Abstract:The infrastructure-vehicle cooperative autonomous driving approach depends on the cooperation between intelligent roads and intelligent vehicles. This approach is not only safer but also more economical compared to the traditional on-vehicle-only autonomous driving approach. In this paper, we introduce our real-world deployment experiences of cooperative autonomous driving, and delve into the details of new challenges and opportunities. Specifically, based on our progress towards commercial deployment, we follow a three-stage development roadmap of the cooperative autonomous driving approach:infrastructure-augmented autonomous driving (IAAD), infrastructure-guided autonomous driving (IGAD), and infrastructure-planned autonomous driving (IPAD).
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to improve the safety, efficiency and economy of autonomous driving through the cooperation between infrastructure and vehicles (i.e., vehicle - road cooperation) in the field of autonomous driving. Specifically:
1. **Safety**: The paper points out that 1.35 million people die in traffic accidents globally every year, and another 20 million people suffer non - fatal injuries. Although the accident rate of autonomous vehicles is already lower than that of human drivers, there are still some blind spots that cannot be completely avoided either by human driving or by existing autonomous driving technologies. Through vehicle - road cooperation, these blind spots can be eliminated, the perception distance of vehicles can be extended, thus more reaction time can be provided and the occurrence of accidents can be reduced. The paper mentions that in actual deployment, through vehicle - road cooperation, the disengagement rate of autonomous vehicles has been reduced by more than 90%.
2. **Efficiency**: Traffic congestion brings huge costs to society. For example, the US trucking industry loses about $74.1 billion annually due to congestion, most of which occurs in urban areas. In addition, there are other costs such as pollution and accidents. Through intelligent roadside systems (SoRs), comprehensive and detailed traffic condition information can be obtained, thereby improving traffic efficiency. The paper mentions that at their deployment sites, through intelligent traffic signal control, traffic efficiency has been increased by 40%.
3. **Cost**: Currently, there are approximately 1.2 billion cars on the roads globally, and it is expected to increase to 2 billion by 2035. However, road capacity is limited, and the construction cost of autonomous vehicles is still high. Therefore, investing in intelligent infrastructure and gradually transferring more computing tasks from vehicles to roadside systems can significantly reduce the hardware and energy costs of autonomous vehicles. The goal of the paper is to finally enable autonomous vehicles to only have basic perception and control capabilities, while most of the active perception and planning tasks are completed by the roadside systems. It is expected that this will reduce the cost of autonomous vehicles by more than 50%.
To achieve these goals, the paper proposes a three - stage development roadmap:
- **Stage 1 (IAAD, Infrastructure - Augmented Autonomous Driving)**: Autonomous vehicles integrate the perception results from the vehicle side and the infrastructure side to improve safety and achieve maximum computational and energy efficiency.
- **Stage 2 (IGAD, Infrastructure - Guided Autonomous Driving)**: Autonomous vehicles can offload all active perception tasks to the infrastructure to reduce the deployment cost of each vehicle. The focus of this stage is to design safe, reliable and efficient infrastructure support.
- **Stage 3 (IPAD, Infrastructure - Planned Autonomous Driving)**: The infrastructure is responsible for perception and planning, thereby achieving maximum traffic efficiency and cost efficiency. Autonomous vehicles at this stage only need to have basic perception and computing capabilities, and the central system will plan the optimal path for each vehicle.
The paper also discusses some challenges faced in the process of achieving these goals, including issues in network performance and security, reliability and efficiency.