3 results
Now showing1 - 3 of 3
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Comparative analysis of local trajectory planning algorithms in ROS2(IEEE, 2025-04); ; Local path planners are essential for real-time obstacle avoidance, particularly in unpredictable industrial environments. This study compares three ROS 2 local path-planning algorithms: Dynamic Window Approach (DWB), Model Predictive Path Integral (MPPI), and Regulated Pure Pursuit (RPP). These planners are part of the Robot Operating System 2 (ROS 2) navigation stack (Nav2), with DWB and MPPI as built-in controllers, and RPP integrated as a controller option.While previous studies have focused on simulations or simple real-world environments with static obstacles, few have examined dynamic and unpredictable scenarios. To address this, experiments were conducted with a Clearpath Jackal robot in environments with static and dynamic obstacles. Key metrics, including distance to obstacles, frequency of recovery behaviors, and path smoothness, were used to compare the planners performance in both real-world and simulated settings.MPPI offers a strong balance across all metrics, making it ideal for dynamic environments requiring both safety and efficiency. DWB excels in fast navigation, though with closer proximity to obstacles, while RPP produces smooth paths but struggles in highly dynamic environments. These results provide valuable insights for selecting the right local path planner for autonomous robots in industrial applications like manufacturing and warehousing.Publicationtype: Conference Paper1037 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Analysing visual-inertial odometry algorithms for the localization of industrial autonomous mobile robots in intralogistics and manufacturing(Wissenschaftliche Gesellschaft für Technische Logistik e.V. (WGTL), 2024-09-26); ; ; The use of Autonomous Mobile Robots (AMR) plays asignificant role inthe automation of intralogisticsprocesses. For safe operation and navigation, high locali-zation accuracy is required. Common AMR systems rely oncost-intensive sensors such as LIDAR scanners. To en-able widespread use of AMRs the industry alternative so-lutions are required. This study explores stereo camera-based visual SLAM as a cost-effective alternative to con-ventional 3D LIDAR-based localization solutions for an industrial robot application. Using Stereolabs ZED 2I and Intel RealSense D455 cameras with ORB-SLAM3 and OpenVINS algorithms, we evaluated Mean Absolute Pose Error (APE) and Root Mean Square Pose Error (RPE). The highest accuracy was achieved with the ZED 2I with OpenVINS with an APE of 0.17m and an RPE of 0.02m while the use of a RealSense D455 showed an APE of 0.33m with an RPE of 0.02mPublicationtype: Journal Article75 37 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Estimating the Pose of a Euro Pallet with an RGB Camera based on Synthetic Training DataE stimating the pose of a pallet and other logistics objects is crucial for various use cases, such as automatized material handling or tracking. Innovations in computer vision, computing power, and machine learning open up new opportunities for device-free localization based on cameras and neural networks. Large image datasets with annotated poses are required for training the network. Manual annotation, especially of 6D poses, is an extremely labor-intensive process. Hence, newer approaches often leverage synthetic training data to automatize the process of generating annotated image datasets. In this work, the generation of synthetic training data for 6D pose estimation of pallets is presented. The data is then used to train the Deep Object Pose Estimation (DOPE) algorithm [1]. The experimental validation of the algorithm proves that the 6D pose estimation of a standardized Euro pallet with a Red-Green-Blue (RGB) camera is feasible. The comparison of the results from three varying datasets under different lighting conditions shows the relevance of an appropriate dataset design to achieve an accurate and robust localization. The quantitative evaluation shows an average position error of less than 20 cm for the preferred dataset. The validated training dataset and a photorealistic model of a Euro pallet are publicly provided [2].Publicationtype: Journal Article97