An autonomous vehicle (AV) requires an accurate perception of its surrounding environment to operate reliably. The perception system of an AV, which normally employs machine learning (e.g., deep learning), transforms sensory data into semantic information that enables autonomous driving. Object detection is a fundamental function of this perception system, which has been tackled by several works, most of them using 2D detection methods. However, the 2D methods do not provide depth information, which is required for driving tasks, such as path planning, collision avoidance, and so on. Alternatively, the 3D object detection methods introduce a third dimension that reveals more detailed object's size and location information. Nonetheless, the detection accuracy of such methods needs to be improved. To the best of our knowledge, this is the first survey on 3D object detection methods used for autonomous driving applications. This paper presents an overview of 3D object detection methods and prevalently used sensors and datasets in AVs. It then discusses and categorizes the recent works based on sensors modalities into monocular, point cloud-based, and fusion methods. We then summarize the results of the surveyed works and identify the research gaps and future research directions.

3D object detection is a common function within the perception system of an autonomous vehicle and outputs a list of 3D bounding boxes around objects of interest. Various 3D object detection methods have relied on fusion of different sensor modalities to overcome limitations of individual sensors. However, occlusion, limited field-of-view and low-point density of the sensor data cannot be reliably and cost-effectively addressed by multi-modal sensing from a single point of view. Alternatively, cooperative perception incorporates information from spatially diverse sensors distributed around the environment as a way to mitigate these limitations. This article proposes two schemes for cooperative 3D object detection using single modality sensors. The early fusion scheme combines point clouds from multiple spatially diverse sensing points of view before detection. In contrast, the late fusion scheme fuses the independently detected bounding boxes from multiple spatially diverse sensors. We evaluate the performance of both schemes, and their hybrid combination, using a synthetic cooperative dataset created in two complex driving scenarios, a T-junction and a roundabout. The evaluation shows that the early fusion approach outperforms late fusion by a significant margin at the cost of higher communication bandwidth. The results demonstrate that cooperative perception can recall more than 95% of the objects as opposed to 30% for single-point sensing in the most challenging scenario. To provide practical insights into the deployment of such system, we report how the number of sensors and their configuration impact the detection performance of the system.

Real-time registration of partially overlapping point clouds has emerging applications in cooperative perception for autonomous vehicles and multi-agent SLAM. The relative translation between point clouds in these applications is higher than in traditional SLAM and odometry applications, which challenges the identification of correspondences and a successful registration. In this paper, we propose a novel registration method for partially overlapping point clouds where correspondences are learned using an efficient point-wise feature encoder, and refined using a graph-based attention network. This attention network exploits geometrical relationships between key points to improve the matching in point clouds with low overlap. At inference time, the relative pose transformation is obtained by robustly fitting the correspondences through sample consensus. The evaluation is performed on the KITTI dataset and a novel synthetic dataset including low-overlapping point clouds with displacements of up to 30 m. The proposed method achieves on-par performance with state-of-the-art methods on the KITTI dataset, and outperforms existing methods for low overlapping point clouds. Additionally, the proposed method achieves significantly faster inference times, as low as 410 ms, between 5 and 35 times faster than competing methods. Our code and dataset are available at https://github.com/eduardohenriquearnold/fastreg .

Lane change (LC) is one of the safety-critical manoeuvres in highway driving according to various road accident records. Thus, reliably predicting such manoeuvre in advance is critical for the safe and comfortable operation of automated driving systems. The majority of previous studies rely on detecting a manoeuvre that has been already started, rather than predicting the manoeuvre in advance. Furthermore, most of the previous works do not estimate the key timings of the manoeuvre (e.g., crossing time), which can actually yield more useful information for the decision making in the ego vehicle. To address these shortcomings, this paper proposes a novel multi-task model to simultaneously estimate the likelihood of LC manoeuvres and the time-to-lane-change (TTLC). In both tasks, an attention-based convolutional neural network (CNN) is used as a shared feature extractor from a bird’s eye view representation of the driving environment. The spatial attention used in the CNN model improves the feature extraction process by focusing on the most relevant areas of the surrounding environment. In addition, two novel curriculum learning schemes are employed to train the proposed approach. The extensive evaluation and comparative analysis of the proposed method in existing benchmark datasets show that the proposed method outperforms state-of-the-art LC prediction models, particularly considering long-term prediction performance.

While extracting meaningful information from big data is getting relevance, literature lacks information on how to handle sensitive data by different project partners in order to collectively answer research questions (RQs), especially on impact assessment of new automated driving technologies. This paper presents the application of an established reference piloting methodology and the consequent development of a coherent, robust workflow. Key challenges include ensuring methodological soundness and data validity while protecting partners’ intellectual property. The authors draw on their experiences in a 34-partner project aimed at assessing the impact of advanced automated driving functions, across 10 European countries. In the first step of the workflow, we captured the quantitative requirements of each RQ in terms of the relevant data needed from the tests. Most of the data come from vehicular sensors, but subjective data from questionnaires are processed as well. Next, we set up a data management process involving several partners (vehicle manufacturers, research institutions, suppliers and developers), with different perspectives and requirements. Finally, we deployed the system so that it is fully integrated within the project big data toolchain and usable by all the partners. Based on our experience, we highlight the importance of the reference methodology to theoretically inform and coherently manage all the steps of the project and the need for effective and efficient tools, in order to support the everyday work of all the involved research teams, from vehicle manufacturers to data analysts.

As industrial research in automated driving is rapidly advancing, it is of paramount importance to analyze field data from extensive road tests. This paper investigates the design and development of a toolchain to process and manage experimental data to answer a set of research questions about the evaluation of automated driving functions at various levels, from technical system functioning to overall impact assessment. We have faced this challenge in L3Pilot, the first comprehensive test of automated driving functions (ADFs) on public roads in Europe. L3Pilot is testing ADFs in vehicles made by 13 companies. The tested functions are mainly of Society of Automotive Engineers (SAE) automation level 3, some of them of level 4. In this context, the presented toolchain supports various confidentiality levels, and allows cross-vehicle owner seamless data management, with the efficient storage of data and their iterative processing with a variety of analysis and evaluation tools. Most of the toolchain modules have been developed to a prototype version in a desktop/cloud environment, exploiting state-of-the-art technology. This has allowed us to efficiently set up what could become a comprehensive edge-to-cloud reference architecture for managing data in automated vehicle tests. The project has been released as open source, the data format into which all vehicular signals, recorded in proprietary formats, were converted, in order to support efficient processing through multiple tools, scalability and data quality checking. We expect that this format should enhance research on automated driving testing, as it provides a shared framework for dealing with data from collection to analysis. We are confident that this format, and the information provided in this article, can represent a reference for the design of future architectures to implement in vehicles.