arXiv subjects

Jongwoo Lee

Publications and source records attributed to Jongwoo Lee.

3 recordsLinked to original sources

arXiv2026

Characterizing and modeling the patterns of vehicle movement on road networks

Understanding vehicle movement on road networks is closely related to various practical and theoretical issues. While recent works have focused on which cost vehicles minimize while moving, how they move to minimize that cost remains less explored. In this work, we analyze large-scale data of individual vehicle trajectories in real-world road networks to identify cost-minimizing movement patterns of vehicles and the influence of road network structure on such movement. We observed that vehicle movements exhibit three phases: the beginning, middle, and end of trips. At the beginning and end, vehicles detour more, lose directional memory quickly, and travel at lower speeds than during the middle. In contrast, during the middle, they tend to detour less, maintain directional memory, and travel faster than at the beginning and end. Finally, at the beginning and end, vehicles exhibit similar detour and velocity patterns, except the direction of movement. To understand these patterns, we propose a double-layered network model mimicking the hierarchical structure of real-world road networks. We found that when vehicles move across our model network while minimizing travel time, they tend to concentrate on high-level roads, and the three observed movement phases are reproduced. Consequently, when a vehicle moves between a given origin-destination pair, it must enter and exit these high-level roads. This causes it to deviate from the trajectory that minimizes travel distance between the same origin-destination pair -- particularly at the beginning and end of the trip. Our results reveal common patterns underlying individual vehicle movements that appear highly diverse at first glance, demonstrating that these patterns emerge because vehicles leverage the characteristics of hierarchical road networks to minimize travel time.

physics.soc-ph↗

arXiv2026

Depth-Guided Privacy-Preserving Visual Localization Using 3D Sphere Clouds

The emergence of deep neural networks capable of revealing high-fidelity scene details from sparse 3D point clouds has raised significant privacy concerns in visual localization involving private maps. Lifting map points to randomly oriented 3D lines is a well-known approach for obstructing undesired recovery of the scene images, but these lines are vulnerable to a density-based attack that can recover the point cloud geometry by observing the neighborhood statistics of lines. With the aim of nullifying this attack, we present a new privacy-preserving scene representation called \emph{sphere cloud}, which is constructed by lifting all points to 3D lines crossing the centroid of the map, resembling points on the unit sphere. Since lines are most dense at the map centroid, the sphere cloud mislead the density-based attack algorithm to incorrectly yield points at the centroid, effectively neutralizing the attack. Nevertheless, this advantage comes at the cost of i) a new type of attack that may directly recover images from this cloud representation and ii) unresolved translation scale for camera pose estimation. To address these issues, we introduce a simple yet effective cloud construction strategy to thwart new attack and propose an efficient localization framework to guide the translation scale by utilizing absolute depth maps acquired from on-device time-of-flight (ToF) sensors. Experimental results on public RGB-D datasets demonstrate sphere cloud achieves competitive privacy-preserving ability and localization runtime while not excessively compensating the pose estimation accuracy compared to other depth-guided localization methods.

cs.CV↗

arXiv2020

How does the structure embedded in learning policy affect learning quadruped locomotion?

Reinforcement learning (RL) is a popular data-driven method that has demonstrated great success in robotics. Previous works usually focus on learning an end-to-end (direct) policy to directly output joint torques. While the direct policy seems convenient, the resultant performance may not meet our expectations. To improve its performance, more sophisticated reward functions or more structured policies can be utilized. This paper focuses on the latter because the structured policy is more intuitive and can inherit insights from previous model-based controllers. It is unsurprising that the structure, such as a better choice of the action space and constraints of motion trajectory, may benefit the training process and the final performance of the policy at the cost of generality, but the quantitative effect is still unclear. To analyze the effect of the structure quantitatively, this paper investigates three policies with different levels of structure in learning quadruped locomotion: a direct policy, a structured policy, and a highly structured policy. The structured policy is trained to learn a task-space impedance controller and the highly structured policy learns a controller tailored for trot running, which we adopt from previous work. To evaluate trained policies, we design a simulation experiment to track different desired velocities under force disturbances. Simulation results show that structured policy and highly structured policy require 1/3 and 3/4 fewer training steps than the direct policy to achieve a similar level of cumulative reward, and seem more robust and efficient than the direct policy. We highlight that the structure embedded in the policies significantly affects the overall performance of learning a complicated task when complex dynamics are involved, such as legged locomotion.

cs.RO↗