Researchers from the University of California, Berkeley, and the University of Michigan have developed a new scheduling algorithm aimed at improving collaborative perception (CP) systems, which are crucial for applications like autonomous driving and smart cities. Collaborative perception involves multiple sensors sharing and fusing information to overcome individual limitations such as blind spots and range restrictions. The study, titled “Timeliness-Oriented Scheduling and Resource Allocation in Multi-Region Collaborative Perception,” was published in the IEEE Transactions on Mobile Computing.
The researchers identified two primary challenges in CP systems: the timeliness of transmitted information and the efficient use of limited computational and wireless bandwidth resources. To address these issues, they developed the Timeliness-Aware Multi-region Prioritized (TAMP) scheduling algorithm. This algorithm aims to balance perception accuracy and communication resource usage by considering the utility of information, which decays over time, and the impact of the Age of Information (AoI) and communication volume on perception performance.
The TAMP algorithm is designed to minimize a timeliness-oriented penalty in the long term, recognizing that scheduling decisions have a cumulative effect on subsequent system states. It uses a Lyapunov-based optimization policy to decompose the long-term average objective into a per-slot prioritization problem, balancing the scheduling worth against resource cost. The researchers validated the algorithm in both intersection and corridor scenarios using the real-world Roadside Cooperative perception (RCooper) dataset. Extensive simulations demonstrated that TAMP outperformed the best-performing baseline, achieving an Average Precision (AP) improvement of up to 27% across various configurations.
For the energy sector, this research could have practical applications in improving the efficiency and reliability of sensor networks used in smart grids and other energy infrastructure. By optimizing the scheduling and resource allocation of sensor data, energy companies could enhance their ability to monitor and manage energy distribution systems, leading to more efficient and reliable energy delivery. Additionally, the timeliness-oriented approach could be valuable in emergency response scenarios, where quick and accurate perception of system states is critical.
Source: IEEE Transactions on Mobile Computing
This article is based on research available at arXiv.