Biological Neural Network Models TRIZ Cases
- Vehicle Driving Control Using Object Passability Classification
Front-field sensor and camera data classify known and unknown objects by passability, enabling safer vehicle maneuver control.
- Parking Assistance ROI Classification for Special Slot Detection
Camera ROI analysis links map-designated parking areas with ground symbols to filter special slots and no-parking zones.
- Lidar Background Imaging for More Reliable Range Detection
Background light intensity images help correct weather and reflection errors, adding more reliable range data to lidar scans.
- Road User Behavior Prediction With Lane-Segment Neural Encoding
Lane-segment and historical state encoding improves near-future road user intention and trajectory prediction for autonomous driving.
- Image-Level Road Element Representation for Real-Time Driving Interaction
Transformer tokenization of road elements improves real-time driving interaction detection while reducing storage and processing demands.
- Driving Scene Token Selection for Efficient Road Element Prediction
Transformer-based token selection focuses on relevant road element attributes by scenario, improving driving prediction accuracy with lower compute use.