KIVA-NET

AI-based Traffic Analysis with Networked Acoustic Sensors
(ZIM, 2025-2027)

Urban traffic areas are increasingly suffering from rising traffic volumes, which has a negative impact on infrastructure and the environment. A promising solution is the digitization of traffic infrastructure, which must be implemented quickly, cost-effectively, energy-efficiently, and robustly. Acoustic sensors meet these requirements, but until now could not be used effectively due to technical hurdles. Thanks to advances in AI, real-time analysis of acoustic signals is now possible.

As part of the KIVA-NET research project, a transferable demonstrator for an AI-supported acoustic sensor system is to be developed in the V2X test field in Magdeburg. With the help of state-of-the-art AI algorithms, road users are to be counted and classified, and the degree of utilization of the infrastructure determined. To increase scalability and energy efficiency, a hybrid computing architecture will be implemented that flexibly distributes calculations between edge devices and the cloud. This innovative sensor infrastructure forms the basis for intelligent, adaptive, and efficient traffic control to reduce traffic congestion and emissions.

Project Partners