Recent visual autonomous perception systems achieve remarkable performances with deep representation learning. However, they fail in scenarios with challenging illumination. While event cameras can mitigate this problem, there is a lack of a large-scale dataset to develop event-enhanced deep visual perception models in autonomous driving scenes. To address the gap, we present theeAP(event-enhancedAutonomousPerception) dataset, the largest dataset with event cameras for autonomous perception. We demonstrate howeAPcan facilitate the study of different autonomous perception tasks, including 3D vehicle detection and object time-to-contact (TTC) estimation, through deep representation learning. Based oneAP, we demonstrate the first successful use of events to improve a popular 3D vehicle detection network in challenging illumination scenarios.eAPalso enables a devoted study of the representation learning problem of object TTC estimation. We show how a geometry-aware representation learning framework leads to the best event-based object TTC estimation network that operates at 200 FPS. The dataset, code, and pre-trained models will be made publicly available for future research.

Trajectory prediction for surrounding agents is a challenging task in autonomous driving due to its inherent uncertainty and underlying multimodality. Unlike prevailing data-driven methods that primarily rely on supervised learning, in this paper, we introduce a novel Graph-oriented Inverse Reinforcement Learning (GoIRL) framework, which is an IRL-based predictor equipped with vectorized context representations. We develop a feature adaptor to effectively aggregate lane-graph features into grid space, enabling seamless integration with the maximum entropy IRL paradigm to infer the reward distribution and obtain the policy that can be sampled to induce multiple plausible plans. Furthermore, conditioned on the sampled plans, we implement a hierarchical parameterized trajectory generator with a refinement module to enhance prediction accuracy and a probability fusion strategy to boost prediction confidence. Extensive experimental results showcase our approach not only achieves state-of-the-art performance on the large-scale Argoverse & nuScenes motion forecasting benchmarks but also exhibits superior generalization abilities compared to existing supervised models.

In the practical application of point cloud completion tasks, real data quality is usually much worse than the CAD datasets used for training. A small amount of noisy data will usually significantly impact the overall system's accuracy. In this paper, we propose a quality evaluation network to score the point clouds and help judge the quality of the point cloud before applying the completion model. We believe our scoring method can help researchers select more appropriate point clouds for subsequent completion and reconstruction and avoid manual parameter adjustment. Moreover, our evaluation model is fast and straightforward and can be directly inserted into any model's training or use process to facilitate the automatic selection and post-processing of point clouds. We propose a complete dataset construction and model evaluation method based on ShapeNet. We verify our network using detection and flow estimation tasks on KITTI, a real-world dataset for autonomous driving. The experimental results show that our model can effectively distinguish the quality of point clouds and help in practical tasks.

Due to the inherent ill-posed nature of 2D-3D projection, monocular 3D object detection lacks accurate depth recovery ability. Although the deep neural network (DNN) enables monocular depth-sensing from high-level learned features, the pixel-level cues are usually omitted due to the deep convolution mechanism. To benefit from both the pow-erful feature representation in DNN and pixel-level geomet-ric constraints, we reformulate the monocular object depth estimation as a progressive refinement problem and propose a joint semantic and geometric cost volume to model the depth error. Specifically, we first leverage neural networks to learn the object position, dimension, and dense normal-ized 3D object coordinates. Based on the object depth, the dense coordinates patch together with the corresponding object features is reprojected to the image space to build a cost volume in a joint semantic and geometric error man-ner. The final depth is obtained by feeding the cost volume to a refinement network, where the distribution of semantic and geometric error is regularized by direct depth supervision. Through effectively mitigating depth error by the re-finement framework, we achieve state-of-the-art results on both the KITTI and Waymo datasets.11Code available at https://github.com/lianqingll/MonoJSG

Segmenting moving objects in the 3D LiDAR point cloud can provide important guidance to localization, mapping and decision-making for self-driving vehicles. As for the conventional approaches to point cloud segmentation, they rely on semantic-level information, which makes it inevitable for long-tail problems to arise as there are always unseen types of objects on the road. To achieve moving segmentation while avoiding the reliance on the object category, the point motion is identified in this paper by fully exploring and aggregating the point-level geometric consistency in sequential point clouds. More specifically, an end-to-end point-wise uncertainty weighted aggregation approach known as PUA-MOS is proposed to segment the moving points in 3D LiDAR Data. Our method is applicable to estimate point-wise moving mask, scene flow and rigid-body transformation simultaneously in a coarse- to-fine network, where the relations between each prediction are implicitly learned. To explicitly model the inner and inter relations across these predictions among all points, the point- wise estimation and the average value of the same motion points are aggregated according to a predicted uncertainty. Then, the aggregated estimation is fed again into the next-level fusion, where the points will be re-segmented using the aggregated mask from the last level. Through iterative joint aggregation, our PUA-MOS outperforms the previous methods significantly on both KITTI [4] and Waymo [26] datasets. The code will be provided to generate the moving segmentation labels on both datasets for reproduction.

Motion prediction for traffic participants is essential for a safe and robust automated driving system, especially in cluttered urban environments. However, it is highly challenging due to the complex road topology as well as the uncertain intentions of the other agents. In this paper, we present a graph-based trajectory prediction network named the Dual Scale Predictor (DSP), which encodes both the static and dynamical driving context in a hierarchical manner. Different from methods based on a rasterized map or sparse lane graph, we consider the driving context as a graph with two layers, focusing on both geometrical and topological features. Graph neural networks (GNNs) are applied to extract features with different levels of granularity, and features are subsequently aggregated with attention-based inter-layer networks, realizing better local-global feature fusion. Following the recent goal-driven trajectory prediction pipeline, goal candidates with high likelihood for the target agent are extracted, and predicted trajectories are generated conditioned on these goals. Thanks to the proposed dual-scale context fusion network, our DSP is able to generate accurate and human-like multi-modal trajectories. We evaluate the proposed method on the large-scale Argoverse motion forecasting benchmark, and it achieves promising results, outperforming the recent state-of-the-art methods. We release the code on our project website.11https://github.com/HKUST-Aerial-Robotics/DSP

Directly learning multiple 3D objects motion from sequential images is difficult, while the geometric bundle adjustment lacks the ability to localize the invisible object centroid. To benefit from both the powerful object understanding skill from deep neural network meanwhile tackle precise geometry modeling for consistent trajectory estimation, we propose a joint spatial-temporal optimization-based stereo 3D object tracking method. From the network, we detect corresponding 2D bounding boxes on adjacent images and regress an initial 3D bounding box. Dense object cues (local depth and local coordinates) that associating to the object centroid are then predicted using a region-based network. Considering both the instant localization accuracy and motion consistency, our optimization models the relations between the object centroid and observed cues into a joint spatial-temporal error function. All historic cues will be summarized to contribute to the current estimation by a per-frame marginalization strategy without repeated computation. Quantitative evaluation on the KITTI tracking dataset shows our approach outperforms previous image-based 3D tracking methods by significant margins. We also report extensive results on multiple categories and larger datasets (KITTI raw and Argoverse Tracking) for future benchmarking.

We propose a 3D object detection method for autonomous driving by fully exploiting the sparse and dense, semantic and geometry information in stereo imagery. Our method, called Stereo R-CNN, extends Faster R-CNN for stereo inputs to simultaneously detect and associate object in left and right images. We add extra branches after stereo Region Proposal Network (RPN) to predict sparse keypoints, viewpoints, and object dimensions, which are combined with 2D left-right boxes to calculate a coarse 3D object bounding box. We then recover the accurate 3D bounding box by a region-based photometric alignment using left and right RoIs. Our method does not require depth input and 3D position supervision, however, outperforms all existing fully supervised image-based methods. Experiments on the challenging KITTI dataset show that our method outperforms the state-of-the-art stereo-based method by around 30% AP on both 3D detection and 3D localization tasks. Code will be made publicly available.

The monocular visual-inertial system (VINS), which consists one camera and one low-cost inertial measurement unit (IMU), is a popular approach to achieve accurate 6-DOF state estimation. However, such locally accurate visual-inertial odometry is prone to drift and cannot provide absolute pose estimation. Leveraging history information to relocalize and correct drift has become a hot topic. In this paper, we propose a monocular visual-inertial SLAM system, which can relocalize camera and get the absolute pose in a previous-built map. Then 4-DOF pose graph optimization is performed to correct drifts and achieve global consistent. The 4-DOF contains x, y, z, and yaw angle, which is the actual drifted direction in the visual-inertial system. Furthermore, the proposed system can reuse a map by saving and loading it in an efficient way. Current map and previous map can be merged together by the global pose graph optimization. We validate the accuracy of our system on public datasets and compare against other state-of-the-art algorithms. We also evaluate the map merging ability of our system in the large-scale outdoor environment. The source code of map reuse is integrated into our public code, VINS-Monol11https://github.com/HKUST-Aerial-Robotics/VINS-Mono.

One camera and one low-cost inertial measurement unit (IMU) form a monocular visual-inertial system (VINS), which is the minimum sensor suite (in size, weight, and power) for the metric six degrees-of-freedom (DOF) state estimation. In this paper, we present VINS-Mono: a robust and versatile monocular visual-inertial state estimator. Our approach starts with a robust procedure for estimator initialization. A tightly coupled, nonlinear optimization-based method is used to obtain highly accurate visual-inertial odometry by fusing preintegrated IMU measurements and feature observations. A loop detection module, in combination with our tightly coupled formulation, enables relocalization with minimum computation. We additionally perform 4-DOF pose graph optimization to enforce the global consistency. Furthermore, the proposed system can reuse a map by saving and loading it in an efficient way. The current and previous maps can be merged together by the global pose graph optimization. We validate the performance of our system on public datasets and real-world experiments and compare against other state-of-the-art algorithms. We also perform an onboard closed-loop autonomous flight on the microaerial-vehicle platform and port the algorithm to an iOS-based demonstration. We highlight that the proposed work is a reliable, complete, and versatile system that is applicable for different applications that require high accuracy in localization. We open source our implementations for both PCs (https://github.com/HKUST-Aerial-Robotics/VINS-Mono) and iOS mobile devices (https://github.com/HKUST-Aerial-Robotics/VINS-Mobile).

Mobile phones equipped with a monocular camera and an inertial measurement unit (IMU) are ideal platforms for augmented reality (AR) applications, but the lack of direct metric distance measurement and the existence of aggressive motions pose significant challenges on the localization of the AR device. In this work, we propose a tightly-coupled, optimization-based, monocular visual-inertial state estimation for robust camera localization in complex indoor and outdoor environments. Our approach does not require any artificial markers, and is able to recover the metric scale using the monocular camera setup. The whole system is capable of online initialization without relying on any assumptions about the environment. Our tightly-coupled formulation makes it naturally robust to aggressive motions. We develop a lightweight loop closure module that is tightly integrated with the state estimator to eliminate drift. The performance of our proposed method is demonstrated via comparison against state-of-the-art visual-inertial state estimators on public datasets and real-time AR applications on mobile devices. We release our implementation on mobile devices as open source software1.