A Computational Experimental and Simulation Framework for Robust Imagination and Distribution-Shifted Autonomous Driving

However, Model-Based Reinforcement Learning (MBRL) suffers from failure of long horizon planning due to the progressive latent collapse of differences between physically dissimilar states. In this paper, we study attractor bias, which is an emergent failure in which new latent states become increasingly biased towards known regions, leading to self-consistent yet possibly erroneous predictions. We define Local Latent Sensitivity (LLS) measures of change in predicted next state distributions in response to small perturbations of the deterministic recurrent state, as well as the Real-Time Attractor-Bias Monitor (RT-ABM), an aggregate measure of sensitivity loss, trajectory divergence, ensemble disagreement, replay buffer novelty, latent collapse, and reward overestimate. In addition, we present Sensitivity-Preserving DreamerV$3$ (SPDreamerV$3$) with one-sided sensitivity preservation loss. In our controlled fault injection experiment, sensitivity loss exceeded its diagnostic threshold at step $52$, occurring $14$ steps before trajectory divergence at step $66$. Robustness of the correlation coefficient was evaluated across different random seed runs by both Spearman and Pearson correlations, with thresholds established via held-out calibration. In our MetaDrive experiments, SP-DreamerV$3$ performed at $88$ success and $6$ collision rates, whereas DreamerV$3$ had $78$ success and $12$ collision rates. These experiments show proof-of-concept results that local latent sensitivity can help diagnose degradation of the learned world model.