Name: Mao Zhi; Student ID: 21021110040;
College: School of Electronic Engineering.
?Reprinted from /html/itswiki/library/2020_05_110167.html
Introduction to embedded cattle This article mainly introduces the application scenario of intelligent transportation with integrated communication and perception.
Chiniubi 6G? Intelligent transportation
Chiniu asked what is intelligent transportation? What functions can intelligent transportation achieve that cannot be achieved now? What is the current level?
Embedded text
1. Development background
(1) Brief description of the development of intelligent transportation systems and Internet of Vehicles
Intelligent transportation systems (ITS) is a general term for the integrated application of communication, control and information processing technologies in transportation systems. It is an integrated transportation system that ensures safety, improves efficiency, improves the environment and saves energy through the close cooperation of people, vehicles and roads.
The development of my country's intelligent transportation system is divided into three stages: the initial stage (before 2000), the substantive construction stage (2000-2005), and the rapid development stage (2005 to the present). In the initial stage, basic research on urban traffic signal control was mainly carried out, and demonstration sites such as electronic toll collection systems and traffic management systems were further established, bringing the intelligent transportation system into the stage of promotion, application and improvement, but the overall level lagged behind. In the substantive construction stage, the state invested a large amount of money in the research and development, production and popularization of ITS, creating favorable conditions for the development of ITS. In the rapid development stage, with the rapid development of artificial intelligence, autonomous driving, Internet of Vehicles and other technologies, with the goal of building "smart cities", "green cities" and "safe cities", my country's ITS technology has been further developed and expanded application.
In recent years, emerging technologies represented by autonomous driving have developed rapidly and have become one of the indispensable key technologies in future intelligent transportation systems. The Society of Automotive Engineers (SAE) divides autonomous driving into six levels from 0 to 5***. The higher the level, the higher the degree of autonomous driving. In order to improve the safety of autonomous vehicles, vehicles are usually equipped with a variety of sensors, such as optical cameras, ultrasonic radars, millimeter wave radars, and laser radars, etc., to improve the bicycle's environmental perception capabilities and contribute to the vehicle's travel control and safety. Driving prediction and other operations. In addition, the development of technologies such as 5G Internet of Vehicles has also provided a variety of communication technology means for intelligent collaboration between vehicles, assisting the development of autonomous driving technology.
2. Why Intelligent Transportation is Needed
In recent years, major car companies and research institutes around the world have equipped vehicles with a variety of sensors to enhance their environmental perception capabilities and collect road condition data. , and use algorithms such as machine learning to combine offline learning and online decision-making to achieve the goal of improving the safety and reliability of autonomous driving. However, because vehicle sensors (such as radar, optical cameras) are easily affected by obstacles, rain, snow, strong and weak light and other factors, the environmental information perception capabilities based on bicycle sensors are limited, and vehicle collisions and accidents are prone to occur. Autonomous driving accidents caused by object recognition failures. Therefore, there is an urgent need to enhance beyond-line-of-sight sensing capabilities through intelligent vehicle connectivity technology, break through the technical bottleneck of limited environmental sensing capabilities of single-vehicle sensors, and improve the safety and reliability of autonomous driving. At the same time, in order to meet the low-latency and high-rate transmission requirements for workshop information sharing for beyond-line-of-sight sensing, this paper proposes a perception and communication integration based on dynamic sharing of time-domain resources in the millimeter wave band. Intelligent vehicle-linked transmission system to ensure reliable broadband sharing of workshop sensing information. Based on the business priorities of sensing and communication, dynamic time allocation and flexible beam control algorithms are designed to optimize the overall performance of the integrated sensing and communication system. Designed and developed a verification platform for integrated sensing and communication intelligent vehicle-linked systems based on millimeter wave technology, achieving principle verification of core functions and key technologies.
3. Challenges faced by the design of intelligent vehicle-connected systems integrating perception and communication
In order to improve the beyond-line-of-sight environmental perception capabilities of autonomous vehicles, the fusion of sensory information is achieved through multi-vehicle collaboration. is one of the ways to achieve it. In order to overcome the current problems faced by multi-vehicle sensor information fusion such as different information formats and low fusion efficiency, it is urgent to improve the intelligent level of information fusion between vehicles through the joint design of sensing and communication systems to ensure the ability of autonomous driving to perceive the environment. Improve the requirements for integration with information timeliness. The following introduces the typical application scenarios of intelligent connected car systems for beyond-line-of-sight sensing and the challenges faced by the integrated design of sensing and communication.
(1) Typical application scenarios of intelligent connected vehicle systems for beyond-line-of-sight sensing
Figure 1 shows typical application scenarios of multi-vehicle collaborative intelligent connected systems for beyond-line-of-sight sensing . Among them, vehicles B, D and E are vehicle targets that can be directly detected by the radar sensor of vehicle A. However, due to the obstruction by vehicles B and D in front, vehicle A's perception range is greatly limited, causing vehicles C and F to be in vehicle A's blind spot.
Therefore, in order to expand the detection distance and range of vehicle A's radar sensor, millimeter wave broadband transmission technology is used to transmit the radar sensing information of vehicles B and D back to vehicle A, and vehicle A performs multi-source information fusion to improve vehicle A. beyond line-of-sight sensing capabilities, thereby improving the safety and reliability of intelligent car-linked systems.
(2) Challenges faced by the design of intelligent vehicle-connected systems integrating perception and communication
Although the intelligent vehicle-connected system using multi-vehicle collaboration can improve the ability to perceive environmental information, there are many challenges in the perception and communication The integrated design of the system faces many challenges. First, the signal forms, signal processing mechanisms, and system performance evaluation parameters of sensing and communication are different. Therefore, how to design an effective system performance evaluation method is crucial. Secondly, the fusion of various delay-sensitive sensing data between vehicles is limited by a variety of software and differentiated hardware platforms. How to achieve rapid fusion of sensing information to meet the requirements of low-latency and high-reliability information transmission is also an integrated system. One of the problems faced by design. Finally, for high-mobility Internet of Vehicles scenarios, how to achieve fast beam alignment and beam tracking for millimeter-wave broadband communications is another difficult problem to ensure the reliability of broadband transmission of sensing data.
4. Design Framework of Intelligent Vehicle Connected System with Integrated Perception and Communication
(1) Design Framework of Integrated Perception and Communication System
In view of the challenges faced in the design of intelligent vehicle interconnection system To address the challenges of high-rate, low-latency sensing information fusion, this paper proposes an intelligent vehicle-connected system framework (see Figure 2) based on the integrated design of perception and communication to achieve the goal of multi-vehicle collaborative over-the-horizon sensing.
First of all, the environmental information obtained by the vehicle through multiple sensors has different priorities. The delay-sensitive sensing information is divided into high-priority data and low-priority data, and is separated through communication technologies with different capabilities. Make the transfer. For example: high-priority data has higher requirements on delay and data rate, and can be transmitted through the vehicle-to-vehicle direct broadband link; low-priority data has lower requirements on delay and data rate, and can be transmitted through the vehicle. Transport on low to medium speed links to the infrastructure. In addition, the frame length ratio of the sensing and communication systems can be dynamically and flexibly configured based on the different delay sensitivities of sensing information, and a dynamically adjustable frame structure method of time slots in the sensing and communication integrated system is proposed. In addition to the time slots used for control signaling transmission in the frame structure, short subframes are used for information with high delay sensitivity, and long subframes are used for information with low delay sensitivity. The information transmission process can also be based on the service. It is necessary to dynamically configure the subframe length, and combined with the characteristics of millimeter wave technology used in workshop communication, a millimeter wave beam rapid alignment and tracking technology is proposed to optimize the beam search space dimension and algorithm complexity to meet the fast and reliable requirement of delay-sensitive information. transmission requirements.
(2) Evaluation indicators of integrated perception and communication systems
In view of the many challenges faced by the design of integrated perception and communication systems, in order to effectively evaluate the performance of the designed integrated system, it is urgently needed Performance indicators that can scientifically analyze and measure the performance improvement and overhead brought about by the integration of the two systems of perception and communication. The traditional evaluation method for the integration of two systems is to convert the performance indicators of one system into the indicators of the other system. Considering the diversity of sensing information types and methods, taking radar sensing data as an example, the radar information estimation rate can be expressed by the entropy of random parameters and the entropy of radar estimation uncertainty, which can be analogous to the data rate of communication systems based on information entropy. representation theory. In addition, a variant form of the communication metric based on the minimum mean square error can convert the communication metric into an effective metric similar to the radar estimated Cramero bound form. Therefore, the unified measurement and evaluation indicators in the integrated sensing and communication system are indispensable key indicators for the performance evaluation of the integration of the two systems. It is necessary to consider the multiple functions of the integrated system for joint design.
(3) Task-driven dynamic time slot allocation frame structure
Different sensing information transmission methods depend to a large extent on the delay sensitivity and priority of the service. For example, car collision and road safety hazard alarms are emergency sensing information and have high priority for autonomous vehicles. Low-latency information transmission requirements are ensured through millimeter wave link transmission between vehicles. On the other hand, low-priority information such as traffic jams, optimal route planning, and entertainment videos can be transmitted through vehicle-to-infrastructure communication links because their priority is lower than emergency sensing information. At the same time, by using a dynamic frame structure based on flexible time slot allocation, it can meet different priority, delay-sensitive and non-sensitive business needs, ensuring low delay and highly reliable data transmission. To this end, this article proposes a new frame structure based on 5G new air interface for the integrated intelligent vehicle connection system of sensing and notification. It provides a flexible frame structure configuration method to realize flexible and dynamic configuration of on-demand time slots for sensing and communication functions. The frame structure The design is shown in Figure 2. In addition, considering the need to transmit emergency information (such as traffic accidents, pedestrians crossing the road and other emergencies), this paper also designs a dynamic subframe time slot configuration method based on tiny sub-slots to ensure low-latency information transmission.
(4) Resource allocation method based on game theory
In order to make the transmission time occupied by sensing and communication in the sensing and communication integrated system according to the needs of delay-sensitive services Perform dynamic adjustment and divide the frame structure into dynamically changing sensing function subframes and communication function subframes, as shown in Figure 2. Taking radar sensing as an example, for one frame, if the radar detection duration is longer, the communication transmission duration will be shorter; on the other hand, the time occupied by the two different functions of sensing and communication is closely related to its performance. In addition, the information sensed by the radar needs to be transmitted effectively within the subsequent communication transmission time as much as possible, otherwise the timeliness of the radar sensed information will be lost. Therefore, there is a mutually restrictive relationship between radar duration and communication transmission time. Non-cooperative game theory and methods can be used to optimize the allocation of time resources and optimize the performance of the integrated sensing and communication system. To allocate time resources for a time-division-based perception and communication integrated system, it is necessary to optimize the amount of radar information under different radar duration ratios and find the optimal radar under the constraint that the amount of radar information is not greater than the amount of communication information. The communication duration ratio achieves joint optimization of the amount of radar and communication transmission information, improving the vehicle's environmental perception performance and the transmission and fusion performance of multi-vehicle perception data.
(5) Flexible beam control method based on reinforcement learning
In order to solve the problem of effective transmission and fusion of high-bandwidth and large-traffic sensing information between vehicles, this paper proposes that millimeter waves can be used to Beam control methods and technologies enable reliable information transmission in the workshop. Millimeter wave communication technology uses large-scale phased array antennas and beamforming technology to enhance signal strength at the receiver to overcome signal loss and attenuation problems. In addition, the radar sensing information in the previous time slot can be used to assist the beam alignment and beam tracking process between vehicles, effectively reducing the time overhead of beam control. During the beam alignment process, the positional relationship between vehicles can be obtained from radar sensing information. Utilizing the sensing information can minimize the beam search space and effectively reduce the beam alignment time. Based on the vehicle speed and trajectory contained in the radar sensing information, a beam tracking algorithm based on reinforcement learning is designed to achieve rapid beam switching in vehicle movement scenarios and ensure the reliability of communication links and link connection stability between vehicles. .
4. Perception and communication integrated intelligent vehicle connectivity system verification platform
Because short-range radar and medium-range radar systems operating in the 20~30GHz millimeter wave band have been used in vehicle collision avoidance and blind spots It is widely used in detection. To this end, this paper designs and builds an integrated perception and communication verification platform working in the 26 GHz millimeter wave frequency band. By aggregating eight 100 MHz carrier frequency bands, an 800 MHz broadband millimeter wave communication bandwidth is obtained to verify the proposed perception and communication integration. The core functions and key technologies of the system. A millimeter-wave phased array antenna with 64 array elements is used at both the transmit and receive ends of the platform to verify the performance and feasibility of fast beam alignment and beam tracking algorithms, as shown in Figure 3.
The results of the integrated sensing and communication test platform are shown in Figure 4. The radar sensing results of vehicle A indicate that vehicles B, D and E exist at a distance of 18 meters, 14 meters and 30 meters from vehicle A respectively. , but because the vehicle blocks the radar detection signal, vehicles C and F are in the blind spot of vehicle A. Vehicles B and D sense the position information of vehicles C and F respectively through their respective radars, and share the information of vehicles B and D with vehicle A*** through a millimeter wave broadband communication link. Finally, by integrating the radar sensing information from vehicles B and D, vehicle A's environmental sensing capabilities are improved and beyond-visual-range sensing is achieved.
5. Future Technology Outlook
As artificial intelligence technology continues to mature and be widely used, technologies such as autonomous driving, multi-vehicle collaborative virtual reality, and augmented reality information fusion are expected to expand The bicycle senses the field of vision, improves the bicycle's perception ability, and improves the safety and intelligence level of the connected vehicle system. Taking into account the rapid development momentum of autonomous vehicles, the integrated perception and communication intelligent vehicle connection system for beyond-line-of-sight sensing will comprehensively break through the bottleneck of single-vehicle perception capabilities and improve the intelligent vehicle connection system through multi-vehicle collaboration, perception and communication system integration, etc. safety and reliability, and will become a research hotspot in this field in the next five to ten years.