Autonomous Driving (L2+) model training

Autonomous Driving (L2+) model training

Lane Change

Teraki lane change monitors the vehicles current lane along with traffic situation and provides real-time feedback to the driver or the autonomous driving stack for action. The predictive alert and the continuous monitoring effectively delivers better safety and improved attention to the driver and helps develop the ADAS/AD stack as well.

Driver monitoring

Drive monitoring consists of models that detect behaviour such as hard manoeuvres, driver distraction, driver drowsiness, in-car smartphone usage, etc.

Traffic Sign

Recognizing traffic signs and their various statuses (red, amber, green) is part of this model. The model can be easily trained to function better under specific circumstances customer wants to train the model for; e.g. geographical differences, rainy or foggy conditions, glare, etc.

General Teraki model training

An AI-model is applied at the edge of the car to provide real-time alerts as one or more of these events have been detected. From the sensors data the AI-models accurately detect events and can provide real-time alerts in driver assistance systems or can be used in further analytics. A continuous influx of the specific data relevant to the AI-model helps to retrain these AI-models to achieve higher accuracy.

Teraki Platform automates this process of smart sensor data selection and ingestion. Teraki SDK requires less processing power and hence lower the OEM’s bill of material and saves energy at the car, drone or robot.

Above use cases will lower the costs of insurane premiums; make the roads safer and can save lives.

How to Implement with Teraki:

ROI-models (“selectors”) and Lane Change, Driver Monitoring and Traffic Sign TOI-models are available on the Platform: