检索结果-内蒙古大学图书馆

Uncovering Variability in Human Driving Behavior Through Automatic Extraction of Similar Traffic Scenes from Large Naturalistic Datasets

学校读者我要写书评

暂无评论

Uncovering Variability in Human Driving Behavior Through Aut...

2023 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2023

作者： Siebinga, Olger Zgonnikov, Arkady Abbink, David Delft University of Technology Human-Robot Interaction Group Faculty of Mechanical Engineering Department of Cognitive Robotics Delft Netherlands

ISBN: (纸本)9798350337020

Recently, multiple naturalistic traffic datasets of human-driven trajectories have been published (e.g., highD, NGSim, and pNEUMA). These datasets have been used in studies that investigate variability in human driving behavior, for example for scenario-based validation of autonomous vehicle (AV) behavior, modeling driver behavior, or validating driver models. Thus far, these studies focused on the variability on an operational level (e.g., velocity profiles during a lane change), not on a tactical level (i.e., to change lanes or not). Investigating the variability on both levels is necessary to develop driver models and AV s that include multiple tactical behaviors. To expose multi-level variability, the human responses to the same traffic scene could be investigated. However, no method exists to automatically extract similar scenes from datasets. Here, we present a four-step extraction method that uses the Hausdorff distance, a mathematical distance metric for sets. We performed a case study on the highD dataset that showed that the method is practically applicable. The human responses to the selected scenes exposed the variability on both the tactical and operational levels. With this new method, the variability in operational and tactical human behavior can be investigated, without the need for costly and time-consuming driving-simulator experiments. © 2023 IEEE.

关键词：

Using Off-the-Shelf Deep Neural Networks for Position-Based Visual Servoing

学校读者我要写书评

暂无评论

Using Off-the-Shelf Deep Neural Networks for Position-Based...

27th RoboCup International Symposium, 2024

作者： Tschesche, Matteo Hofmann, Till Ferrein, Alexander Lakemeyer, Gerhard Mobile Autonomous Systems and Cognitive Robotics Institute FH Aachen University of Applied Sciences Aachen Germany Knowledge-Based Systems Group RWTH Aachen University Aachen Germany

ISBN: (纸本)9783031858581

Visual servoing is a well-established technique for object grasping and controls the robot in a closed-loop fashion. It typically uses hand-crafted features or a neural network that directly learns the control output. We propose an alternative approach that uses an off-the-shelf neural network object classifier and can therefore compute target poses without manually selected features while also not requiring training from scratch. Instead, the object classifier only needs to be fine-tuned on a domain-specific dataset, significantly reducing the amount of required training data. We describe a task sequencing approach that can control a robot with a mobile base and an additional gripper, which is a typical setup in many robotics applications. We evaluate the approach in the RoboCup Logistics League and demonstrate the reliability and speed of the proposed approach. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Mobile robots

Real-time 3D semantic occupancy prediction for autonomous vehicles using memory-efficient sparse convolution

学校读者我要写书评

暂无评论

Real-time 3D semantic occupancy prediction for autonomous ve...

IEEE Symposium on Intelligent Vehicle

作者： Samuel Sze Lars Kunze Department of Engineering Science Cognitive Robotics Group Oxford Robotics Institute University of Oxford

ISBN: (数字)9798350348811

ISBN: (纸本)9798350348828

In autonomous vehicles, understanding the surrounding 3D environment of the ego vehicle in real-time is essential. A compact way to represent scenes while encoding geometric distances and semantic object information is via 3D semantic occupancy maps. State of the art 3D mapping methods leverage transformers with cross-attention mechanisms to elevate 2D vision-centric camera features into the 3D domain. However, these methods encounter significant challenges in real-time applications due to their high computational demands during inference. This limitation is particularly problematic in autonomous vehicles, where GPU resources must be shared with other tasks such as localization and planning. In this paper, we introduce an approach that extracts features from front-view 2D camera images and LiDAR scans, then employs a sparse convolution network (Minkowski Engine), for 3D semantic occupancy prediction. Given that outdoor scenes in autonomous driving scenarios are inherently sparse, the utilization of sparse convolution is particularly apt. By jointly solving the problems of 3D scene completion of sparse scenes and 3D semantic segmentation, we provide a more efficient learning framework suitable for real-time applications in autonomous vehicles. We also demonstrate competitive accuracy on the nuScenes dataset.

关键词： Three-dimensional displays Convolution Semantic segmentation Semantics Feature extraction Cameras Transformers

A ROS 2-Based Navigation and Simulation Stack for the Robotino

学校读者我要写书评

暂无评论

A ROS 2-Based Navigation and Simulation Stack for the Rob...

27th RoboCup International Symposium, 2024

作者： Borse, Saurabh Viehmann, Tarik Ferrein, Alexander Lakemeyer, Gerhard The Mobile Autonomous Systems and Cognitive Robotics Institute FH Aachen University of Applied Science Aachen52066 Germany Knowledge-Based Systems Group RWTH Aachen University Aachen52074 Germany

ISBN: (纸本)9783031858581

The Robotino, developed by Festo Didactic, serves as a versatile platform in education and research for mobile robotics tasks. However, there currently is no ROS 2 integration for the Robotino available. In this paper we describe our work on a Webots simulation environment for a Robotino platform extended by LIDAR sensors. A ROS 2 integration and a pre-configured setup for localization and navigation using existing ROS packages from the Nav2 suite is provided. We validate our setup by comparing simulations with real-world experiments conducted by three Robotinos in a logistics environment in our lab. Additionally, we tested the setup using a ROS 2 hardware driver for the Robotino developed by team GRIPS of the RoboCup Logistics League. The results demonstrate the feasibility of using ROS 2 and Nav2 for navigation tasks on the Robotino platform showing great consistency between simulation and real-world performance. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Mobile robots

Real-time 3D semantic occupancy prediction for autonomous vehicles using memory-efficient sparse convolution

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sze, Samuel Kunze, Lars The Cognitive Robotics Group Oxford Robotics Institute Department of Engineering Science University of Oxford United Kingdom

关键词： Autonomous vehicles

DPFT: Dual Perspective Fusion Transformer for Camera-Radar-based Object Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Fent, Felix Palffy, Andras Caesar, Holger Germany Group Department of Microelectronics Netherlands Section Department of Cognitive Robotics Netherlands

The perception of autonomous vehicles has to be efficient, robust, and cost-effective. However, cameras are not robust against severe weather conditions, lidar sensors are expensive, and the performance of radar-based perception is still inferior to the others. Camera-radar fusion methods have been proposed to address this issue, but these are constrained by the typical sparsity of radar point clouds and often designed for radars without elevation information. We propose a novel camera-radar fusion approach called Dual Perspective Fusion Transformer (DPFT), designed to overcome these limitations. Our method leverages lower-level radar data (the radar cube) instead of the processed point clouds to preserve as much information as possible and employs projections in both the camera and ground planes to effectively use radars with elevation information and simplify the fusion with camera data. As a result, DPFT has demonstrated state-of-the-art performance on the K-Radar dataset while showing remarkable robustness against adverse weather conditions and maintaining a low inference time. The code is made available as open-source software under https://***/TUMFTM/DPFT. © 2024, CC BY.

关键词： Radar target recognition

See Further Than CFAR: a Data-Driven Radar Detector Trained by Lidar

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Roldan, Ignacio Palffy, Andras Kooij, Julian F.P. Gavrila, Dariu M. Fioranelli, Francesco Yarovoy, Alexander Group Department of Microelectronics Netherlands Group Department of Cognitive Robotics Delft University of Technology Delft Netherlands

In this paper, we address the limitations of traditional constant false alarm rate (CFAR) target detectors in automotive radars, particularly in complex urban environments with multiple objects that appear as extended targets. We propose a data-driven radar target detector exploiting a highly efficient 2D CNN backbone inspired by the computer vision domain. Our approach is distinguished by a unique cross-sensor supervision pipeline, enabling it to learn exclusively from unlabeled synchronized radar and lidar data, thus eliminating the need for costly manual object annotations. Using a novel large-scale, real-life multi-sensor dataset recorded in various driving scenarios, we demonstrate that the proposed detector generates dense, lidar-like point clouds, achieving a lower Chamfer distance to the reference lidar point clouds than CFAR detectors. Overall, it significantly outperforms CFAR baselines detection accuracy. Copyright © 2024, The Authors. All rights reserved.

关键词： Deep learning

Evaluation of Motion Comfort using Advanced Active Human Body Models and Efficient Simplified Models

学校读者我要写书评

暂无评论

Evaluation of Motion Comfort using Advanced Active Human Bod...

International Conference on Intelligent Transportation

作者： Raj Desai Marko Cvetkovié Georgios Papaioannou Riender Happee Intelligent-vehicles group Cognitive Robotics (CoR) TU Delft

Active muscles are crucial for maintaining postural stability when seated in a moving vehicle. Advanced active 3D non-linear full body models have been developed for impact and comfort simulation, including large numbers of individual muscle elements, and detailed non-linear models of the joint structures. While such models have an apparent potential to provide insight into postural stabilization, they are computationally demanding, making them less practical in particular for driving comfort where long time periods are to be studied. In vibrational comfort and in general biomechanical research, linearized models are effectively used. This paper evaluates the effectiveness of simplified 3D full-body human models to capture comfort provoked by whole-body vibrations. An efficient seated human body model is developed and validated using experimental data. We evaluate the required complexity in terms of joints and degrees of freedom for the spine, and explore how well linear spring-damper models can approximate reflexive postural stabilization. Results indicate that linear stiffness and damping models can well capture the human response. However, the results are improved by adding proportional integral derivative (PID) and head-in-space (HIS) controllers to maintain the defined initial body posture. The integrator is shown to be essential to prevent drift from the defined posture. The joint angular relative displacement is used as the input reference to each PID controller. With this model, a faster than real-time solution is obtained when used with a simple seat model. The paper also discusses the advantages and disadvantages of various models and provides insight into which models are more appropriate for motion comfort analysis. For designers and researchers in the automotive and seating industries, the findings given in this paper provide useful insights that will help them improve the comfort and safety of both vehicle occupants and seats.

关键词：

SEM-GAT: Explainable Semantic Pose Estimation using Learned Graph Attention

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Panagiotaki, Efimia De Martini, Daniele Pramatarov, Georgi Gadd, Matthew Kunze, Lars Cognitive Robotics Group United Kingdom Mobile Robotics Group Oxford Robotics Institute Department of Engineering Science University of Oxford United Kingdom

This paper proposes a Graph Neural Network (GNN)-based method for exploiting semantics and local geometry to guide the identification of reliable pointcloud registration candidates. Semantic and morphological features of the environment serve as key reference points for registration, enabling accurate lidar-based pose estimation. Our novel lightweight static graph structure informs our attention-based node aggregation network by identifying semantic-instance relationships, acting as an inductive bias to significantly reduce the computational burden of pointcloud registration. By connecting candidate nodes and exploiting cross-graph attention, we identify confidence scores for all potential registration correspondences and estimate the displacement between pointcloud scans. Our pipeline enables introspective analysis of the model’s performance by correlating it with the individual contributions of local structures in the environment, providing valuable insights into the system’s behaviour. We test our method on the KITTI odometry dataset, achieving competitive accuracy compared to benchmark methods and a higher track smoothness while relying on significantly fewer network parameters. Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics