检索结果-内蒙古大学图书馆

Odometry-Less Indoor Dynamic Object Detection and Localization with Spatial Anticipation

学校读者我要写书评

暂无评论

Odometry-Less Indoor Dynamic Object Detection and Localizati...

IEEE/SICE International Symposium on System Integration

作者： Zhengcheng Shen Yi Gao Linh Kästner Jens Lambrecht Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

The continually evolving image segmentation methods in computer vision can further broaden the cognitive abilities of the robot. As humans, we won't judge if the object is moving by accuracy speed estimation but through semantic information based on visual input. With the existing video segmentation method, the robot can pay more attention to the foreground objects, which are more important for path planning and navigation in most cases. The proposed methods try to deploy the first-person view segmentation method with an RGBD sensor and get the location information and locally egocentric map. The error of the segmentation methods will cause unexpected objects in the egocentric map, and the self-occlusion will influence the location and distance estimation. To further improve the methods, a spatial anticipate unit is embedded into the framework. We improve 5.6% the localization accuracy(RMSE), 12.9% the distance estimation(MAE) with the trained object, and 28.9% the recall for the trained object. The methods also show the margin on the untrained object dataset for similar objects.

关键词：

Towards Real-Time Motion Planning for Industrial Robots in Collaborative Environments

学校读者我要写书评

暂无评论

Towards Real-Time Motion Planning for Industrial Robots in C...

Annual Conference of Industrial Electronics Society

作者： Teham Bhuiyan Benno Kutschank Karim Prüter Huy Flach Linh Kästner Jens Lambrecht Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

In collaborative environments, real-time motion planning is crucial for industrial robots to navigate safely and efficiently. Traditional planning algorithms, such as Rapidly-exploring Random Trees (RRT) or Probabilistic Roadmaps (PRM), often face challenges in coping with dynamic environments due to their inherent computational complexity. To address this issue, we propose an approach based on Deep Reinforcement Learning (DRL) for real-time motion planning of industrial robots. Our method leverages the power of machine learning and neural networks to enable robots to make intelligent decisions in real-time, ensuring prompt and adaptive navigation. However, applying DRL to industrial robots poses unique challenges, as vision-based training is difficult and distance sensors commonly used in mobile robots are unavailable. To overcome these challenges, we employ depth cameras to generate distance information and convert the obtained point cloud into voxels using the Open3D library. The obstacles are then loaded into the simulation environment in real-time, allowing the agent to perceive and react to the dynamic environment. To achieve a low simulation-to-real-gap, we propose a hardware-in-the-loop (HIL) approach, where the real robot mimics the movements of the simulated robot. We demonstrate the effectiveness of our system through real-world experiments. Our code is available on GitHub [1].

关键词：

MIMO Virtual Array Design for mmWave 4D-Imaging Radar Sensors 31

学校读者我要写书评

暂无评论

MIMO Virtual Array Design for mmWave 4D-Imaging Radar Sensor...

31st European Signal Processing Conference, EUSIPCO 2023

作者： Sichani, Nazila Karimian Alaee-Kerahroodi, Mohammad Raei, Ehsan Bhavani Shankar, M.R. Mehrshahi, Esfandiar Ghorashi, Seyyed Ali Department of Telecommunications Faculty of Electrical Engineering Shahid Beheshti University Tehran*** Iran University of Luxembourg Luxembourg Department of Computer Science and Digital Technologies School of Architecture Computing and Engineering University of East London LondonE16 2RD United Kingdom

ISBN: (纸本)9789464593600

The recently developed 4D-Imaging mmWave MIMO radars have significant advantages over conventional radar sensors because they offer a large virtual array in both azimuth and elevation. However, the physical placement of transmit and receive antennas to achieve the desired virtual array, while considering cost and efficiency is not intuitive. In this paper, we propose an optimization framework for optimally placing the transmit and receive array physical elements. The proposed framework obtains a sequence of optimal placements based on the coordinate descent by defining a desired virtual array and number of transmit and receive antenna elements as the objective. The objective function is then monotonically minimized in each iteration of the proposed algorithm, converging to a solution that guarantees the desired number of transmit and receive antenna elements, while the obtained virtual array is as close to the desired virtual array as possible. The simulation results are validated by simulating the virtual array of multiple commercially available 4D-Imaging radar products. © 2023 European Signal Processing Conference, EUSIPCO. All rights reserved.

关键词： Radar imaging

Deep-Reinforcement-Learning-based Path Planning for Industrial Robots using Distance Sensors as Observation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Bhuiyan, Teham Kästner, Linh Hu, Yifan Kutschank, Benno Lambrecht, Jens Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

Industrial robots are widely used in various manufacturing environments due to their efficiency in doing repetitive tasks such as assembly or welding. A common problem for these applications is to reach a destination without colliding with obstacles or other robot arms. Most commonly used sampling-based path planning approaches such as RRT require long computation times, especially in complex environments. Furthermore, the environment in which they are employed needs to be known beforehand. When utilizing the approaches in new environments, a tedious engineering effort in setting hyperparameters needs to be conducted, which is time- and cost-intensive. On the other hand, Deep Reinforcement Learning has shown remarkable results in dealing with unknown environments, generalizing new problem instances, and solving motion planning problems efficiently. On that account, this paper proposes a Deep-Reinforcement-Learning-based motion planner for robotic manipulators. We evaluated our model against state-of-the-art sampling-based planners in several experiments. The results show the superiority of our planner in terms of path length and execution time. Copyright © 2023, The Authors. All rights reserved.

关键词： Reinforcement learning

Integrative object and pose to task detection for an augmented-reality-based human assistance system using neural networks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Kästner, Linh Eversberg, Leon Mursa, Marina Lambrecht, Jens Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

As a result of an increasingly automatized and digitized industry, processes are becoming more complex. Augmented Reality has shown considerable potential in assisting workers with complex tasks by enhancing user understanding and experience with spatial information. However, the acceptance and integration of AR into industrial processes is still limited due to the lack of established methods and tedious integration efforts. Meanwhile, deep neural networks have achieved remarkable results in computer vision tasks and bear great prospects to enrich Augmented Reality applications . In this paper, we propose an Augmented-Reality-based human assistance system to assist workers in complex manual tasks where we incorporate deep neural networks for computer vision tasks. More specifically, we combine Augmented Reality with object and action detectors to make workflows more intuitive and flexible. To evaluate our system in terms of user acceptance and efficiency, we conducted several user studies. We found a significant reduction in time to task completion in untrained workers and a decrease in error rate. Furthermore, we investigated the users learning curve with our assistance system. Copyright © 2020, The Authors. All rights reserved.

关键词： Augmented reality

Robust Few-Shot Pose Estimation of Articulated Robots using Monocular Cameras and Deep-Learning-based Keypoint Detection

学校读者我要写书评

暂无评论

Robust Few-Shot Pose Estimation of Articulated Robots using ...

International Conference on Robot Intelligence Technology and Applications (RiTA)

作者： Jens Lambrecht Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Technische Universitt Berlin Berlin Germany

Camera-based pose estimation is a necessity for flexible applications in robotics, especially interaction between robots and mobile entities. Inspired by recent advancements in human pose estimation based on Convolutional Neural Networks, it is aspired to substitute the usage of artificial marker by automatically detecting inherent keypoints of the robot representing its 2D skeleton model. In addition, current encoder readings of the robot are utilized establishing the corresponding 3D skeleton model through forward kinematics. With the help of these 2D - 3D point correspondences, an estimation of the translation and orientation deviation between robot and camera is derived solving the perspective-n-point problem. An adequate approach for markerless keypoint detection of an UR5 robot is presented and evaluated in terms of precision and pose dispersion considering a dynamically moving robot. The promising results show that the novel method works robustly and reliably as a few-shot approach and copes with false positives as well as with partly occlusions and non-detected keypoints. Further potential is identified regarding enhancing the accuracy through the use of synthetic data.

关键词： Robot kinematics Cameras Pose estimation Robot vision systems Three-dimensional displays Two dimensional displays

Spatial imagination with semantic cognition for mobile robots

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Shen, Zhengcheng Kästner, Linh Lambrecht, Jens The Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

The imagination of the surrounding environment based on experience and semantic cognition has great potential to extend the limited observations and provide more information for mapping, collision avoidance, and path planning. This paper provides a training-based algorithm for mobile robots to perform spatial imagination based on semantic cognition and evaluates the proposed method for the mapping task. We utilize a photo-realistic simulation environment, Habitat, for training and evaluation. The trained model is composed of Resent-18 as encoder and Unet as the backbone. We demonstrate that the algorithm can perform imagination for unseen parts of the object universally, by recalling the images and experience and compare our approach with traditional semantic mapping methods. It is found that our approach will improve the efficiency and accuracy of semantic mapping. © 2021, CC BY-NC-ND.

关键词： Mobile robots

Holistic Deep-Reinforcement-Learning-based Training for Autonomous Navigation in Crowded Environments

学校读者我要写书评

暂无评论

Holistic Deep-Reinforcement-Learning-based Training for Auto...

IEEE/ASME (AIM) International Conference on Advanced Intelligent Mechatronics

作者： Linh Kästner Marvin Meusel Teham Bhuiyan Jens Lambrecht Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

In recent years, Deep Reinforcement Learning emerged as a promising approach for autonomous navigation of robots and has been utilized in various areas of navigation such as obstacle avoidance, motion planning, or decision making in crowded environments. However, most research works either focus on providing an end-to-end solution training the whole system using Deep Reinforcement Learning or focus on one specific aspect such as local motion planning. This however, comes along with a number of problems such as catastrophic forgetfulness, inefficient navigation behavior, and non-optimal synchronization between different entities of the navigation stack. In this paper, we propose a holistic Deep Reinforcement Learning training approach in which the training procedure is involving all entities of the navigation stack. This should enhance the synchronization between- and understanding of all entities of the navigation stack and as a result, improve navigational performance in crowded environments. We trained several agents with a number of different observation spaces to study the impact of different input on the navigation behavior of the agent. In profound evaluations against multiple learning-based and classic model-based navigation approaches, our proposed agent could outperform the baselines in terms of efficiency and safety attaining shorter path lengths, less roundabout paths, and less collisions especially in situations with a high number of pedestrians.

关键词：

Connecting deep-reinforcement-learning-based obstacle avoidance with conventional global planners using waypoint generators

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Kästner, Linh Buiyan, Teham Zhao, Xinlin Shen, Zhengcheng Lambrecht, Jens Marx, Cornelius Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

Deep Reinforcement Learning has emerged as an efficient dynamic obstacle avoidance method in highly dynamic environments. It has the potential to replace overly conservative or inefficient navigation approaches. However, the integration of Deep Reinforcement Learning into existing navigation systems is still an open frontier due to the myopic nature of Deep-Reinforcement-Learning-based navigation, which hinders its widespread integration into current navigation systems. In this paper, we propose the concept of an intermediate planner to interconnect novel Deep-Reinforcement-Learning-based obstacle avoidance with conventional global planning methods using waypoint generation. Therefore, we integrate different waypoint generators into existing navigation systems and compare the joint system against traditional ones. We found an increased performance in terms of safety, efficiency and path smoothness especially in highly dynamic environments. © 2021, CC BY-NC-ND.

关键词： Deep learning