Peripheral Object Simulation Control Using External Trigger Inputs

The patent uses neural network models to copy and learn from real human driving behavior data. Instead of programming explicit rules, the system trains neural networks on datasets of actual human driving patterns, enabling the simulation to reproduce irregular human behaviors such as restarting after stopping attempts or returning to lanes during lane changes, thereby resolving the contradiction between accuracy and complexity

The patent replaces traditional mechanical rule-based control systems with neural network-based intelligent systems. The neural networks process sensor inputs and generate driving commands through learned patterns rather than predefined rules, substituting the rigid mechanical approach with a flexible intelligent approach that can handle irregular human driving behaviors

The patent creates virtual copies of real-world driving environments, vehicles, and scenarios in a simulation system. By copying real driving conditions into a virtual space, the system enables safety testing without the time, cost, and risks associated with physical test drives, while maintaining the realism needed for accurate safety validation

The patent performs preliminary safety testing and validation in the virtual simulation environment before deploying autonomous vehicles in the real world. By conducting extensive testing in advance in a risk-free virtual setting, the system identifies and resolves safety issues beforehand, reducing the need for repeated real-world testing and minimizing overall testing time and risk