Deep reinforcement learning (DRL) integrates the feature representation ability of deep learning with the decision-making ability of reinforcement learning so that it can achieve powerful end-to-end learning control capabilities. In the past decade, DRL has made substantial advances in many tasks that require perceiving high-dimensional input and making optimal or near-optimal decisions. However, there are still many challenging problems in the theory and applications of DRL, especially in learning control tasks with limited samples, sparse rewards, and multiple agents. Researchers have proposed various solutions and new theories to solve these problems and promote the development of DRL. In addition, deep learning has stimulated the further development of many subfields of reinforcement learning, such as hierarchical reinforcement learning (HRL), multiagent reinforcement learning, and imitation learning. This article gives a comprehensive overview of the fundamental theories, key algorithms, and primary research domains of DRL. In addition to value-based and policy-based DRL algorithms, the advances in maximum entropy-based DRL are summarized. The future research topics of DRL are also analyzed and discussed.

Accurate maneuver prediction for surrounding vehicles enables intelligent vehicles to make safe and socially compliant decisions in advance, thus improving the safety and comfort of the driving. The main contribution of this paper is proposing a practical, high-performance, and low-cost maneuver-prediction approach for intelligent vehicles. Our approach is based on a dynamic Bayesian network, which exploits multiple predictive features, namely, historical states of predicting vehicles, road structures, as well as traffic interactions for inferring the probability of each maneuver. The paper also presents algorithms of feature extraction for the network. Our approach is verified on real traffic data in large-scale publicly available datasets. The results show that our approach can recognize the lane-change maneuvers with an F1 score of 80% and an advanced prediction time of 3.75 s, which greatly improves the performance on prediction compared to other baseline approaches.

Multi-object tracking is a crucial problem for autonomous vehicle. Most state-of-the-art approaches adopt the tracking-by-detection strategy, which is a two-step procedure consisting of the detection module and the tracking module. In this paper, we improve both steps. We improve the detection module by incorporating the temporal information, which is beneficial for detecting small objects. For the tracking module, we propose a novel compressed deep Convolutional Neural Network (CNN) feature based Correlation Filter tracker. By carefully integrating these two modules, the proposed multi-object tracking approach has the ability of re-identification (ReID) once the tracked object gets lost. Extensive experiments were performed on the KITTI and MOT2015 tracking benchmarks. Results indicate that our approach outperforms most state-of-the-art tracking approaches.

Loop closing and relocalization are crucial techniques to establish reliable and robust long-term SLAM by addressing pose estimation drift and degeneration. This article begins by formulating loop closing and relocalization within a unified framework. Then, we propose a novel multihead network, LCR-Net, to tackle both tasks effectively. It exploits novel feature extraction and a pose-aware attention mechanism to precisely estimate similarities and 6-DoF poses between pairs of LiDAR scans. In the end, we integrate our LCR-Net into a SLAM system and achieve robust and accurate online LiDAR SLAM in outdoor driving environments. We thoroughly evaluate our LCR-Net through three setups derived from loop closing and relocalization, including candidate retrieval, closed-loop point cloud registration, and continuous relocalization using multiple datasets. The results demonstrate that LCR-Net excels in all three tasks, surpassing the state-of-the-art methods and exhibiting a remarkable generalization ability. Notably, our LCR-Net outperforms baseline methods without using a time-consuming robust pose estimator, rendering it suitable for online SLAM applications. To the best of the authors' knowledge, the integration of LCR-Net yields the first LiDAR SLAM with the capability of deep loop closing and relocalization.

Multi-camera 3D perception has emerged as a prominent research field in autonomous driving, offering a viable and cost-effective alternative to LiDAR-based solutions. The existing multi-camera algorithms primarily rely on monocular 2D pre-training. However, the monocular 2D pre-training overlooks the spatial and temporal correlations among the multi-camera system. To address this limitation, we propose the first multi-camera unified pre-training framework, called UniScene, which involves initially reconstructing the 3D scene as the foundational stage and subsequently fine-tuning the model on downstream tasks. Specifically, we employ Occupancy as the general representation for the 3D scene, enabling the model to grasp geometric priors of the surrounding world through pre-training. A significant benefit of UniScene is its capability to utilize a considerable volume of unlabeled image-LiDAR pairs for pre-training purposes. The proposed multi-camera unified pre-training framework demonstrates promising results in key tasks such as multi-camera 3D object detection and surrounding semantic scene completion. When compared to monocular pre-training methods, UniScene shows a significant improvement of about 2.0% in mAP and 2.0% in NDS for multi-camera 3D object detection, as well as a 3% increase in mIoU for surrounding semantic scene completion. By adopting our unified pre-training method, a 25% reduction in 3D training annotation costs can be achieved, offering significant practical value for the implementation of real-world autonomous driving.