Skip to main content
grasp.study
About
Explore
Pricing
Log in
Try it for free
Back to course
👋
Welcome to a Grasp lesson
Create a free account
to keep studying and create your own study path.
Generative AI for Autonomous Driving
·
Module 12
Safety Alignment for Autonomous Driving via GRPO
1
Imitation Learning's Unsafe Replications
Explain why imitation learning on human driving data can replicate unsafe behaviors like speeding
2
Plan-R1: Two-Stage Training (Imitation + GRPO)
Describe Plan-R1's two-stage framework: imitation pretraining followed by GRPO fine-tuning
3
Rule-Based Reward Functions for Safe Reinforcement Learning
Design a rule-based reward function to penalize collisions and enforce hard safety constraints
4
Crafting Rule-Based Reward Functions for Autonomous Driving
Design rule-based reward functions for lane compliance and adherence to speed limits
5
Reward Function Design for Safe and Comfortable Navigation
Design rule-based reward functions to encourage progress towards the goal and ensure passenger comfort via jerk minimization
6
Understanding VD-GRPO and Centering for Safe Reinforcement Learning
Explain Variance-Decoupled GRPO (VD-GRPO) and why replacing normalization with centering prevents reinforcing unsafe actions
7
GRPO vs. PPO: Verifiable Rewards for Reference-Free Safety Alignment
Compare GRPO and PPO for trajectory planning, explaining how verifiable rewards enable reference-free safety alignment
8
Open- vs. Closed-Loop Metrics for Autonomous Driving Planners
Distinguish open-loop from closed-loop evaluation metrics for autonomous driving planners (NAVSIM, nuPlan)
End of course,
Back to course
Previous module
Generative Planning for Autonomous Driving
Next module