A simulation scenario for the self-driving route in Changhua Coastal Industrial Park has been established. The roads, markings and signs according to the route have been set and calibrated with high-precision maps. The virtual scenario mileage has accumulated up to 12km. A WinBus vehicle dynamic model is also included in the environment, which integrate the self-driving car control strategy for tracking control verification. Besides, traffic conditions can be added to the simulation environment to verify decision-making control; the influence of different weather condition on the detection ability of the sensor is also considered, and the abnormal or even failure of the sensor signal is simulated to verify the resilience of the self-driving car. In terms of actual vehicle verification, based on the self-driving test draft, taking into account Taiwan’s road usage habits, adding common violations and emergencies, established the first version of the basic functional test standard for self-driving cars, and conducting basic functional tests in the closed ARTC field with WinBus. The standard will continue to be modified in accordance with the pilot run results.

1.Scenario and sensor model (PreScan), Vehicle dynamics model (Carsim), Traffic model (Vissim)
2.Real-time simulating platform, Six degree of freedom motion platform
3.Steering, Brake, and Accelerating control devices
4.RTK DGPS, Vehicle performance measuring devices

This technique can be applied in V&V of passenger or commercial vehicles for advanced driver assistance system and autonomous driving system.

More and more manufactures start to develop ADAS and ADS, but most of them lack of a comprehensive environment to verify the control algorithm. The establishment can assist system suppliers to verify system functions for problem shooting and algorithm improving. Not only can the competitiveness be strengthened, but also the development cycle and testing risk could be reduced.