AI Perception Solution

The AI perception solution, named as OAD (Object Detection, Action Prediction, and Distance Estimation), utilizes the RGB and thermal images to detect common objects on the road.

This solution is primarily based on deep learning algorithm for object detection, combined with  object tracking and action prediction to obtain object information, such as category, bounding box, action, and distance.

Thermal Sensing Solutions :

• AEB automatic emergency braking
• FCW forward collision warning
• Level 2 and 2.5 hands-free driving
• Level 3 eyes off hands off system
• Level 4 Full self-driving cars

Compliant with FMVSS127

Key solution for L2-L4 Autonomous Driving

Compared to other solutions, our solution can address challenging situations, such as glare and  low-light conditions, providing more stable and accurate object detection, enhancing driving safty. It can be applied to ADASs such as AEB(Autonomous Emergency Braking) and  PAEB(Pedestrian-AEB).

Object Detection & Distance Estimation

The object detection and distance estimation algorithm are specifically designed for front view ADAS (Advanced Driver Assistance Systems). By processing images with deep learning models, the algorithm can obtain relevant information about target objects, such as their category, bounding box, and distance, so as to prevent potential accidents.

Currently, the algorithm can recorgnize object categories such as pedestrians, motorcycle, car, bike, truck, and bus.

AD (Action Detection) Behavior Detection

This system utilizes deep learning models to identify actions of known objects, determining whether they will invade the main lane (ego-vehicle's driving lane) in 2-seconds ahead. Once the system detects an object will invade main lane, it issues a warning signal to driver thereby  enhancing road safety.

The system used IMU (Inertial Measurement Unit) and GPS to predict ego-vehicle's movement trajectory. Through ego-vehicle trajectory and road object's action, the system assigns different levels of risk to the road objects,

• Green: Safe
• Yellow: Caution
• Red: Dangerous

Model Deployment