We present an approach for estimating a mobile robot's pose w.r.t. the allocentric coordinates of a network of static cameras using multi-view RGB images. The images are processed online, locally on smart edge sensors by deep neural networks to detect the robot and estimate 2D keypoints defined at distinctive positions of the 3D robot model. Robot keypoint detections are synchronized and fused on a central backend, where the robot's pose is estimated via multi-view minimization of reprojection errors. Through the pose estimation from external cameras, the robot's localization can be initialized in an allocentric map from a completely unknown state (kidnapped robot problem) and robustly tracked over time. We conduct a series of experiments evaluating the accuracy and robustness of the camera-based pose estimation compared to the robot's internal navigation stack, showing that our camera-based method achieves pose errors below 3 cm and 1° and does not drift over time, as the robot is localized allocentrically. With the robot's pose precisely estimated, its observations can be fused into the allocentric scene model. We show a real-world application, where observations from mobile robot and static smart edge sensors are fused to collaboratively build a 3D semantic map of a ~240 m2indoor environment.

Haptic perception is highly important for immersive teleoperation of robots, especially for accomplishing manipulation tasks. We propose a low-cost haptic sensing and rendering system, which is capable of detecting and displaying surface roughness. As the robot fingertip moves across a surface of interest, two microphones capture sound coupled directly through the fingertip and through the air, respectively. A learning-based detector system analyzes the data in real time and gives roughness estimates with both high temporal resolution and low latency. Finally, an audio-based vibrational actuator displays the result to the human operator. We demonstrate the effectiveness of our system through lab experiments and our winning entry in the ANA Avatar XPRIZE competition finals, where briefly trained judges solved a roughness-based selection task even without additional vision feedback. We publish our dataset used for training and evaluation together with our trained models to enable reproducibility of results.

One-shot action recognition allows the recognition of human-performed actions with only a single training example. This can influence human-robot-interaction positively by enabling the robot to react to previously unseen behaviour. We formulate the one-shot action recognition problem as a deep metric learning problem and propose a novel image-based skeleton representation that performs well in a metric learning setting. Therefore, we train a model that projects the image representations into an embedding space. In embedding space similar actions have a low euclidean distance while dissimilar actions have a higher distance. The one-shot action recognition problem becomes a nearest-neighbor search in a set of activity reference samples. We evaluate the performance of our proposed representation against a variety of other skeleton-based image representations. In addition we present an ablation study that shows the influence of different embedding vector sizes, losses and augmentation. Our approach lifts the state-of-the-art by 3.3% for the one-shot action recognition protocol on the NTU RGB+D 120 dataset under a comparable training setup. With additional augmentation our result improved over 7.7%.

We propose an approach for action segmentation by representing motions as images. A transformer object detection network is used to segment the sequences from the representation images. We examine different encoding approaches, normalization strategies and skeleton joint orders in an extensive experiment study. Our approach is evaluated on skeleton sequences from the PKU-MMD dataset. We successfully apply transformer networks for action segmentation on skeleton sequences. Our proposed approach achieves high class accuracies while start and end-time estimation of the action segments are subject to further improvement.

Recognizing an activity with a single reference sample using metric learning approaches is a promising research field. The majority of few-shot methods focus on object recognition or face-identification. We propose a metric learning approach to reduce the action recognition problem to a nearest neighbor search in embedding space. We encode signals into images and extract features using a deep residual CNN. Using triplet loss, we learn a feature embedding. The resulting encoder transforms features into an embedding space in which closer distances encode similar actions while higher distances encode different actions. Our approach is based on a signal level formulation and remains flexible across a variety of modalities. It further outperforms the baseline on the large scale NTU RGB+D 120 dataset for the one-shot action recognition protocol by 5.6%. With just 60% of the training data, our approach still outperforms the baseline approach by 3.7%. With 40% of the training data, our approach performs comparably well to the second follow up. Further, we show that our approach generalizes well in experiments on the UTD-MHAD dataset for inertial, skeleton and fused data and the Simitate dataset for motion capturing data. Furthermore, our inter-joint and inter-sensor experiments suggest good capabilities on previously unseen setups.

We present a simple, yet effective and flexible method for action recognition supporting multiple sensor modalities. Multivariate signal sequences are encoded in an image and are then classified using a recently proposed EfficientNet CNN architecture. Our focus was to find an approach that generalizes well across different sensor modalities without specific adaptions while still achieving good results. We apply our method to 4 action recognition datasets containing skeleton sequences, inertial and motion capturing measurements as well as Wi-Fi fingerprints that range up to 120 action classes. Our method defines the current best CNN-based approach on the NTU RGB+D 120 dataset, lifts the state of the art on the ARIL Wi-Fi dataset by +6.8%, improves the UTD-MHAD inertial baseline by +14.4%, the UTD-MHAD skeleton baseline by +0.5% and achieves 96.1% on the Simitate motion capturing data (80/20 split). We further demonstrate experiments on both, modality fusion on a signal level and signal reduction to prevent the representation from overloading.

We present Simitate - a hybrid benchmarking suite targeting the evaluation of approaches for imitation learning. A dataset containing 1938 sequences where humans perform daily activities in a realistic environment is presented. The dataset is strongly coupled with an integration into a simulator. RGB and depth streams with a resolution of 960×540 at 30Hz and accurate ground truth poses for the demonstrator's hand, as well as the object in 6 DOF at 120Hz are provided. Along with our dataset we provide the 3D model of the used environment, labeled object images and pre-trained models. A benchmarking suite that aims at fostering comparability and reproducibility supports the development of imitation learning approaches. Further, we propose and integrate evaluation metrics on assessing the quality of effect and trajectory of the imitation performed in simulation. Simitate is available on our project website: https://agas.uni-koblenz.de/simitate/.

The support through robots in orthopaedic rehabilitation is an opportunity to relieve physiotherapists. However, to be able to provide a control in the robot-patient cooperation in the therapy process a certain standard in interpreting the exercise has to be established. In this paper we present an evaluation approach of the health subject performance in a tibiofemoral rehabilitation on the example of squat exercises. The proposed method utilizes only depth images for the performance evaluation and any human-robot interaction system for the performance correction. Thus, this method can be easily applied to a mobile service robot in the robot-aided physical therapy. The patient is observed while performing the exercise and the motion is evaluated and segmented using Motion History Images. Concrete, depth images are used to monitor local points of interest on the performer during the exercise. The proposed approach was evaluated on custom image sequences with a multitude of varying subjects and shows the suitable performance for assisting in the correctness of the exercise execution.

In this paper, we propose an integrated system to detect and track a single operator that can switch off and on when it leaves and (re-)enters the scene. Our method is based on a set-valued Bayes-optimal state estimator that integrates RGB-D detections and image-based classification to improve tracking results in severe clutter and under long-term occlusion. The classifier is trained in two stages: First, we train a deep convolutional neural network to obtain a feature representation for person re-identification. Then, we bootstrap a classifier that discriminates the operator from remaining people on the output of the state-estimator. We evaluate the approach on a publicly available multi-target tracking dataset as well as custom datasets that are specific to our problem formulation. Experimental results suggest reliable tracking accuracy in crowded scenes and robust re-detection after long-term occlusion.

Sensors for RGB-D data have gained high popularity in the computer vision community. We present an efficient ensemble classifier that combines visual and depth data and achieves higher recognition rates than the individual classifiers or a classifier exploiting visual and depth data at the same time. The presented approach was evaluated in practice on a mobile robot during the RoCKIn robotics challenge in 2014.

In this paper we present the approaches and contributions of team homer@UniKoblenz that were developed for and applied during the RoboCup@Home competitions. In particular, we highlight the different abstraction layers of our software architecture that allows for rapid application development based on the ROS actionlib. This architectural design enables us to focus on the development of new algorithms and approaches and significantly helped us in winning the RoboCup@Home competition in 2015. We further give an outlook on recently published open-source software for service robots that can be downloaded from our ROS package repository on http://wiki.ros.org/agas-ros-pkg .