas near highways. 4. Expand the range of autonomous driving to include major roads only. 5. Use a multicopter to automatically navigate between parking areas near the highway. (This can be applied to flying cars.) ---------------

[0089] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Summary If we paint a single-color transparent ladder road marking on the road, embed an RFID tag at the beginning, and give Fumi-san information such as the speed limit (this can also be stored in the RFID tag at the beginning. If stored, this algorithm will provide a symbol for speed changes), and add an ultraviolet line scan camera and a camera to the underside of a conventional self-driving car, we should be able to realize a self-driving car with a top speed of 300 km / h, but only on long straight roads. Since autonomous vehicles only connect expressways and areas immediately after exiting interchanges, they do not encounter people, making it clear who is responsible for any accidents. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ [Explanation of symbols]

[0090] (Explanation 1) none

Claims

1. An algorithm that represents a line or a closed loop from a point, The algorithm is: The closed loop such as an intersection is a contour line, and the single line with thickness such as a road is a center line. From one point, it is expressed with continuous 1-bit information of rotation and forward movement, and is characterized by expressing up to the end, the first point, or a specific symbol. The characteristic of this is that the amount of rotation is not constant when rotating continuously. The feature of expressing a symbol on the line or the closed loop is that the symbol returns to its original direction by continuously rotating; An algorithm for expressing a line or a closed loop from a single point.

2. 2. The algorithm for expressing a line or a closed loop as viewed from one point according to claim 1, wherein the amount of rotation is 90 degrees for the first rotation and 180 degrees for the second rotation.

3. The algorithm for expressing a line or a closed loop as viewed from one point according to claim 2, characterized in that the amount of advancement in one step is 25 cm.

Images