Search

Microscopic traffic simulation plays a crucial role in transportation engineering by providing insights into individual vehicle behavior and overall traffic flow. However, creating a realistic simulator that accurately replicates human driving behaviors in various traffic conditions presents significant challenges. Traditional simulators relying on heuristic models often fail to deliver accurate simulations due to the complexity of real-world traffic environments. Due to the covariate shift issue, existing imitation learning-based simulators often fail to generate stable long-term simulations. In this paper, we propose a novel approach called learner-aware supervised imitation learning to address the covariate shift problem in multi-agent imitation learning. By leveraging a variational autoencoder simultaneously modeling the expert and learner state distribution, our approach augments expert states such that the augmented state is aware of learner state distribution. Our method, applied to urban traffic simulation, demonstrates significant improvements over existing state-of-the-art baselines in both short-term microscopic and long-term macroscopic realism when evaluated on the real-world dataset pNEUMA., Comment: Accepted by CVPR 2024. arXiv admin note: text overlap with arXiv:2306.06401

Artificial light plays an integral role in modern cities, significantly enhancing human productivity and the efficiency of civilization. However, excessive illumination can lead to light pollution, posing non-negligible threats to economic burdens, ecosystems, and human health. Despite its critical importance, the exploration of its causes remains relatively limited within the field of artificial intelligence, leaving an incomplete understanding of the factors contributing to light pollution and sustainable illumination planning distant. To address this gap, we introduce a novel framework named Causally Aware Generative Adversarial Networks (CAGAN). This innovative approach aims to uncover the fundamental drivers of light pollution within cities and offer intelligent solutions for optimal illumination resource allocation in the context of sustainable urban development. We commence by examining light pollution across 33,593 residential areas in seven global metropolises. Our findings reveal substantial influences on light pollution levels from various building types, notably grasslands, commercial centers and residential buildings as significant contributors. These discovered causal relationships are seamlessly integrated into the generative modeling framework, guiding the process of generating light pollution maps for diverse residential areas. Extensive experiments showcase CAGAN's potential to inform and guide the implementation of effective strategies to mitigate light pollution. Our code and data are publicly available at https://github.com/zhangyuuao/Light_Pollution_CAGAN., Comment: 9pages, 9figures, accepted by AAAI2024, AI for Social Impact (Special Track)

Building a multi-modality multi-task neural network toward accurate and robust performance is a de-facto standard in perception task of autonomous driving. However, leveraging such data from multiple sensors to jointly optimize the prediction and planning tasks remains largely unexplored. In this paper, we present FusionAD, to the best of our knowledge, the first unified framework that fuse the information from two most critical sensors, camera and LiDAR, goes beyond perception task. Concretely, we first build a transformer based multi-modality fusion network to effectively produce fusion based features. In constrast to camera-based end-to-end method UniAD, we then establish a fusion aided modality-aware prediction and status-aware planning modules, dubbed FMSPnP that take advantages of multi-modality features. We conduct extensive experiments on commonly used benchmark nuScenes dataset, our FusionAD achieves state-of-the-art performance and surpassing baselines on average 15% on perception tasks like detection and tracking, 10% on occupancy prediction accuracy, reducing prediction error from 0.708 to 0.389 in ADE score and reduces the collision rate from 0.31% to only 0.12%.

A realistic long-term microscopic traffic simulator is necessary for understanding how microscopic changes affect traffic patterns at a larger scale. Traditional simulators that model human driving behavior with heuristic rules often fail to achieve accurate simulations due to real-world traffic complexity. To overcome this challenge, researchers have turned to neural networks, which are trained through imitation learning from human driver demonstrations. However, existing learning-based microscopic simulators often fail to generate stable long-term simulations due to the \textit{covariate shift} issue. To address this, we propose a history-masked multi-agent imitation learning method that removes all vehicles' historical trajectory information and applies perturbation to their current positions during learning. We apply our approach specifically to the urban traffic simulation problem and evaluate it on the real-world large-scale pNEUMA dataset, achieving better short-term microscopic and long-term macroscopic similarity to real-world data than state-of-the-art baselines., Comment: update

Imitation learning holds great promise for addressing the complex task of autonomous urban driving, as experienced human drivers can navigate highly challenging scenarios with ease. While behavior cloning is a widely used imitation learning approach in autonomous driving due to its exemption from risky online interactions, it suffers from the covariate shift issue. To address this limitation, we propose a context-conditioned imitation learning approach that employs a policy to map the context state into the ego vehicle's future trajectory, rather than relying on the traditional formulation of both ego and context states to predict the ego action. Additionally, to reduce the implicit ego information in the coordinate system, we design an ego-perturbed goal-oriented coordinate system. The origin of this coordinate system is the ego vehicle's position plus a zero mean Gaussian perturbation, and the x-axis direction points towards its goal position. Our experiments on the real-world large-scale Lyft and nuPlan datasets show that our method significantly outperforms state-of-the-art approaches., Comment: Accepted by Robotics: Science and Systems

In this paper, we aim to forecast a future trajectory distribution of a moving agent in the real world, given the social scene images and historical trajectories. Yet, it is a challenging task because the ground-truth distribution is unknown and unobservable, while only one of its samples can be applied for supervising model learning, which is prone to bias. Most recent works focus on predicting diverse trajectories in order to cover all modes of the real distribution, but they may despise the precision and thus give too much credit to unrealistic predictions. To address the issue, we learn the distribution with symmetric cross-entropy using occupancy grid maps as an explicit and scene-compliant approximation to the ground-truth distribution, which can effectively penalize unlikely predictions. In specific, we present an inverse reinforcement learning based multi-modal trajectory distribution forecasting framework that learns to plan by an approximate value iteration network in an end-to-end manner. Besides, based on the predicted distribution, we generate a small set of representative trajectories through a differentiable Transformer-based network, whose attention mechanism helps to model the relations of trajectories. In experiments, our method achieves state-of-the-art performance on the Stanford Drone Dataset and Intersection Drone Dataset., Comment: CVPR2022

In this paper, a multi-parameterized proximal point algorithm combining with a relaxation step is developed for solving convex minimization problem subject to linear constraints. We show its global convergence and sublinear convergence rate from the prospective of variational inequality. Preliminary numerical experiments on testing a sparse minimization problem from signal processing indicate that the proposed algorithm performs better than some well-established methods, Comment: 7 pages

Environment is one of the key external drivers of the galaxies, while active galactic nucleus (AGN) is one of the key internal drivers. Both of them play fundamental roles in regulating the formation and evolution of galaxies. We explore the interrelationship between environment and AGN in SDSS. At a given stellar mass, the specific star formation rate distribution of the AGN host galaxies remains unchanged with over-density, with the peak of the distribution around the Green Valley. We show that, at a given stellar mass, the AGN fraction that has been commonly used in previous studies (defined as the number of AGNs relative to all galaxies including passive and star forming ones) does decrease with increasing over-density for satellites. This is largely due to the fact that the fraction of passive galaxies strongly depends on environment. In order to investigate the intrinsic correlation between AGN and environment, especially under the assumption that AGN feedback is responsible for star formation quenching, the AGN fraction should be defined as the number of AGNs relative to the star-forming galaxies only. With the new definition, we find little dependence of AGN fraction on over-density, central/satellite, and group halo mass. There is only marginal evidence that AGN may prefer denser regions, which is possibly due to more frequent interaction of galaxies or higher merger rate in groups. Our results support the scenario that internal secular evolution is the predominant mechanism of triggering AGN activity, while external environment related processes only play a minor role., Comment: Accepted for publication in MNRAS, submitted on June 4. 11 pages, 7 figures

We explore the interrelationships between the galaxy group halo mass and various observable group properties. We propose a simple scenario that describes the evolution of the central galaxies and their host dark matter halos. Star formation quenching is one key process in this scenario, which leads to the different assembly histories of blue groups (group with a blue central) and red groups (group with a red central). For blue groups, both the central galaxy and the halo continue to grow their mass. For red groups, the central galaxy has been quenched and its stellar mass remains about constant, while its halo continues to grow by merging smaller halos. From this simple scenario, we speculate about the driving properties that should strongly correlate with the group halo mass. We then apply the machine learning algorithm the Random Forest (RF) regressor to blue groups and red groups separately in the semianalytical model L-GALAXIES to explore these nonlinear multicorrelations and to verify the scenario as proposed above. Remarkably, the results given by the RF regressor are fully consistent with the prediction from our simple scenario and hence provide strong support for it. As a consequence, the group halo mass can be more accurately determined from observable galaxy properties by the RF regressor with a 50% reduction in error. A halo mass more accurately determined in this way also enables more accurate investigations on the galaxy-halo connection and other important related issues, including galactic conformity and the effect of halo assembly bias on galaxy assembly., Comment: Accepted for publication in The Astrophysical Journal, submitted on Feb 24. 16 pages, 6 figures. Grammar mistakes are corrected, results unchanged

The alternating direction method of multipliers (ADMM) were extensively investigated in the past decades for solving separable convex optimization problems. Fewer researchers focused on exploring its convergence properties for the nonconvex case although it performed surprisingly efficient. In this paper, we propose a symmetric ADMM based on different acceleration techniques for a family of potentially nonsmooth nonconvex programing problems with equality constraints, where the dual variables are updated twice with different stepsizes. Under proper assumptions instead of using the so-called Kurdyka-Lojasiewicz inequality, convergence of the proposed algorithm as well as its pointwise iteration-complexity are analyzed in terms of the corresponding augmented Lagrangian function and the primal-dual residuals, respectively. Performance of our algorithm is verified by some preliminary numerical examples on applications in sparse nonconvex/convex regularized minimization signal processing problems., Comment: 20 pages, submitted