Write in front of the text

Intelligent driving has been one of the hottest topics in the automotive industry in recent years. After the intelligent driving system has rapidly developed to L2 level, it seems to have hit a hard wall under the influence of technology development, driving environment infrastructure construction, laws and regulations and other factors, so that people in the industry have begun to carefully examine the implementation path of L3 and above high-level intelligent driving systems. As a traditional Autobot, this article is just a summary of my superficial thinking, and I will share and discuss it with my colleagues in the industry.

text

Intelligent driving automobile is one of the most rapidly developing emerging industries in the information age of the fourth industrial revolution. With the rapid development of new sensor technology, information and communication technology, automatic control technology, computer technology and artificial intelligence, intelligent driving system has also developed rapidly. At present, the three main technical links of intelligent driving system are perception layer, decision-making layer and control execution layer. The correct perception of the driving environment by various sensors in the perception layer is the prerequisite for intelligent driving and the basis for ensuring the follow-up intelligent driving control. However, due to the limitations of sensor technology, environmental interference, perception accuracy, recognition algorithm, data fusion technology and other factors, it has become the main challenge to achieve high-level intelligent driving,

1、 Current situation and challenges of driving environment perception

As shown in the architecture of the intelligent driving control system (Figure 1), the sensing layer of the intelligent driving control system uses various sensors, such as visual sensors (vehicle mounted cameras), radar sensors (ultrasonic radar, millimeter wave radar, laser radar, etc.), positioning sensors (ground map, satellite positioning), etc., to detect and identify various information in the vehicle interior and driving environment. Through data processing and data fusion technology, It provides the decision-making basis for the driving control behavior of the vehicle, which is the basis for realizing the automatic driving control of the vehicle. It is equivalent to the eyes of the vehicle and determines the overall efficiency and performance of the auto drive system.

8d7d0bca-b0d5-11ec-aa7f-dac502259ad0. png

Figure 1 Intelligent driving control system architecture

In order to provide comprehensive and accurate decision-making basis for intelligent driving control system, the driving environment information that needs to be perceived mainly includes the following categories:

Road parameters: Road physical parameters, paths, lanes, road facilities, etc;

Traffic participants: vehicles, people, other objects (animals, road debris, etc.);

Various traffic behavior rules: road speed limit, no traffic, zebra crossing, single line, signal light status, etc;

Natural weather affecting vehicle handling: wind, fog, rain, snow, etc.

Other factors affecting vehicle driving behavior: occasional or temporary abnormal traffic conditions, such as temporary traffic control, safety inspection of drivers and passengers, giving way to fire vehicles or ambulances, etc.

At present, there are great difficulties and challenges in perceiving the above driving environment information through various sensors and providing comprehensive and accurate decision-making basis for the decision-making layer of intelligent driving control system by relying on data processing and data fusion technology, mainly including:

The sensing accuracy of the sensor is insufficient: for example, the data error of sensing the speed and relative distance of the vehicle ahead through the radar sensor.

The sensing data of the sensor is vulnerable to natural environment interference: for example, the radar is disturbed by smoke and dust, rain, snow, fog and other adverse weather conditions, and the visual sensor is affected by many uncertain factors such as illumination, angle of view, scale, shadow, dirt, background interference and target occlusion.

At present, some driving environment elements cannot be perceived, or the perceived cost is very high: for example, vehicles or pedestrians covered by foreground vehicles or obstacles cannot be perceived and recognized through vision or radar sensors, or the use of multi eye stereo vision technology or hybrid solid-state laser radar to improve the perception ability has greatly increased the complexity and cost of perception technology.

Perceptual data processing, data fusion technology is immature, unreliable and unstable: for example, the massive image information obtained through visual sensors needs more advanced digital image processing technology and higher computer hardware computing power; In addition, in the complex traffic environment, the visual sensor still has the problems of difficult target detection, large amount of image calculation and difficult algorithm to achieve. When dealing with the traffic environment with complex road structure and mixed people and vehicles, the perceived and analyzed information is not reliable and stable. The data fusion and intelligent learning algorithm of multi-sensor sensing driving environment information are not mature and stable.

It is precisely because of the above technical and capacity limitations of sensor sensing accuracy, data fusion, recognition algorithm, deep learning algorithm, etc., that the driving environment information provided to the decision-making layer of the intelligent driving control system is seriously distorted from the real driving environment. The reliability of intelligent driving decisions is greatly reduced due to the uncertainty and inaccuracy of driving environment perception, as shown in Figure 2. This is also one of the major obstacles encountered in the progress of intelligent driving from low-level to high-level.

8d8f23c8-b0d5-11ec-aa7f-dac502259ad0. png

Figure 2 Making unreliable and unreliable control decisions based on distorted driving environment data

In order to improve the accuracy and reliability of driving environment perception, the technical route adopted includes the development of sensors with higher accuracy, the optimization of identification algorithms, the development of AI intelligent algorithms, etc., which brings about a huge increase in uncontrollable system complexity, system cost, technology development and iteration cycle, which is also another major obstacle to the progress of intelligent driving from low-level intelligent driving to high-level intelligent driving.

2、 Construction of “driving environment holographic model” based on accurate driving environment data information

Accurate driving environment information data input is the basis for making accurate and reliable intelligent driving control decisions. At present, due to the uncertainty and inaccuracy of driving environment data sensed and identified by various sensors, it is impossible to make credible and reliable intelligent driving control decisions.

We know that the driving environment data related to vehicle driving accurately and truly exist in the real world and are mastered and understood by relevant entities, such as:

Road manager: accurate information of road physical parameters, paths, lanes, road facilities, etc;

Traffic manager: traffic rule information such as road speed limit, no traffic, zebra crossing, one-way traffic, signal light status, etc;

Traffic participants: real-time vehicle status information such as vehicle speed, position, attitude and heading;

Climate monitor: climate information affecting the driving environment such as wind, fog, rain and snow;

Party involved: occasional or temporary abnormal traffic conditions, such as temporary traffic control (traffic police), giving way to fire vehicles or ambulances (drivers of special vehicles), traffic accident avoidance (drivers of accident vehicles), etc.

If the entity mastering these driving environment data inputs accurate driving environment data to the decision-making layer of the intelligent driving control system through the communication device, it will largely ensure the accuracy and reliability of intelligent driving decisions. Before the decision-making layer of the intelligent driving control system architecture, it is necessary to build an accurate holographic model of the driving environment based on these accurate driving environment data information, This model provides accurate decision-making basis for decision-makers.

8db42092-b0d5-11ec-aa7f-dac502259ad0. png

Figure 3 Building a holographic model of driving environment with accurate driving environment data

In the area where intelligent driving is allowed, the road manager can transmit the static accurate geographic information of the detailed road type, pavement width, lane width, lane line type, Lane speed limit, path, geographic coordinates and other information of the road to the controlled vehicle for automatic driving through the roadside RSU equipment to build the holographic model of the driving environment for automatic driving.

Other vehicles in the driving environment send non private vehicle driving control information such as the overall dimension, positioning, speed, heading, posture, driving intention (acceleration, steering, braking, etc.) of the vehicle through V2V equipment in the form of broadcast. The controlled vehicle of automatic driving receives and monitors these information, and according to its own position information, vehicle speed, posture The lane and other information are used to calculate and construct the dynamic space-time positioning of the vehicle in the holographic model of the driving environment (such as the relative distance and relative speed of the vehicle or obstacles in all directions of the vehicle). At the same time, the non private vehicle driving control information such as the overall dimension, positioning, speed, heading, posture, driving intention (acceleration, steering, braking, etc.) of the vehicle is also transmitted through V2V equipment in the form of broadcast, It is used to collect driving environment information of other intelligent driving vehicles.

The roadside unit RSU can also send real-time environmental climate information, intersection signal light status information, pedestrian, non motor vehicle and other dynamic information sensed by roadside sensing equipment in the form of broadcast, which is received and monitored by the automatic driving controlled vehicle to supplement all element information of the intelligent driving vehicle driving environment model.

At the same time, the party involved in an emergency or temporary event in the driving environment or the management and maintenance personnel of the driving environment can broadcast the special event through the roadside RSU or V2V, which is received and monitored by the controlled vehicle for automatic driving and used for the control decision input of the automatic driving vehicle.

After building a driving environment model with all element information based on accurate data, the decision-making level of the intelligent driving control system can observe its own space-time position in the driving environment from the “God’s perspective”, so as to make more accurate and reliable control decision planning.

3、 Change and value of intelligent driving technology route based on “holographic model of driving environment”

The essence of the intelligent driving technology route based on the “driving environment holographic model” is to change the way the driving environment is perceived, build the “driving environment holographic model” with more accurate data, and observe their own space-time positioning in the driving environment from the “God’s perspective”, so as to make more accurate and reliable control decision planning.

This change in the architecture of intelligent driving control system will also bring about corresponding changes in the business model and technology development of the industry, and promote and accelerate the landing and implementation of high-level intelligent driving. This change and its promotion to the development of intelligent driving technology are mainly reflected in:

3.1 lowering the “threshold” and removing the “barrier”

Intelligent driving is a multidisciplinary and interdisciplinary technical complex, which requires strong comprehensive technical strength in many fields, such as high-precision map, high-precision positioning, vision, radar, artificial intelligence, data fusion, simulation verification, etc. The weakness in any field may lead to the unreliability of the intelligent driving control system developed by it. Therefore, the industry threshold of intelligent driving is very high. At the same time, it will form very high exclusive technical barriers and industry monopoly enterprises, which is extremely detrimental to the technological progress and healthy development of the industry.

The control system of the intelligent driving technology route using the “driving environment holographic model” takes the accurate data received by the v2x device as the main input, supplemented by other high-precision maps, high-precision positioning, vision, radar and other perceptual data to build the control system “driving environment holographic model”, when the high-precision maps, high-precision positioning, vision, radar, artificial intelligence, data fusion Simulation verification and other fields require strong comprehensive technical strength, which is greatly reduced. Enterprises in the industry can give full play to their technical expertise and advantages in one or some fields to develop intelligent driving control systems, so as to create a relatively fair development environment for the industry.

3.2 safety redundancy design of intelligent driving control system

The intelligent driving control system is a control system with a very high safety level. The safety redundancy design is a technical means often used by the safety system to ensure the functional safety of the system. The control system using the intelligent driving technology route of “driving environment holographic model” takes the accurate data received by v2x equipment as the main input for the control system to build the “driving environment holographic model”, and other technologies such as vision The data parsed by radar and other sensing methods can be used as the accuracy calibration, data backup and safety redundancy of the “driving environment holographic model” data, so as to further improve the accuracy of the built driving model and the safety of the intelligent driving control system.

For example, the active collision avoidance function needs to collect the external dimension, speed (relative speed), heading, position information (relative distance) and other data of the vehicle in front. These accurate information of the vehicle in front is sent in the form of broadcast through the V2V equipment of the vehicle in front and is received by the controlled vehicle for automatic driving and processed to make control decisions on detour, deceleration or emergency braking; At the same time, the external dimension, relative speed, heading, relative distance and other data of the front vehicle sensed and identified by the on-board visual sensor and radar sensor of the vehicle can be compared and verified with the data broadcast by the V2V equipment of the front vehicle and received by the controlled vehicle. In addition to further improving the accuracy of these data, the data of these two channels can also be used as a redundant backup of the V2V data channel, When the data of a channel is invalid or missing, the system can quickly switch the data of other backup channels, which greatly improves the security performance of the system.

3.3 loose coupling of system software and hardware

When traditional positioning, vision, radar and other sensors are used to perceive and identify driving environment data, due to different technical characteristics of different sensors, their perception algorithms and data fusion methods have relatively unique technical characteristics, which are closely coupled with their adaptive intelligent driving decision-making and control execution algorithms, Thus, it is difficult for the intelligent driving control system integrator to flexibly adapt to the sensing devices (modules) of different suppliers; At the same time, the supplier of a certain sensing device (module) needs to spend a lot of energy and cost to adapt to the intelligent driving control system of different integrators.

The intelligent driving control system architecture based on the “driving environment holographic model” modularizes the architecture of the sensing layer, and adds a layer of “driving environment holographic model” between the sensing layer and the decision-making layer. The sensor sensing module only needs to transfer data according to the uniformly defined interface definition and data format.

8de2bccc-b0d5-11ec-aa7f-dac502259ad0. png

Figure 4 The standardized driving environment data format and interface definition of “driving environment holographic model” are conducive to the loose coupling of system software and hardware

3.4 cost reduction

The intelligent control system adopting the intelligent driving technology route of “driving environment holographic model” reduces the complexity of the system. In addition to reducing the cost of system development and test verification (see the following introduction for details), the requirements for the accuracy and performance of the sensor module are reduced, and the hardware cost of the intelligent driving control system is also greatly reduced.

As the intelligent driving control system architecture is more clearly layered, and the functional components are modular and standardized, for the sensor module suppliers and intelligent driving control system integrators, unified and standard data interface definitions and data formats are adopted, which further reduces the R & D investment in sensor module adaptation and system integration.

3.5 accelerate the realization and landing operation of high-level intelligent driving system

3.5.1 intelligent driving control system development stage:

Because the driving environment perception layer is the foundation of the decision-making layer and the control execution layer, the development of more accurate and reliable intelligent driving decisions and control algorithms is subject to the technical difficulties of vision, radar and other perception accuracy and performance. The intelligent driving control system using the “driving environment holographic model” takes the “driving environment holographic model” as the input of the intelligent driving control system decision-making and control execution layer, which can make full use of the digital simulation technology, Simulate various driving environments and driving scenarios, provide accurate and comprehensive control input for the software control algorithm and logic of intelligent driving control system decision-making and control execution layer, and shorten the development cycle of intelligent driving control system.

3.5.2 intelligent driving control system test verification stage:

Before the intelligent driving system goes on the road, a lot of testing and verification must be done. In addition to the real vehicle testing in the real driving environment, virtual simulation testing is very important and essential. At present, simulation test accounts for most of the workload of the whole test work, especially the effectiveness test of autonomous vehicle function and performance under complex environment and extreme conditions. Virtual simulation test is an efficient test method. However, due to the great technical difficulties in signal simulation of visual sensors and radar sensors, it is generally difficult to connect the real visual sensor and radar sensor signals to the virtual simulation test equipment for virtual simulation test and verification in the test and verification stage of intelligent driving control system. In this case, it is very difficult to create various visual or physical driving environments and driving scenes to give the visual or radar sensor perception input, It makes the test and verification of intelligent driving control system inefficient and lengthy.

The intelligent driving control system adopting “driving environment holographic model” can make full use of digital simulation technology in the test and verification stage to simulate various driving environments. Driving scenes (including simulating various complex environments and extreme working conditions), namely “driving environment holographic model”, are used in virtual simulation test equipment, which greatly improves the efficiency of test and verification and accelerates the mature landing of control strategy algorithm.

At the same time, during the real vehicle test of the intelligent driving control system, the driving environment data sensed by various sensors in the real driving process are recorded. Through the comparison and verification with the “driving environment model” data constructed by v2x equipment, the sensing algorithm can be optimized and improved to further improve the sensing accuracy.

3.5.3 intelligent driving on road operation stage:

Several important problems faced by intelligent driving vehicles on the road:

1) Intelligent transportation laws, regulations and infrastructure are a process of gradual construction and improvement. In this process of realization and construction, intelligent driving vehicles are faced with problems of vehicle operation safety and traffic safety.

2) Vehicles of different intelligent driving levels coexist with non intelligent driving vehicles for a long time. The randomness of human driving behavior of non intelligent driving vehicles or low-level intelligent driving vehicles and the problems of operation safety and traffic safety of high-level autonomous vehicles caused by uncertainty.

Vehicles of all intelligent driving levels that adopt the intelligent driving solution of “holographic model of driving environment”, including the stock of non intelligent driving vehicles, can operate under control in areas or roads with different levels of intelligent driving infrastructure construction and perfection, and improve the safe operation level of vehicles at all levels.

Safe operation in areas with incomplete intelligent driving infrastructure

When a high-level intelligent driving vehicle enters an area or road where laws and regulations and road infrastructure are not perfect, or is not suitable for the operation of high-level autonomous vehicles, human intervention was required in the past, and the vehicle was passively switched or degraded to a low-level manned driving state. This process of passive switching or driver taking over vehicle control has great uncertainty, which brings many problems of vehicle operation safety and traffic safety.

The intelligent driving technology route based on the “holographic model of driving environment” can identify the roads and areas that allow automatic driving response in the road model data transmitted to the intelligent driving vehicle through the roadside unit. Before leaving the area, the driver is actively prompted to take over the vehicle control. When the vehicle is not taken over, the auto drive system actively controls the vehicle to slow down until it stops in a safe area. If you enter the roads and areas where the traffic infrastructure is imperfect or not suitable for automatic driving, or the laws and regulations do not allow automatic driving, the automatic driving function software of the decision-making control layer of the auto drive system will be prohibited or restricted.

Safe operation in a driving environment where high-level intelligent driving vehicles and non intelligent driving vehicles coexist

Vehicles of different intelligent driving levels and non intelligent driving vehicles will run together with high-level intelligent driving vehicles for a long time. The randomness and uncertainty of human driving behavior of non intelligent driving vehicles or low-level intelligent driving vehicles bring many problems of operation safety and traffic safety of high-level automatic driving vehicles.

The v2x device that meets the data format requirements of “driving environment holographic model” can not only improve the operation safety level of high-level intelligent driving vehicles, but also improve the intelligent driving ability level and operation safety of non intelligent driving or low-level intelligent driving vehicles at low cost, The benefits brought by this technical route also provide consumption power for the v2x equipment of low-level or non intelligent driving vehicles in stock, and indirectly promote the construction of the whole intelligent driving infrastructure and environment.

For high-level intelligent driving vehicles, low-level or non intelligent driving vehicles transmit the physical dimensions, driving position, speed, driving intention (steering, braking, etc.) of the vehicle to high-level intelligent driving vehicles through v2x devices installed in the rear that meet the data format requirements of “driving environment holographic model”, so as to improve the safety level of intelligent driving.

8e0cca58-b0d5-11ec-aa7f-dac502259ad0. png

Figure 5 Intelligent driving technology route based on “holographic model of driving environment” to improve the operation safety of intelligent driving

For low-level intelligent driving vehicles, the driving environment data of the intelligent driving control system can be changed from the original map, visual The sensing mode of sensors such as radar is switched to the mode of using v2x accurate data to build a “holographic model of driving environment”. Instead of replacing and upgrading more advanced vision, radar and other sensor hardware, their intelligent driving ability and operation reliability can be improved.

For non intelligent driving vehicles, the physical dimensions, driving position, speed, driving intention (steering, braking, etc.) of the vehicle are broadcast to other vehicles through the v2x equipment through the v2x device installed in the rear that meets the data format requirements of the “driving environment holographic model”. At the same time, the physical dimensions, driving position, speed, driving intention (steering, braking, etc.) of other vehicles broadcast by other v2x equipment are received, Realize low-level intelligent driving functions at low cost through software mode, such as collision warning, lane departure warning, etc.

8e1e12d6-b0d5-11ec-aa7f-dac502259ad0. png

Figure 6 Non intelligent or low-level intelligent driving vehicles improve intelligent driving ability and operation safety level through rear mounted v2x equipment

3.6 promote the realization of the business model of automobile software charging

As the intelligent driving technology route based on the “holographic model of driving environment” greatly reduces its dependence on the performance of radar, vision and other sensors, the upgrade of intelligent driving control function mainly depends on the realization of software functions, and the upgrade of vehicle firmware greatly reduces its dependence. With the continuous improvement of the infrastructure of intelligent driving environment, the intelligent driving function continues to mature, Release and upgrade the low-level to high-level intelligent driving function software in stages, making the real transformation of the automobile business model from hardware consumption to software consumption.

To sum up, it can be seen that the intelligent driving technology route based on the “driving environment holographic model” builds the “driving environment holographic model” by changing the way of driving environment data perception, broadcasting and collecting accurate driving environment data by means of communication, so as to provide reliable data input for the decision-making and control implementation of the intelligent driving control system. This method ensures the comprehensiveness and high precision of intelligent driving control system decision-making and control execution layer data input, improves the reliability and safety of intelligent driving control decision-making, promotes the development of intelligent driving control system technology, reduces the hardware and R & D cost of intelligent driving control system, and shortens the R & D and marketing cycle. At the same time, it solves the problems of the operation and operation safety of high-level intelligent driving vehicles in the long-term process of the gradual construction and improvement of intelligent driving infrastructure and the simultaneous operation with the stock of non intelligent driving vehicles and all levels of intelligent driving vehicles on the road. While improving the intelligent driving level of intelligent driving vehicles, it also improves the driving safety level of the stock of non intelligent driving vehicles and low-level intelligent driving vehicles.

4、 Key technologies of intelligent driving technology route based on “holographic model of driving environment”

4.1 road model and high-precision map

The road model of intelligent driving technology route based on “holographic model of driving environment” focuses on various element information of the road driven by vehicles, such as road types (asphalt, cement, stones, bridges, tunnels, etc.); Width of pavement; The number, width and position of lanes, and the type of lane line; Road driving conditions (speed limit, restricted traffic, etc.); Type of roadside isolation belt (rigid, flexible, no physical isolation, etc.). It does not involve the data of road surrounding facilities irrelevant to vehicle driving, such as the name, location and other geographical information of roadside buildings.

The road information of the road model is more accurate than the high-precision map. The geographic positioning information can be set as the relative position information based on the positioning base point, and it is not necessary to use the absolute high-precision longitude and latitude information. For example, the coordinate data attributes of the road model can be designed as “G3 expressway”, “uplink”, 360.26km “and” second lane “, which is very beneficial to the security of national geographic information data.

The area where the road model is released is the area where intelligent driving vehicles are allowed to run. In some sensitive areas or areas where intelligent driving is not allowed, the road model data is not released, that is, the area where intelligent driving vehicles are prohibited to run. The vehicles with intelligent driving function in these areas must be taken over and controlled by the driver, and the control system cannot collect the road geographic information of the area, It is conducive to the geographic information data security of these sensitive areas, as well as the operation safety and road traffic safety of intelligent driving vehicles in these areas.

Intelligent driving infrastructure is a process of gradual construction and improvement. In this process, intelligent driving vehicles can be opened to operate in the areas that have been completed. The construction of intelligent driving infrastructure can generate benefits at the same time. Ordinary consumers can enjoy the driving experience brought by intelligent driving in advance, so as to stimulate the market consumption demand of intelligent driving vehicles and promote the technical development and progress of intelligent driving vehicles from the perspective of the market.

In addition, the suppliers of high-precision maps are a small number of enterprises with map publishing qualifications. The road information collection, production and publishing cycle of high-precision maps is long, the cost is high, the update and maintenance are not timely, and there are hidden dangers in the safety of geographic information data; The publisher of the road model is the national road or the specialized agency of the traffic management department. The acquisition, production, release and update of road information are more timely and accurate. At the same time, the national geographic information data are controlled by the state organs rather than commercial institutions; Moreover, the release of road models can be controlled and authorized, and only the road models allowed to be released can be released, which is also very beneficial to the confidentiality and security of national geographic information data.

In some specific areas, such as the container port terminals, such as the underground garages that cannot be covered by satellite navigation signals or wireless communication signals, it can also realize the innovative application of the business scope of non high-precision map enterprises such as intelligent unmanned operation of vehicles in the port area or unmanned intelligent parking lot at low cost. The production, updating and maintenance of the road model of the port terminal or underground garage can be implemented by a qualified technical service company under the authorization of the owner.

4.2 communication mode of v2x equipment

The on-board v2x equipment sends its own vehicle geometric dimension, driving speed, position information, steering, braking and other driving intentions, faults, specific driving conditions and other non exclusive public information to the outside in a periodic and broadcast manner. At the same time, the on-board v2x equipment receives the broadcast information of other nodes (such as vehicles, roadside units, etc.) to build the “driving environment holographic model” of the vehicle.

The data transmitted by the broadcast is the non exclusive public information of the vehicle, that is, the open information used to build the driving environment model, excluding the vehicle license plate number and other personal information that may involve privacy. It is conducive to the driver’s personal information security. Broadcast rather than interactive data transmission mode is conducive to vehicle information security and function security. The vehicle only receives broadcast data in standard format, and does not receive and process data in non-standard format. Avoid tampering with vehicle information or vehicle hijacking through physical isolation.

V2x data communication is transmitted in the form of broadcast. According to the communication protocol and signal transmission power of v2x equipment, broadcast data is received and analyzed by other nodes within the range of signal transmission power (such as 500m radius) to build an “all information model of driving environment” for intelligent driving. Firstly, the coverage of the signal can ensure that the information of all elements of the driving environment required by the current intelligent driving vehicle can be received and collected; Secondly, it can increase the difficulty of illegally obtaining and tracking vehicle data and ensure the safety of vehicle information.

The data communication of v2x equipment is transmitted in the form of broadcast, and is received and analyzed by other nodes within a certain distance, so that the data communication of intelligent driving vehicles in the same specific area is not affected by the poor or shielding of satellite signals or base station transmitted signals, such as tunnels, underground garages and other special areas. Through this data communication mode of v2x equipment, Combined with the positioning technology of the vehicle inertial navigation system, it can still build an accurate “holographic model of driving environment”, so as to achieve accurate intelligent driving.

8e3b9126-b0d5-11ec-aa7f-dac502259ad0. png

Figure 7 The broadcast data of v2x equipment in a specific area is used to build a “holographic model of driving environment”

On board v2x equipment broadcasts non exclusive public information such as its own vehicle’s geometric dimension, driving speed, position information, steering, braking and other driving intentions, faults, specific driving conditions, and in some accidental or temporary abnormal traffic conditions, such as temporary traffic control, safety inspection of drivers and passengers, or when special vehicles such as fire vehicles or ambulances are operating, or vehicle faults In the case of road accidents, the safety fence or avoidance area can be quickly set up through the built-in vehicle data under special conditions in the system.

8e54f7ba-b0d5-11ec-aa7f-dac502259ad0. png

Figure 8 V2x equipment sets security fence by broadcasting specific data

V2x data communication is transmitted in the form of broadcast. The broadcast data is received and analyzed by other nodes within a limited distance range. Generally, within this distance range, these data are only received and used to meet the control requirements of intelligent driving vehicles. However, some data under special circumstances, such as the safety zone set under the above special circumstances, even include some data that v2x cannot express, It is the driving environment data identified and identified by the vision or radar sensing system of intelligent driving vehicles, such as road scattering objects. Through the v2x communication protocol and message format design, the vehicle that first identifies the data is sent out through the v2x device broadcast, received and forwarded by other nodes, providing earlier or longer warning time and distance for other intelligent driving vehicles.

5、 Summary

The intelligent driving control system is a complex large-scale system. Based on the intelligent driving technology route of “driving environment holographic model”, it relies on v2x and other devices to transmit accurate driving environment data to build an accurate “driving environment holographic model”, which provides accurate decision-making basis for the intelligent driving control system, and solves the problem that the traditional driving environment perception layer is subject to sensor technology, environmental interference, perception accuracy, recognition algorithm The uncertainty and inaccuracy of driving environment data caused by the restriction of data fusion technology and other factors lead to the unreliability and unreliability of intelligent driving decision-making, which promotes the development, progress, realization and implementation of high-level intelligent driving system technology.

In addition, the “driving environment holographic model” is an ecological platform, which fully takes into account the concerns of all parties involved in the intelligent driving environment (technical environment and social environment). The participants in the intelligent driving environment include the technology R & D department for the development, testing and verification of the intelligent driving system, the system parts supplier, the vehicle manufacturer, the construction management department of the intelligent driving infrastructure, and the traffic management department, Users of intelligent driving vehicles, coexisting non intelligent driving vehicles and other vehicles, etc. The rights and interests of all parties involved in the intelligent driving environment can be reflected in a relatively balanced way on the ecological platform, and appropriate solutions can be found for their concerns. Therefore, the participants of the ecological platform can give full play to their respective technologies and capabilities, and contribute to the construction, development and progress of the intelligent driving environment.

Reviewed by: Li Qian

Leave a Reply

Your email address will not be published. Required fields are marked *