Sangjae Bae

Title: Sangjae Bae: Innovations in Autonomous Driving and Motion Forecasting.

Introduction

Sangjae Bae is an inventor based in San Jose, California. His work includes technologies for autonomous driving, motion forecasting, and vehicle planning. He is associated with Honda Motor Co., Ltd., and holds 3 patents.

Latest Patents

Bae’s recent patent work includes methods for predicting the movement of road users and planning safe, efficient paths for autonomous vehicles.

“Adaptive Prediction Ensemble for Motion Forecasting” describes a system that combines neural-network predictions with rule-based predictions. It processes road-map images and historical trajectories, generates candidate forecasts, and uses a routing function to rank the predictions. The neural network and routing function are trained using ground-truth trajectories and ranking results.

“Multi-Profile Quadratic Programming (MPQP) for Optimal Gap Selection and Speed Planning of Autonomous Driving” presents a method for identifying usable driving areas along an autonomous vehicle’s path. The method builds a space-time graph that represents the vehicle’s route and the trajectories of intersecting road users. It divides the graph into cells representing viable, unoccupied spaces, identifies possible passageways, and uses quadratic programming to select a desired passageway when multiple options are available.

Career Highlights

Bae’s listed inventions address key challenges in autonomous driving: anticipating how other agents may move and selecting a safe, practical route through available space. His work at Honda connects predictive modeling with vehicle planning and decision-making.

Collaborations

Bae’s listed coworker is Faizan M. Tariq. Their association reflects collaborative work in the development of autonomous-driving technologies.

Conclusion

Sangjae Bae’s patent portfolio focuses on combining motion prediction, trajectory analysis, and optimization to support safer and more capable autonomous vehicles.