In the fast-paced world of autonomous and connected vehicles, the SELFY Vehicle Situational Awareness Tool (VSAT) stands out as a pivotal innovation. VSAT is a centralized, real-time software component that integrates and stores data from various sensors and sources, including Vehicle-to-Everything (V2X) communication, creating a structured georeferenced local database. This data fusion process enables the VSAT to serve as a common layer accessed by all applications within the vehicle.
VSAT excels in integrating heterogeneous data sources, providing driver-friendly visual traffic information using detections in the ego-car coordinates. This information is input into a database that agnostic applications can access to develop a multitude of use cases. Each vehicle, both transmitter and receiver, uses the VSAT as a centralized database for managing data, focusing on both dynamic and static information.
Our VSAT implementation leverages InfluxDB, an open-source time-series database optimized for fast writes and queries, ideal for real-time data management needs. This setup allows for real-time visualization of accessible data via tools like RVIZ or open-source dashboard and analytics solutions such as Grafana.
By integrating data from diverse sources, the vehicle’s perception system creates a “360-degree situational awareness,” generating a local dynamic map that accurately positions detected elements and road users in a 3D space.
The VSAT tool enables the ego vehicle and other cooperative systems to make informed decisions in real-time, adapting to various traffic scenarios.
In this tool, developed as part of SELFY’s Situational Awareness and Cooperative Platform, each output and intermediate data point generated within the architecture is stored along with the vehicle’s GPS position. This robust database supports the development of Advanced Driver Assistance Systems (ADAS) applications, including real-time visualizations of detections from each source using Grafana.
The tool also emphasizes a layered structure inspired by the SAFESPOT framework, comprising four layers:
- Static Layer: static map information, including road topology.
- Quasi-static Layer: semi-permanent information like traffic signals and infrastructure.
- Transient Layer: dynamic transient information such as traffic light phases, traffic congestion, and parking statuses.
- Dynamic Layer: highly dynamic data, including vehicles and pedestrians.
Our focus is on the static layer and the dynamic layer. For the static layer, we utilize OpenStreetMaps for road topology data, while the dynamic layer integrates data from multiple sources, including the ego vehicle’s GPS, perception module detections, and V2X communication data. This data is inserted into InfluxDB at a high frequency, ensuring timely updates and relevance.
The VSAT tool continuously accesses the ego vehicle’s position from DGPS to transform local coordinate objects into global coordinates, ensuring georeferenced detections from the perception module and cooperative perception messages. These real-time data queries can be visualized in Grafana, providing insightful dashboards for positioning, detections, and V2X messaging.
In conclusion, the Vehicle Situational Awareness Tool can be an integral component of connected autonomous vehicles. It ensures efficient data fusion, real-time updates, and a comprehensive situational awareness that enhances decision-making processes and adapts to evolving traffic conditions.
Author: Ficosa ADAS
