Unified Map Prior Encoder for Mapping and Planning

Abstract

Online mapping and end-to-end (E2E) planning in autonomous driving are still largely sensor-centric, leaving rich map priors—HD/SD vector maps, rasterized SD maps, and satellite imagery—underused due to heterogeneity, pose drift, and inconsistent availability at test time. We present emph{UMPE}, a Unified Map Prior Encoder that can ingest any subset of four priors and fuse them with BEV features for both mapping and planning. emph{UMPE} has two branches. The vector encoder pre-aligns HD/SD polylines with a frame-wise SE(2) correction, encodes points via multi-frequency sinusoidal features, and produces polyline tokens with confidence scores. BEV queries then apply cross-attention with confidence bias, followed by normalized channel-wise gating to avoid length imbalance and to softly down-weight uncertain sources. The raster encoder shares a ResNet-18 backbone conditioned by FiLM (scaling/shift at every stage), performs SE(2) micro-alignment, and injects priors through zero-initialized residual fusion so the network starts from a do-no-harm baseline and learns to add only useful prior evidence. A vector-then-raster fusion order reflects the inductive bias of “geometry first, appearance second.” On nuScenes mapping, emph{UMPE} lifts MapTRv2 from 61.5 → 67.4 mAP (+5.9) and MapQR from 66.4 → 71.7 mAP (+5.3). On Argoverse2, emph{UMPE} adds +4.1 mAP over strong baselines. emph{UMPE} is compositional: when trained with all priors, it outperforms single-prior models even when only one prior is available at test time, demonstrating powerset robustness. For E2E planning (VAD backbone, nuScenes), emph{UMPE} reduces trajectory error from 0.72 → 0.42 m L2 (avg. −0.30 m) and collision rate from 0.22% → 0.12% (−0.10%), surpassing recent prior-injection methods. These results show that a unified, alignment-aware treatment of heterogeneous map priors yields better mapping and better planning.