Part 6/13:
Lane navigation around puddles: The system understands that venturing into oncoming traffic may be preferable to navigating through a puddle, based on scene context.
Understanding animal and pedestrian intent: The system patiently waits for crossing chickens or geese, interpreting scene semantics better than rule-based systems.
Predicting second-order effects: The neural network can recognize that a spinning-out vehicle will bounce into a lane and proactively brake earlier than traditional systems would, demonstrating an advanced understanding of dynamics.