COGDRIVE

DPI2017-86915-C3-3-R 2018-2020

Goal: The main goal of the project is to design, develop and validate on real vehicles a cognitive-inspired decision system for autonomous vehicles, with the ability to emulate human intuition in different driving environments.

Vision: Architecture for cognitive-inspired decision making, with three ties: a “dorsal stream”, which represents inverse models nested in layers that generate “possibilities” from the sensory input; a mechanism for selecting actions (“basal ganglia”) that operates at various levels of the hierarchy, and a “cerebellum” that learns advanced models.

Total investment: € 209,000.00

Role of AUTOPIA

Demonstrators:

Specialized autonomous driving simulator (TRL 4). Robustness of the components developed under variations of scenario, environment, vehicle, etc.
Mobile robots in controlled environments (TRL 5), to validate cross-cutting technologies (location, interaction, modeling) on relevant scenarios.
Automated vehicles in segregated environments (TRL 6), in which tests can be repeatable, and reliability analysis can be carried out correctly.
Automated vehicles on roads open to traffic (TRL 7), combining the urban environment of the Arganda del Rey test tracks, and the surrounding peri-urban roads.

Partners

1. CONSEJO SUPERIOR DE INVESTIGACIONES CIENTÍFICAS (CSIC)

2. UNIVERSIDAD POLITÉCNICA DE MADRID (UPM)