CN112562373A

CN112562373A - Method for automobile automatic driving lane level positioning and roadside traffic identification and command signal identification

Info

Publication number: CN112562373A
Application number: CN202011380468.XA
Authority: CN
Inventors: 郭荣江; 陈武玲
Original assignee: Individual
Current assignee: Individual
Priority date: 2020-08-28
Filing date: 2020-11-30
Publication date: 2021-03-26
Also published as: WO2022041942A1

Abstract

The invention provides a simple and practical method for automobile automatic driving lane level positioning and roadside traffic identification and command signal identification. The method comprises the steps that a reflector group is arranged on the road surface of a lane, a control unit, a transmitter and a receiver are arranged on a vehicle, a transmitted wave transmitted by the transmitter is reflected by the reflector group, the receiver receives a reflected wave formed by reflecting the transmitted wave by the reflector group, the control unit determines a coded command represented by the reflected wave according to the reflected wave and further determines the command content represented by the coded command, and the vehicle is controlled according to the determined command content, wherein the command content represented by the coded command of the reflector group is roadside static marks such as vehicle positions, lane attributes and the like. The dynamic identification information such as traffic lights and the like is transmitted to the vehicle by the roadside active transmitting device. Depending on the circumstances, a portion of the static identification may also be transmitted to the vehicle by the roadside active transmission device.

Description

Method for automobile automatic driving lane level positioning and roadside traffic identification and command signal identification

Technical Field

The invention relates to a road device and an automobile automatic driving method for directly providing lane-level positioning, roadside traffic identification and command signal electronic instructions for vehicles.

Background

The flying blanket can freely go to the wanted place, which is the myth of human beings for many centuries. Technology development to date, achieving this with automotive autodrive, is a recent goal that the artificial intelligence and automotive industry have considered possible. If the technology is successfully realized, except that few people hope to have private car, most people only need to attract an automatic taxi anytime and anywhere like using Uber or dropping to get a car, and can safely arrive at the taxi by inputting the destination. Private cars are not needed, garages in homes are not needed, and parking lots on the streets of shopping malls are not needed. Passengers can take the vehicle as a place for continuous work, social intercourse, entertainment or rest during the journey. The authors believe this is the most recent future artificial intelligence that may bring the greatest change to human life.

This is why, taking china as an example, capital has entered this area in large quantities starting at 2014. However, the good scene is not long, and after about 3 years, from the next half year of 2017, the Chinese automobile autopilot ring begins to feel the gust of capital, the enterprise financing difficulty is increased, and the failed enterprises are increased rapidly^[1]. By 2019, the industry has more a statement of "automatically driving into the cold winter capital".

Disclosure of Invention

For simple reasons, the potential and progress of a commercial landing of capital interest has not met the requirements^[2]。

Technologically, engineers have introduced into this area the leading technologies of laser, radar, communications and artificial intelligence, such as LIDAR, millimeter wave radar, C-V2X, 5G-V2X, deep learning, CNN, … …, and so on. But why does not solve the problem of commercialization? The authors believe that the technology went wrong with the route.

Take an example for illustration. It is assumed that the goal of the project is to achieve automation of automobile manufacturing. There are 2 routes technically: in the first route, a universal robot capable of walking autonomously is researched and developed, tools such as a welding gun trolley crane and the like used by human beings are identified and used, screw nuts and various automobile parts are identified, and the robot works under the command of software simulating the step of assembling an automobile by workers by utilizing equipment of a manual automobile assembly workshop existing decades ago, so that the automobile is accurately installed by replacing the human beings; the second route is to reform the manual automobile assembly workshop existing decades ago into the automatic production line of most modern automobile factories, various manipulators which are simpler than universal robots are arranged on two sides of the production line, each manipulator completes a simple operation, and finally, the automatic installation of the whole automobile is completed at the terminal of the production line. Clearly, the second route is correct and results in success.

At present, the automobile driving automation is walkingSimilar first route. Why is it said? Please see the method for identifying the traffic lights of the intersection in the current industry: the most widely used are various artificial intelligence-based visual recognition algorithms^[3]. Since recognition using only color and shape information is not effective, researchers have proposed methods such as a "cascade classifier" method, a "feature classification" method, a Convolutional Neural Network (CNN) method, and the like. In the case that a plurality of traffic lights are simultaneously present at the same intersection, it is difficult for the vehicle to determine which traffic light is suitable for the vehicle position in the current lane, so that the information of the position, size and the like of the traffic lights can be stored in the high-precision map only by using the high-precision map. In the running process of the vehicle, the coordinates of the traffic light suitable for the current position are found out from the high-precision map according to the position of the vehicle, and then the coordinates in the map are mapped to the shot image of the visual field by using coordinate conversion. And because the traffic lights only occupy a small area in the whole visual field, a region of interest (ROI) must be cut out from the image, and finally, the ROI is identified by using a target detection network. Even through the complicated process, the highest recognition accuracy reported by various methods can only reach about 99%. What is this concept? That is, an autonomous vehicle may misjudge the traffic light 1 time every 100 times the intersection! The situation is similar for the field of positioning technology. At present, the lane-level positioning technology using artificial intelligent recognition can only achieve about 90% of accuracy^[4]. In addition, centimeter-level positioning position can be obtained by a global satellite positioning system and various enhancing methods, and lane-level positioning accuracy can be achieved by using a high-accuracy map, but the cost is high in economy and technology.

In view of the above, the present invention provides a simple and practical road device and method for automatically driving a vehicle, which can be used for lane-level positioning, roadside traffic identification and electronic command of command signals.

In order to achieve the purpose, the method for automatically positioning the driving lane of the automobile and identifying the roadside traffic signs and command signals comprises the steps of arranging a specially arranged device on the roadside or the road surface of the lane, and directly providing lane-level positioning, roadside traffic signs and command signal electronic instructions required by automatic driving for the automobile by the device. The device specially arranged can be a passive reflector group (explained in detail below), a control unit, a transmitter and a receiver are arranged on a vehicle, a transmitted wave transmitted by the transmitter is reflected by the reflector group, the receiver receives a reflected wave formed by reflecting the transmitted wave by the reflector group, the control unit determines a coded command represented by the reflected wave according to the reflected wave and further determines the command content represented by the coded command, and controls the vehicle according to the determined command content, wherein the command content represented by the coded command of the reflector group is information of a lane line, the position and lane attribute of the vehicle and a roadside traffic mark and a command signal of the road section. In addition, the specially arranged devices can also be active emitting device groups. At the moment, a control unit and a receiver are arranged on the vehicle, the receiver receives the waves transmitted by the transmitting device group to the receiver, the control unit determines the coding instruction represented by the transmitting waves according to the transmitting waves and further determines the instruction content represented by the coding instruction, and the vehicle is controlled according to the determined instruction content, wherein the instruction content represented by the coding instruction of the transmitting device group is the information of the position and lane attribute of the vehicle and the road roadside traffic identification and command signal of the road section.

By adopting the method, the information of the position and the lane attribute of the vehicle and the roadside traffic mark and the command signal of the road section can be provided very effectively and accurately only by arranging the corresponding passive reflector group or the active transmitter on the road. Therefore, the method of the present invention is simple and practical.

Preferably, the lane attribute information includes travel rule information.

Preferably, the lane attribute information includes position information of the own lane in the road.

Preferably, the lane attribute information includes: a) the position of the lane in the road, namely which lane, is convenient for selecting the corresponding traffic light signal in a plurality of traffic light signals at the intersection in a command data summary table (viewing [0056]) received by the vehicle immediately according to the information; b) the left and right turns or straight driving rule allowed by the lane; or c) if the lane does not allow the next operation of driving, it is necessary to cross several lanes to the left or right to reach the target lane.

In the present invention, it is preferable that the control unit determines specific contents corresponding to all the coded commands required for the automatic driving based on a table stored in advance. For example, code 0 represents a lane line marker, and the mirror position represents its straight-line distance and declination angle from the vehicle; code 0002 represents a 75 km speed limit indicator; code 0005 represents left turn lane identification, etc.

Preferably, the control unit determines static-dynamic traffic identification information represented by the coding instruction according to a pre-stored chart; the static-dynamic traffic identification information is determined according to an instruction data summary table (detailed in [0056]) of static-dynamic traffic identification sent by the roadside immediately received by the vehicle.

In the present invention, the emission wave is preferably a physical wave such as an ultrasonic wave, an electromagnetic wave, an optical wave, or a laser.

In the present invention, it is preferable that the mirror group includes 1 or more mirrors, and reflection characteristics of the reflected wave by the mirrors in the plurality of mirror groups are different, so that the coded command signal can be formed by different arrangements of the mirrors.

In the present invention, it is preferable that the reflected waves formed by the reflecting mirror have different frequencies, waveforms, peaks, widths, or waveform integrals according to the difference in size, material, shape, surface finish, or flatness of the reflecting mirror, so that different reflected waves are obtained.

In addition, the invention also provides a road system, a plurality of reflector groups are arranged beside a lane or on the road surface, the reflector groups comprise 1 or a plurality of reflectors, and the reflector groups directly provide lane line information, lane level positioning, roadside traffic identification and command signal electronic instructions required by automatic driving for vehicles.

In the present invention, it is preferable that the reflection characteristics of the reflected wave by the mirrors in the plurality of mirror groups are different from each other, so that the coded signal can be formed by the arrangement of the mirrors being different from each other.

According to the invention, a concave hole is preferably arranged on the road surface, and the reflector group is arranged in the concave hole.

In the present invention, it is preferable that a transmitting device is provided in a roadside or a road surface, the transmitting device stores a data summary table indicating information of static traffic signs and dynamic traffic signs of a road section provided in the transmitting device, and the transmitting device transmits the data summary table to a vehicle traveling on the road section at a constant frequency.

In the present invention, the transmitting device is preferably installed at a place where a static traffic sign changes or a predetermined distance from an intersection exists.

In the present invention, a reflector indicating a parking mark is preferably provided near an entrance of a building.

Preferably, the invention is provided with a reflector for indicating the parking position of the taxi on the road.

Drawings

Fig. 1 is a schematic diagram showing an example of a road scene.

Fig. 2 is a schematic diagram for illustrating an automatic driving method of a vehicle and a road system structure according to an embodiment.

Fig. 3 is a schematic diagram for illustrating an automatic driving method, a lane-level locating method and a road system of a vehicle according to an embodiment of the present invention.

Detailed Description

Traffic control equipment for roads in the world is now designed for human driving, just as automobile assembly plants decades ago were designed for manual assembly by workers. In view of the above-described current methods of identifying traffic lights and other traffic signs that are automatically driven, it is like a universal robot for the first route of automatic automobile production.

If the automatic driving needs to go the second route, the existing road is completely reformed to be suitable for automatic driving, just like the automatic production line of modern automobiles is completely different from the old manual automobile assembly workshop.

What would be a road device suitable for automatic driving? One general principle is that the road should speak directly to the vehicle in the language understood by the vehicle, rather than using artificial intelligence to translate the human-understood language into the vehicle-understood language. That is: the method is characterized in that a code instruction which can be clearly and accurately recognized is transmitted to a vehicle in a hardware, software or hardware and software mixed mode, and the method for recognizing marks and signals from the existing shooting data of roadside traffic marks such as characters, figures or colors for human recognition through artificial intelligent image recognition is replaced. The most straightforward approach is naturally to use a code instruction generator to send all the codes of the traffic signs and command signals that the human driver needs to recognize to the incoming vehicle. But this is difficult to achieve. As a simplest example, it is not possible to install an active code instruction generator on the lane line marking of each small segment, since it is an astronomical number. Therefore, we will discuss the information in terms of both static information and dynamic information.

The static information refers to traffic signs marked on the road surface and the roadside without changing with time, and is called as static traffic signs. Such as lane lines, straight or turning arrows, speed limit and yield stop signs, etc. For this type of identification, as many passive echo reflectors (hereinafter collectively referred to as mirrors) are used as possible to send commands to the incoming vehicles. This reflector should be a small mirror made of metal or other material, mounted near the lane (preferably at the midline) where the incoming vehicle is traveling, and the reflected echo is generated upon illumination by the transmitted waves (ultrasonic, electromagnetic, light, laser, or other) from a transmitter mounted at the head, belly, or tail (or other location) of the vehicle. This transmitted wave has only one fixed type (type, frequency, intensity), but due to the difference in size, material, shape or surface finish and flatness of the mirror, the radiation reflection characteristics (frequency, waveform, peak, width, waveform integration, etc.) are different, and thus different echoes (reflected waves) are obtained. Theoretically, we can obtain 2+2 by using N such mirror arrangements as long as we can obtain 2 different echoes²+…+2^NIs coded to obtain 2+2²+…+2^NAn identified instruction. In practice, as an example, if we can getTo 5 different echoes, with 1,2 and 3 such mirrors, respectively, aligned, we can obtain: 0.5 in total, namely 1,2,3 and 4; 00. 01, … 04, 10, …, 44, i.e. 5²25 in total; 000. 001, …, 444, i.e. 5³125 codes, for a total of 5+25+125 155 identified instructions. Under most road conditions, the road surface cleaning agent is enough to be used.

In the following, we will describe in detail the lane attribute mirror in the static sign: whether human driving or automatic driving, at a busy intersection, we are not interested in the precise values of the longitude and latitude of the vehicle location and surrounding buildings, and we are interested in the lane sidelines, sidewalk sidelines, STOP lines, the distance from the intersection or some exit, which lane to locate, whether to allow the vehicle to turn and go straight next, and the like. This information may be communicated to the vehicle by placing a series of coded mirrors on the lane in which the vehicle is traveling. For example, the vehicle is now in the lane of the few turns, which is the left and right turn and the straight lane. This is the most accurate "lane level positioning". Thus, a high-precision map is not necessary. Although a high-precision map may provide some a priori information to the unmanned vehicle, including road curvature, heading, grade, and bank angle, etc., this information is critical to the safety and comfort of the unmanned vehicle. However, the information does not need the "lane-level precision", and the information is shared by a certain road section, even including severe weather such as heavy rain, heavy snow, dust and the like encountered at the present moment, and the information such as obstacles, passable spaces and the like which are far away are difficult to be sensed by the vehicle-mounted sensor is pushed to all vehicles arriving at the road section through the LTE-V2X communication system arranged at the information center of the road section. Neither a high-precision map nor 5G communication is necessary.

In particular, in the case of adding and discarding roads and lanes, before the adding and discarding are carried out, hardware (reflecting mirror) or software (LTE-V2X information pushing of local road sections) for allowing the entering of the added roads and lanes or forbidding the entering of the discarded roads and lanes is arranged on the relevant lanes at a certain distance before the entering of the added or discarded roads and lanesSend), information of such changes as road and lane addition and abandonment can be notified to the vehicle coming before immediately after the change is executed. This easily solves the problem that the expensive high-precision map needs to be updated at a higher cost after being put into use (as described by the Gade map company)^[5]Mapping is only 10% of the work done, and another 90% of the competitive strength is on data iterative updates).

At intersections with different complexity, the attribute information corresponding to the lane of the vehicle coming along the own lane is notified by using different coded mirrors. The method is as follows.

The following symbolic semantic interpretation is agreed in advance:

1. straightening: a straight-through lane;

2. left: a left turn lane;

3. and (3) right: a right-turn lane;

4. left/straight: left-turn and straight lane;

5. straight/right: right-turn and straight-going lane;

6. p is a natural number 1,2,3, …, q is (r), (c), (…), p or q indicates 0 in the absence thereof (the same applies below). It means that p lanes are crossed from the lane to the right to reach a straight lane, and q lanes are crossed from the lane to the right to reach a right-turn lane. For example: and the left 1 is the lane for left turning, if the vehicle needs to go straight, the vehicle needs to cross 1 lane, and if the vehicle needs to go right, the vehicle needs to cross 2 lanes.

Qp right indicates that p lanes are crossed to the left from the lane to reach the straight lane and q lanes are crossed to the left from the lane to reach the left turn lane. For example: and 2, the right side indicates that the lane is a right-turn lane, if the lane needs to go straight, the lane needs to cross 1 lane, and if the lane needs to turn left, the lane needs to cross 2 lanes.

8.q₁Straight q₂Indicates that q is to be crossed from the lane to the left₁The lane can reach the left-turn lane and the lane will cross q from the right₂The strip lane can reach the right-turn lane. For example, the third step is to indicate that the lane is a straight lane, if a left turn is needed, 3 lanes need to be crossed, and if a right turn is needed, 2 lanes need to be crossed.

9. Symbols and their definitions may also be added if the above definitions cannot cover more complex lanes.

Fig. 1 shows a road scene, which is a bidirectional four-lane scene, wherein a left-turn straight-going sign 151a and a right-turn straight-going sign 151b are respectively disposed on two lanes on the right side, and corresponding to the scene of the figure, there may be the following table:

TABLE 1 significance of mirror group coding example 1

For the same reason, the following table may be provided for other scenarios.

Table 2 mirror group coding meaning example 2

Table 3 meaning example 3 of mirror group coding

The same approach can be used to define the sign of an intersection lane for any number of lanes. With this information, the control system knows whether the current lane allows the next operation in the vehicle's movement plan, and if not, the lane that allows the next operation can be reached by traversing several lanes left or right. The instruction is direct and clear, and does not need to look up a map or calculate.

Similarly, in the multi-layer tunnel and overpass area, the reflector arranged on the lane representing the-2, -1 to 0 (ground) of the tunnel and then the 1 st, 2 nd, … th, 5 th layer of the overpass can be used to inform the vehicle of the layer number.

In addition, when all the reflectors are arranged on the lane center line, the information of the lane center line indication reflector and the information of the reflector with different lane symbols can be combined into a whole. For example, while we use differently coded (i.e., different symbols) mirrors to represent which lane, the positions of these mirrors are again the center line of the lane. In fact, if we agree that all mirrors are exactly mounted on the lane centreline, then all mirrors can provide a specific instruction according to their code, whose position in turn can represent the position of the lane centreline. Similarly, if we agree that all mirrors are exactly mounted on the lane boundary, then all mirrors can provide a specific instruction according to their code, whose position can in turn represent the position of the lane boundary. For a simple road far away from a city, if vehicle-mounted optical camera shooting and millimeter wave radar image processing are used, lane center line (or side line) information can be obtained, and a reflector method can be omitted on the basis of further improving software performance and ensuring that the requirement of automatic driving safety on the road section is met.

The arrangement of the mirrors may be one-dimensional (e.g. bar codes) or 2-dimensional (e.g. 2-dimensional codes). We can imagine our scanning of the road edge (or road surface) mirror as a scanning reader of a bar code (or 2 d code) on the ground while the vehicle is in motion.

In addition, a map (1 st map) indicating the correspondence between the code and the traffic sign may be stored in advance in the storage device of the vehicle, and the control unit may determine the traffic sign indicated by the code corresponding to the reflected wave from the map. It goes without saying that a table (table 2) indicating the correspondence between the reflected wave and the code may be stored in advance in the storage device of the vehicle, and the control unit may specify the code indicated by the received reflected wave based on the table.

Such active transmission and passive reflection systems can be technically and constructively followed by the following principles: 1. the active transmitter-receiver belongs to the vehicle-mounted device, and the cost is not the center of consideration. However, passive reflectors, because of their large number, must be cost-effective. For example, in the lane line marking, considering that 1 is installed on a road at a certain distance (for example, 5 meters), the setting density can be increased (i.e., the setting distance can be decreased) at the turning of the lane. If the lane marking reflector is placed in the center of the lane and an automatic driving lane keeping system (LKA) keeps the center line of the vehicle consistent with the lane marking, then for a two-lane road, 2 lane marking lines are needed. If lane markings are placed at the edges of the lanes, the lane keeping system that is automatically driven keeps the center line of the vehicle equidistant from the lane markings on both sides, then for a two-lane road, 4 lane markings are required. Therefore, the lane line mark is saved by placing the lane line mark at the center of the lane. Even if roads all around the world need to be modified, the lane marking reflectors which need to be installed are astronomical numbers. It uses only one of the most economical single mirrors (e.g., a mirror encoded as a "0") and does not use 2 bits or more of encoding. The cost of such a mirror is preferably around 200 dollars or $ 30 for current prices. 2. If all traffic sign reflectors are mounted on the road surface of each lane (e.g., the center line of the lane) to facilitate the operation of the vehicle-mounted transmitter-receiver, the reflectors can be hidden in small recesses in the road surface that are not touched by the wheels, the wheels cannot hurt the reflectors, and the reflectors do not hinder the vehicle from walking. 3. The reflecting mirror is arranged in a large quantity, and a machine like a rice transplanter can be considered to be invented and automatically arranged on a road. 4. A reflector (for example, the most economical single reflector coded as "1") can be arranged at the doorway (the departure point) of each building such as a home, a shop, an office building and the like as a parking mark, and when a vehicle travels to approach the departure point, the vehicle traveling on the center line of the lane must travel to the departure point under the guidance of the traffic mark of "parking beside" for parking; in addition, a reflector may be provided at the side of a street where traffic is not heavy, so as to guide a free taxi to park and wait for a taxi (so-called "groveling"). The arrangement on the street is to take into account that most parking lots will be cancelled later, so that taxis are parked by using an open street.

Traffic identification of dynamic information is discussed below. The warning signs mainly refer to traffic lights, slippery roads in rainy and snowy days, strong wind, traffic congestion in the front of the roads and the like. The inventor's approach is embodied in mounting the traffic lights' active electronic command transmitters directly on each lane, e.g., the lane center (or roadside), as the passive mirrors described above. The electronic command transmitter of the traffic light receives the electric signal (not the color and image information needing to be identified) representing the red, green and yellow light from the existing signal light system of the intersection, and the vehicle sequentially receives the electronic command signal of the traffic light corresponding to the lane starting from a specified distance (such as 100 meters) away from the intersection and executes the electronic command signal according to the command and other cooperative commands when reaching the intersection. Since the electronic command transmitters of the traffic lights are installed on the respective lanes, the problem of which traffic light signal is to be followed by the intersection with multiple traffic lights mentioned above is solved. Since the number of electronic command transmitters of such dynamic information traffic lights is orders of magnitude less than the static information reflectors mentioned above, cost is not the most significant problem.

In order to save a large amount of reflector manufacturing and installation cost problems, it is also an option to directly provide roadside traffic signs and command signal electronic instruction data summary tables for vehicles by adopting active ground transmission-vehicle reception for most static traffic signs (such as speed limit), all dynamic traffic signs, traffic lights and other signals except reflectors which are only related to positions, such as lane lines, lane attributes, get-off points and the like. The specific manner in which this mix of hardware and software is described in detail below:

on the whole, a sending device which uniformly stores an instruction data general table formed by most static and all dynamic traffic marks arranged on the road section and all traffic control signals such as traffic lights and the like can be arranged at certain intervals on the roadside (particularly when the number of marks such as speed limit and the like is changed), the data general table is sent to all vehicles coming forward at certain frequency by using proper physical waves (such as ultrasonic waves, electromagnetic waves and the like), and after the data general table is received by a vehicle-mounted receiver, an instruction suitable for the vehicle is selected from the instruction data general table according to the characteristics of the vehicle (the vehicle type, the load capacity, the lane confirmed by hardware devices such as a lane attribute mark reflector, the distance from an intersection and the like) and is transmitted to an automatic automobile driving unit. Since at regular intervals along the road there is a transmitting device for traffic identification and command data summary formed by control signals, the vehicles will overlap or mix up signals within the effective transmission range of two or more such transmitting devices. Therefore, each data summary table must have the link number of the transmitting device, and the vehicle can select the matched instruction data summary table from the link numbers acquired from the navigation map (or from the road side 'link number' reflector) during operation.

In addition, technically, to realize the practical implementation of the present invention, it is important to have the above-mentioned on-board transmission wave generator, mirror, reflected wave receiving and processing device. The inventors have a preliminary consideration that existing millimeter wave radars may be used as the point of entry. Because the electromagnetic wave in the millimeter wave band cannot be influenced by common environmental factors such as rain, fog, dust and the like, the millimeter wave has all-weather strong anti-interference detection performance. In addition, in detection distance, because the attenuation of the millimeter wave in the atmosphere is weak, a longer distance can be detected and perceived, and the detection distance of the medium-long range millimeter wave radar can reach 250 meters. Therefore, during high-speed travel, the millimeter wave radar can detect various mirrors in front of the own lane earlier. In terms of manufacturing process, the millimeter wave radar has the characteristics of small volume, light weight, easy integration and the like, and is easy to install on an automobile; in terms of cost, the price of the millimeter wave radar can be controlled to be about one thousand yuan. It is also a good feature that millimeter waves can penetrate plastics but are very sensitive to metals. If different metals are used as the reflector and the metal reflector is arranged in the plastic box and is buried on the lane line, the reflector can be well protected.

< embodiment mode 1: manner of biasing hardware >

The method is characterized in that all static marks use a passive reflector to transfer the codes of the marks to the vehicle; all dynamic identifiers are coded by respective active transmitting devices to transfer the identifiers to the vehicle.

Fig. 2 is a schematic diagram showing an automatic driving method and a road system configuration of an automobile according to embodiment 1.

As shown in fig. 2, in the present embodiment, a case where a vehicle is going to pass through the intersection will be described as an example. Specifically, taking a bidirectional four-lane example, the vehicle 200 travels on a road 101, the other side of the road 101 is an oncoming road 102, the front side is an intersection 103, a normal lane boundary sign 11 is drawn on the road 101, and a sidewalk sign 12 is drawn between the road 101 and the intersection 103. In the present embodiment, a plurality of mirror groups 11a are provided on the road surface along the center line X (virtual line) of the lane. The mirror group 11a includes a lane center line mirror having only one mirror (for example, a "0" mirror) at the lowest cost, and 2 or more different mirrors having different reflection characteristics, which are arranged in different order to generate codes.

In addition, the mirror cluster is placed near the lane center line in fig. 2, as compared with the mirror cluster disposed along the lane edge mark 11, in order to avoid the situation that the vehicle may be covered by the vehicle on the adjacent lane, and in the case that the vehicle is covered, the transmission-reflection is blocked. In addition, mirror groups are also provided on the oncoming road 102, but these mirror groups are not shown in the figure. In addition, when all the reflectors are arranged on the center line of the lane, the information of indicating the position of the center line of the lane and different coded reflectors can be combined into a whole. While we use differently coded mirrors to represent different commands, the position of these mirrors is again the center line of the lane. In fact, if we agree that all mirrors are exactly mounted on the lane centreline, then all mirrors can provide a specific instruction according to their code, whose position in turn can represent the position of the lane centreline. Similarly, if we agree that all mirrors are exactly mounted on the lane boundary, then all mirrors can provide a specific instruction according to their code, whose position can in turn represent the position of the lane boundary. Of course, the reflector can also be hung right above each lane, which is most beneficial to scanning and signal receiving and only needs to increase the cost.

In the present embodiment, a pit (not shown) is provided on the road surface of the road, and the mirror group 11a is provided in the pit.

In addition, the vehicle 200 has a transmitter-receiver 201 and a control unit 202 connected to the transmitter-receiver 201. The transmitter-receiver 201 transmits a transmission wave which is reflected by the mirror group 11a placed in the vicinity of the center line of the lane to form a prescribed reflected wave, the transmitter-receiver 201 receives the reflected wave and transmits information indicating the reflected wave to the control unit 202, the control unit 202 obtains a code, and determines a road sign corresponding to the reflected wave based on the code to perform automatic driving of the vehicle 200, specifically, to run the vehicle 200 "on" the recognized road sign. This is a mode in which the mirror group 11a is provided on the lane center line X.

In addition, in the present embodiment, a plurality of transmitting device groups 13 (including transmitters and storage devices) are provided beside (on the side of) the road 101, and the transmitting device groups 13 are provided at predetermined distances (for example, 100m) from the intersection 103 and emit electronic command information indicating traffic light states and other dynamic signs indicating the traffic rules of the road 101 at the intersection 103. The vehicle 200 receives the electronic instruction information (via an information transceiver not shown), and executes automatic driving of the vehicle 200 based on the electronic instruction information.

Referring to fig. 3, a concrete example of the lane attribute mirror will be described, and in this example, a vehicle 200 travels on a lane 101a dedicated for left turn, a left turn sign 101b is provided on a ground of the lane 101a near an intersection, and a mirror 111a whose code indicates that the lane 101a is a lane dedicated for left turn, cannot go straight, and cannot turn right is provided near the left turn sign 101b, with reference to fig. 3.

The above is an implementation with a bias towards hardware. The safest should be the way in which all static identifications are biased by the mirror/mirror combination hardware, as compared to the two following ways.

< embodiment mode 2: software and hardware mixing mode >

The method is characterized in that all static marks related to positions (such as lane lines, lane attributes, crosswalks and the like) are coded and transmitted to the vehicle by using the passive mirror; all static signs (such as speed limit) and all dynamic signs are all active ground emitting-vehicle receiving emitting devices to transmit command data summary tables to vehicles.

In order to save a lot of problems of manufacturing and installing cost of the reflector, a software and hardware mixing mode is adopted, namely, the sending device group 13 of fig. 2 is integrated into a sending device, the sending device group is stored with an instruction data total table which represents most static traffic identifications and all dynamic traffic identification information of the road section arranged by the sending device 13 except the reflector of the static identifications such as the lane line related to the position, and the sending device 13 sends the data total table to the running vehicle 200 at a certain frequency. Vehicle 200 selects information suitable for the vehicle from the data summary table according to the situation of the vehicle, and control section 202 controls the vehicle according to the information. The "self-situation" referred to herein includes, for example, a self-vehicle type, a load capacity, lane and link information determined by a reflector such as a lane attribute mark and a link mark, a distance from an intersection, or any combination thereof. In addition, the link number of the link set by the transmission device 13 may be included in the data summary table.

< embodiment 3>

1. The lane line mark needing the daily setting adopts a passive single-chip metal reflector. As described in [0058], it is located with a millimeter wave radar device on board.

2. In addition to the lane marking mentioned in item 1, the static dynamic marking and the command signal all use active transmitters powered by the roadside to send coded commands specific to each marking to the vehicle, instead of using mirror arrays with different reflection wave characteristics to form different codes as in < mode 1> and < mode 2>, and then the different codes are received by a vehicle-mounted receiver and converted into coded commands.

3. The method in the above 2, can be realized by using a roadside-powered active Radio Frequency Identification (RFID) method. For example, vehicle-mounted Ultra High Frequency (UHF) or microwave frequency and rfid reading devices with an effective reading range of 1-2 meters are used to obtain information such as lane attributes from rfid tags on roadsides or roadways. Of course, such rfid readers and tags must meet automotive autodrive accuracy and longevity requirements.

Other roadside commands besides the above 1./2./3. can be included in the command data summary table and sent to the vehicle by the LTE-V2X method.

< summary of the embodiments >

Since the traffic guidance system equipment of modern highways is based on human driving of vehicles, almost all signs and guidance signals are characters, figures, shapes, colors and the like which are recognized by human drivers by eyes. At present, most of researchers for automatic driving at home and abroad adopt an artificial intelligence recognition method, but the recognition accuracy rate of the method cannot meet the requirement of safe driving in the present and relatively long-term future. The inventor considers this to be an incorrect technical route, and for this reason, the invention proposes a practical and effective technical route.

The invention provides an automatic driving method for an automobile, wherein a specially arranged device is arranged on the roadside or the road surface of a lane, and the device directly provides lane-level positioning, roadside traffic identification and command signal electronic instructions required by automatic driving for the automobile. The device specially arranged can be a passive reflector group, a control unit, a transmitter and a receiver are arranged on a vehicle, a transmitted wave transmitted by the transmitter is reflected by the reflector group, the receiver receives a reflected wave formed by reflecting the transmitted wave by the reflector group, the control unit determines a coded command represented by the reflected wave according to the reflected wave and further determines the command content represented by the coded command, and controls the vehicle according to the determined command content, wherein the command content represented by the coded command of the reflector group is information of a lane line, the position and lane attribute of the vehicle and a roadside traffic mark and command signal of the road section. In addition, the specially arranged devices can also be active emitting device groups. At the moment, a control unit and a receiver are arranged on the vehicle, the receiver receives the waves transmitted by the transmitting device group to the receiver, the control unit determines the coding instruction represented by the transmitting waves according to the transmitting waves and further determines the instruction content represented by the coding instruction, and the vehicle is controlled according to the determined instruction content, wherein the instruction content represented by the coding instruction of the transmitting device group is the information of the position and lane attribute of the vehicle and the road roadside traffic identification and command signal of the road section.

According to the embodiment of the invention, different reflected waves of different passive reflectors form coded signals or different coded signals sent by an active transmitting device group to represent different roadside signs and command signals, and vehicles recognize traffic signs and command signals according to the codes, so that compared with the mode of artificial intelligent image recognition and high-precision map positioning in the prior art, the method can greatly reduce the cost, improve the recognition precision and reliably carry out automatic driving.

The invention particularly provides an automatic driving lane level positioning method, wherein a passive reflector group is arranged on the road surface of a lane, a control unit, a transmitter and a receiver are arranged on a vehicle, the transmitter transmits a transmission wave to be reflected by the reflector group, the receiver receives a reflected wave formed by reflecting the transmission wave by the reflector group, the control unit determines a coding command represented by the reflected wave so as to determine the lane attribute command content represented by the coding command according to the reflected wave, and the vehicle is controlled according to the determined command content. Similarly, different coded signals sent by the active transmitter group can be used to represent different lane attributes, and the coded signals are received by the vehicle-mounted receiver and then sent to the control unit to control the vehicle.

By adopting the method, the lane attribute information can be provided very effectively and accurately only by arranging the corresponding reflector group or the corresponding transmitting device on the lane, so the method is simple and practical.

Preferably, the lane attribute information includes driving rule information. The travel rule information here includes, for example, straight-ahead lane information, left-turn lane information, and the like.

In addition, it is preferable that the lane attribute information includes position information of a lane in a road. The "position information of the lane in the road" herein refers to information that there are several adjacent lanes on the left and/or right side of the own lane and the attributes of these adjacent lanes.

According to the embodiment of the invention, the automatic driving method of the automobile is provided, wherein a plurality of reflector groups are arranged beside a lane or on a road surface, a control unit, a transmitter and a receiver are arranged on the automobile, a transmitted wave emitted by the transmitter is reflected by the reflector groups, the receiver receives a reflected wave formed by reflecting the transmitted wave by the reflector groups, the control unit determines a code represented by the code according to the reflected wave and further determines the content represented by the code, and the automobile is controlled according to the determined content. The transmitter and the receiver may be one device or may be formed independently. Similarly, the vehicle may be controlled by using different coded signals from the active transmitter group, which are received by the on-board receiver and passed to the control unit to determine the content indicated by the code based on the code.

In the method, optionally, a table indicating a correspondence between the code and the traffic sign is stored in advance in a storage device of the vehicle, and the control unit determines the content indicated by the code based on the table stored in advance and controls the content.

In the above method, optionally, the control unit determines the static-dynamic traffic identification of the coded representation from a pre-stored graph.

In the above method, the transmitted wave is optionally an ultrasonic wave, an electromagnetic wave, a light wave, a laser or other physical waves available.

In the above method, optionally, the mirror group includes 1 or more mirrors, and the reflection characteristics of the mirrors for the reflected wave in the mirror groups are different, so that the encoded signal can be formed by different arrangements of the mirrors.

In the above method, optionally, the reflected waves formed by the mirrors have different frequencies, waveforms, peaks, widths, waveform integrals, and the like by the difference in size, quality, shape, surface finish, or flatness, and the like of the mirrors, so that different reflected waves are obtained.

On the other hand, in the above-described software and hardware hybrid system, in the automatic driving method for an automobile, a transmitting device (electronic command transmitter and storage device) is optionally provided on a roadside or a road surface, the transmitting device stores therein a command data summary table indicating information of most of static traffic signs and all of dynamic traffic signs of a road section provided by the transmitting device, excluding a lane line and a parking spot, and the transmitting device transmits the data summary table to a vehicle traveling on the road section at a certain frequency. And the vehicle receiving the data summary table selects instruction information suitable for the vehicle from the data summary table according to the condition of the vehicle, and the control unit controls the vehicle according to the instruction information. The vehicle selects information suitable for the vehicle from the instruction data summary table based on the vehicle type, the load capacity, the lane and link information determined from the mirrors such as the lane attribute and the link identification, the distance from the intersection, or any combination thereof. The data summary table includes link numbers of the links set by the transmitting device.

In the road aspect, the present application provides a road system, wherein a plurality of reflecting mirrors are arranged beside a lane or on a road surface, the reflecting mirrors comprise a plurality of reflecting mirrors, and the reflecting characteristics of the reflecting waves are different by the plurality of reflecting mirrors, so that a coded signal can be formed through reflection.

In the road system, optionally, the reflection waves formed by the plurality of reflection mirrors have different frequencies, waveforms, peaks, widths or waveform integrals due to the difference of the sizes, qualities, shapes, surface finishes or flats of the reflection mirrors, so that the different reflection waves are obtained.

In the road system, optionally, a pit is provided on the road surface, in which pit the mirror cluster is arranged.

In a road system, a software/hardware hybrid system is characterized in that a transmitter is provided on a roadside or a road surface, and the transmitter transmits electronic instruction information indicating (the state of) a static or dynamic traffic sign other than a sign indicating only a position property such as a lane line.

In the road system, the transmitting device is optionally provided at a prescribed distance from the intersection, and at a place where the static traffic sign changes.

Summary of embodiment 3: 1. the lane line mark needing the daily setting adopts a passive single-chip metal reflector, and a millimeter wave radar device is arranged on the vehicle to position the lane line mark. 2. All static and dynamic identification and command signals except for the lane line identification mentioned in the description of 1. adopt active transmitters powered by the roadside to send coded commands specific to each identification to the vehicle, instead of forming different codes by using mirror arrays with different reflection wave characteristics as in the case of < mode 1> and < mode 2>, and then converting the codes into coded commands after being received by a vehicle-mounted receiver. 3. The method in the above 2, can be realized by using a roadside-powered active Radio Frequency Identification (RFID) method. For example, vehicle-mounted Ultra High Frequency (UHF) or microwave frequency and rfid reading devices with an effective reading range of 1-2 meters are used to obtain information such as lane attributes from rfid tags on roadsides or roadways. Of course, such rfid readers and tags must meet automotive autodrive accuracy and longevity requirements. 4. Other roadside commands besides the above 1./2./3. can be included in the command data summary table and sent to the vehicle by the LTE-V2X method.

In addition, the methods of all the embodiments of the present invention may be used in combination with a method of acquiring an instruction by communication, that is, when the vehicle is traveling, there is a vehicle in the left and right lanes to merge into the current lane, and when the driver drives, the taillights are blinked in front of the left and right of the host vehicle to indicate his intention to the host vehicle. In the occasion of the main technical route of automatic driving, the sensing system is also used for searching and capturing and then is identified by an artificial intelligence method. In the method of the present embodiment, if the vehicle that requires the merge directly sends a message to the automated vehicle to request the vehicle to merge with it using V2V communication, it is not necessary to use an artificial intelligence "recognition" method. After the vehicle automatic driving system receives the V2V information of the vehicles requiring the doubling, necessary measures are taken to enable the vehicles requiring the doubling to merge.

That is, in the present invention, the vehicle has 3 methods to obtain the external control signal: 1. a ground "mirror"; 2. individual instructions transmitted by the roadside or "instruction data summary"; 3. other vehicles directly communicate request information sent to the automotive vehicle using V2V; 4. but of course also information from other information sources. The core idea of this patent is that it is desirable that this information is not ambiguous instruction information, rather than source-start information that requires "recognition" and complex computation.

< position of the technology of the present application in the full stack solution for automatic vehicle driving >

The full stack solution for a typical autonomous vehicle is divided into four parts^[6]. (1) And planning a route through a road network. This part of the task can be basically accomplished with the GPS now in widespread use. (2) A behavioral layer that decides local driving tasks that push cars to destinations and obey road regulations. (3) The motion planning module selects a continuous path through the environment to complete the local navigation task. (4) The control system corrects errors in performing the planning movement in real time. The method is mainly applied to directly providing electronic instructions of road side traffic identification static state and traffic light dynamic information for vehicles in the step (2), so that the vehicles strictly follow the road rules in the driving process and are not responsible for detecting moving vehicles, pedestrians and obstacles which randomly appear, and therefore the method is not related to the tasks of the steps (1), (3) and (4).

< analysis of practical applicability and economic feasibility >

Modern highways are built over the last hundred years of accumulation. Is it economically feasible if modified according to the method provided by this patent? The authors here provide a simple set of data. The total highway of China is 500 kilometers; 665 ten thousand kilometers of national roads (2015); the global highway is 6500 kilometers (2013). Estimated in two lanes (which should be the majority): 1 lane marking reflector is arranged at the center line of each lane at a certain distance (for example, 5 meters), and 1 kilometer of the lane is provided400 roads are installed. The total of 8 ten thousand RMB is needed by the calculation of 200 yuan for each reflector. Plus other static, dynamic information devices and engineering costs much less than lane marking mirrors, 100 million RMB per kilometer should be sufficient. On a 500-kilometre highway in China, 5 trillion RMB are required. Calculated as 99 trillion GDPs in 2019, is about 1/20 of GDPs in 1 year nationwide. This algorithm is applied to the total number of U.S. roads and GDP, which is also approximately 1/20 for GDP in the U.S. 1 year. If the road reconstruction is done in 10 years, the cost is 1/200 for the GDP of 1 year. For China, 5000 hundred million RMB are used each year. According to the reference^[2]The annual investment in the automotive industry from 2016 to 2019 in china is 6014, 9938, 8528 and 5940 billion, respectively, which exceed 5000 billion. Profitably, if the method of the patent is implemented after 10 years, it is obtained that most of the garages of homes, almost all shops, streets and parking lots of public places are freed; the national GDP increases by 0.5% every year; the death rate of global traffic accidents is reduced by dozens of times; most importantly, human beings have finally entered the era of auto-driving of automobiles!

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

[ REFERENCE ] to

[1] Auto-pilot 2019: what besides the embarrassment of full screen is what is the frustration of creating and doing business/great?

https://zhidao.baidu.com/question/925163743268075699.htmlfr＝iks&word＝％B1％B1％BE％A9％CA％D0％C6％FB％B3％B5％D7％D4％B6％AF％BC％DD％CA％BB％D1％D0％B7％A2％C7％E9％BF％F6&ie＝gbk

[2]2020 + 2023 research on development trend of high-grade automatic driving industry in China

Self-media synthesis of Xinlang finance and economic benefit in 04/2020

https://finance.sina.com.cn/stock/stockzmt/2020-04-06/doc-iimxyqwa5382413.shtml

[3] Overview of autopilot development and Key technologies

https://new.qq.com/omn/20190604/20190604A0JJYB.html

[4] Lane level positioning method based on vision and millimeter wave radar

https://www.sohu.com/a/224987096_465591

[5] Is a prodigious map a lingdan miao medicine for drawing high-precision maps?

https://www.d1ev.com/kol/81666

[6] Overview of urban road autonomous vehicle motion planning and control technology

https://blog.csdn.net/hgdwdtt/article/details/82052577

Claims

1. A method for automatically locating the driveway of car and recognizing the traffic signs and command signals of roadside features that a special device is arranged at the roadside or road surface of driveway, and said device can directly provide the lane line information, driveway location, roadside traffic signs and electronic command of command signals.

2. The method for automatic driveway-level positioning of automobiles and identification of roadside traffic signs and command signals according to claim 1, it is characterized in that the device specially arranged on the roadside or the road surface of the lane can be a passive reflector group, the vehicle is provided with a control unit, a transmitter and a receiver, the transmitted wave transmitted by the transmitter is reflected by the reflector group, the receiver receives a reflected wave formed by reflecting the transmitted wave by the mirror group, the control unit determines the coded command represented by the reflected wave according to the reflected wave and further determines the command content represented by the coded command, controls the vehicle according to the determined command content, the instruction content represented by the coded instruction of the reflector group is information of a lane line, the position and lane attribute of a vehicle and roadside traffic identification and command signals of the road section.

3. The method according to claim 1, wherein the device is specially installed at the roadside or on the road surface of the lane, or an active transmitting device group, the vehicle is provided with a control unit and a receiver, the receiver receives the wave transmitted by the transmitting device group to the receiver, the control unit determines the coded command represented by the transmitting wave according to the transmitting wave and further determines the command content represented by the coded command, and controls the vehicle according to the determined command content, wherein the command content represented by the coded command of the transmitting device group is the position and the lane attribute of the vehicle and the information of the roadside traffic sign and the guidance signal.

4. The method for automotive autonomous lane-level positioning and roadside traffic marking and directing signals identification according to claim 1,2 or 3, characterized in that the lane attribute information includes driving regulation information, the lane attribute information includes: a) the position of the lane in the road, namely which lane, is convenient for selecting corresponding traffic light signals in a plurality of traffic light signals at the intersection in a command data summary table received by the vehicle immediately according to the information; b) the left-right turn or straight driving rule allowed by the lane; or c) if the lane does not allow the next operation of driving, the target lane can be reached only by crossing several lanes leftwards or rightwards, the position of the vehicle comprises the lane side line, the sidewalk side line, the STOP line, the distance from the vehicle to the intersection and the like, and the roadside traffic marks and command signals refer to all static-dynamic traffic marks, traffic lights and other traffic command signals of the road section.

5. The method for automatic driveway-level positioning of automobiles and identification of roadside traffic signs and command signals according to claims 2 and 3, wherein the control unit determines the contents of static and dynamic traffic signs and command signals represented by coded instructions to control the vehicles according to a pre-stored chart.

6. The method for automatic driveway-level positioning of automobiles and identification of roadside traffic signs and command signals according to claims 2 and 3, wherein the transmitted waves are physical waves such as ultrasonic waves, electromagnetic waves, light waves or laser.

7. The method for automatic driveway-level positioning of automobiles and identification and command signals of roadside traffic signs and signals according to claim 1 or 2, wherein the reflector group comprises 1 or more reflectors, and the reflectors in the reflector group have different reflection characteristics to the reflected waves, so that coded command signals can be formed by different arrangements of the reflectors.

8. The method according to claim 7, wherein the reflecting mirror has different sizes, materials, shapes, surface smoothness or flat recesses, so that the reflecting wave formed by the reflecting mirror has different frequencies, waveforms, peaks, widths or waveform integrals, thereby obtaining different reflecting waves.

9. The method for automatic driveway-level positioning of automobiles and identification of roadside traffic signs and direction signals according to claim 1 or 2, wherein the arrangement of the mirrors in the mirror group is one-dimensional or two-dimensional.

10. The method for automatic lane-level positioning and roadside traffic sign and command signal identification of automobiles as claimed in claim 1 or 2, wherein when all said reflectors are arranged on the center line of the lane, different codes are used to represent different roadside signs, and the positions of the reflectors are the center line of the lane, and similarly, when all reflectors are accurately arranged on the side line of the lane, then all reflectors provide a specific instruction according to the codes, and the positions of the reflectors represent the positions of the side line of the lane; alternatively, the mirror may be suspended directly above each lane.

11. A road system is characterized in that a plurality of passive mirror groups are arranged beside a lane or on a road surface, each mirror group comprises 1 or a plurality of mirrors, and the mirror groups directly provide lane line information, lane level positioning, roadside traffic identification and command signal electronic instructions required by automatic driving for vehicles; a plurality of active transmitting devices can be arranged beside a lane or on a road surface, and the transmitting device group directly provides lane-level positioning, roadside traffic identification and command signal electronic instructions required by automatic driving for the vehicle.

12. The roadway system of claim 11, wherein the mirrors in the plurality of mirror groups have different reflection characteristics for the reflected wave, such that the encoded signal can be formed by different arrangements of mirrors.

13. A road system according to claims 11 and 12, characterised in that a pit is provided in the road surface, in which pit the mirror or the emitting device is arranged.

14. The road system according to claim 11, wherein a transmitting device is provided on a roadside or a road surface, the transmitting device stores therein a data summary table indicating information of static traffic signs and dynamic traffic signs of a road section provided by the transmitting device, and the transmitting device transmits the data summary table to vehicles traveling on the road section at a certain frequency.

15. The road system according to claim 14, characterized in that the transmitting means are arranged at a place where a static traffic sign changes or a prescribed distance from an intersection.

16. A road system according to claim 11, characterised in that a reflector representing a parking sign or a transmitting device of a parking sign is provided near the entrance of the building and at the parking position of the taxi.