Glossario IA
Il dizionario completo dell'Intelligenza Artificiale
ADAS (Advanced Driver Assistance Systems)
Set of electronic technologies assisting the driver in driving and parking, using sensors and AI algorithms to improve safety and comfort.
ECU (Electronic Control Unit)
Electronic control unit embedded in the vehicle, executing AI algorithms in real-time for critical functions such as autonomous braking and obstacle detection.
Sensor Fusion
Artificial intelligence process combining data from multiple sensors (cameras, lidar, radar) to create a complete and reliable environmental perception of the vehicle.
Embedded Neural Networks
Neural networks optimized and deployed directly on the vehicle's embedded chips to execute AI tasks with minimal latency and without cloud connection.
Automotive AI Accelerators
Specialized processors designed to accelerate AI inference calculations in vehicles, enabling real-time processing of autonomous driving data.
V2X Communication
Vehicle-to-Everything (V2X) communication technology allowing vehicles to exchange information with their environment via AI-enhanced protocols.
Fail-safe Systems
Redundant AI-based safety mechanisms ensuring safe vehicle operation in case of failure of main autonomous driving systems.
Driver Monitoring System (DMS)
Computer vision system analyzing the driver's state (fatigue, distraction) in real-time through infrared cameras and deep learning algorithms.
Predictive Maintenance
Onboard AI application analyzing vehicle sensor data to predict failures and optimize preventive maintenance of critical components.
Automotive SoC (System on Chip)
Specialized integrated circuit combining processor, GPU, and AI accelerators to execute complex autonomous driving workloads in real time.
Object Detection Network
Optimized convolutional neural network architecture to identify and locate pedestrians, vehicles, and obstacles in the driving environment.
Lane Keeping Assistance
AI system using computer vision to detect road markings and automatically keep the vehicle in its lane.
Real-time SLAM
Simultaneous Localization and Mapping algorithm executed in real time on onboard computers to build and update the vehicle's environment map.
Over-the-Air AI Updates
Capability to update vehicle's AI models and embedded algorithms remotely, enabling continuous improvement of autonomous features.
CAN Bus AI Integration
Integration of artificial intelligence algorithms with the vehicle's Controller Area Network to optimize communication between different ECUs.
Behavioral Cloning
Supervised learning technique where AI learns driving behaviors by imitating human driver actions from recorded data.
Edge-to-Cloud AI Pipeline
Distributed processing architecture where critical tasks are executed in edge computing while model training and advanced analysis are done in the cloud.