A Transformer parking policy with cross-attention goal fusion and a GRU pedestrian predictor achieves 96.57% success in CARLA parking, but only in simulation with oracle goal and pedestrian states.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ParkFormer: A Transformer-Based Parking Policy with Goal Embedding and Pedestrian-Aware Control
A Transformer parking policy with cross-attention goal fusion and a GRU pedestrian predictor achieves 96.57% success in CARLA parking, but only in simulation with oracle goal and pedestrian states.