The intention-aware system of multi-objections uses a monocular camera to capture image information in front of vehicle. The system applies multi-objections extraction feature method, tracking technology, and smoothing operation to transform images to stably track the features of multi-objections. Moreover, using our intention-aware technology predicts possible positions and speeds for multi-objections in following seconds. The host vehicle can receive real-time warnings in advance in dangerous situations based on our system.

1.Image processing rate: Above 10 frame/sec.
2.Image resolution: 1920x1080
3.Detected Objections: car, motorcycle/bike, pedestrian
4.Operating computer：IPC with WebCam/Nvidia Drive PX2 with GMSL camera
5.Interface: CANBUS including speed, indicator and object position
6.Operating temperature: -40°C~85°C.
7.Operation environment：both daytime and nighttime (above 0.1 lux.) without raining and snowing.
8.Operating current：(12V, 20A)/IPC, (12V, 25A）/Nvidia Drive PX2.

Intelligent Vehicle Safety System

The Intention-aware system predicting multi-objections can effectively reduce the occurrence of dangerous driving. With the Intention-aware system, the driver can figure out in advance whether there exists risk in front of the host car, which has obvious benefits for driving safety. Besides, our system helps the domestic automotive industry level up to high-market-value active safety field, and the market benefit will be increased.