AStriD - Autonomous streetcar in the depot

Problem definition

Due to the operational environment - no public access, low speeds, regulated processes - autonomous streetcars could already increase efficiency in the depot, reduce costs and increase availability. For implementation, it is necessary to answer open research questions regarding the sensor and localization technologies to be used, the data exchange between the systems involved and the approval and legal framework conditions.

Project goals

A fully automated streetcar depot based on an autonomously driving streetcar and a digital depot is to be designed using the example of the Potsdam transport company and the technical feasibility demonstrated as a prototype with autonomous service trips (e.g. through a washing facility to a siding). Based on the experience gained from the implementation, the further scientific/technical/legal need for action - with regard to sensor technology, data quality and availability, data access and the data interfaces to the systems involved - is to be identified.

Implementation

A concept for a digital depot with a focus on data networking of all components involved in automation (control software, systems and vehicles) will be developed. The feasibility will be demonstrated using selected autonomous service journeys by implementing the data link between the depot systems, the digital map and an intelligent streetcar via a data hub. The streetcar will be equipped with a ticket machine and a sensor-based environment detection system.

KIT's contribution

KIT is responsible for analyzing the existing processes in the depot and deriving the use cases based on them. The target state of the digital depot will also be specified. With these findings, KIT will contribute to the creation of the digital map in order to enable automated travel in the depot. The expertise in pre-processing the sensor data, the creation of evaluation metrics for the methodical selection of a suitable procedure for AI-based sensor fusion and the determination of optimization approaches will also be incorporated. This enables precise 3D modeling of the routes and the environment at the depot.
Finally, the knowledge gained from the concept is abstracted for transfer to other depots.