ParkPredict: Motion and Intent Prediction of Vehicles in Parking Lots

TL;DR

This paper presents a study on predicting driver behavior in parking lots using deep learning models and a new dataset, demonstrating high accuracy in intent and trajectory prediction influenced by environmental understanding.

Contribution

The paper introduces a new parking lot dataset and compares deep learning models with a physics-based baseline for vehicle intent and trajectory prediction.

Findings

LSTM and CNN-LSTM models achieve ~85% top-1 intent accuracy

Knowledge of parking spot significantly improves trajectory prediction

Semantic environmental features enhance long-term prediction accuracy

Abstract

We investigate the problem of predicting driver behavior in parking lots, an environment which is less structured than typical road networks and features complex, interactive maneuvers in a compact space. Using the CARLA simulator, we develop a parking lot environment and collect a dataset of human parking maneuvers. We then study the impact of model complexity and feature information by comparing a multi-modal Long Short-Term Memory (LSTM) prediction model and a Convolution Neural Network LSTM (CNN-LSTM) to a physics-based Extended Kalman Filter (EKF) baseline. Our results show that 1) intent can be estimated well (roughly 85% top-1 accuracy and nearly 100% top-3 accuracy with the LSTM and CNN-LSTM model); 2) knowledge of the human driver's intended parking spot has a major impact on predicting parking trajectory; and 3) the semantic representation of the environment improves long term predictions.