A5013517179- Robot Manipulation and Learning
- Modular Robots and Swarm Intelligence
- Robotics and Sensor-Based Localization
- Animal Behavior and Reproduction
- Computational Drug Discovery Methods
- Metabolomics and Mass Spectrometry Studies
- Handwritten Text Recognition Techniques
- Advanced Graph Neural Networks
- Underwater Vehicles and Communication Systems
- Plant and animal studies
- Distributed Control Multi-Agent Systems
- Neurobiology and Insect Physiology Research
- Machine Learning and Algorithms
- Social Robot Interaction and HRI
- Insect and Arachnid Ecology and Behavior
- Reinforcement Learning in Robotics
- Domain Adaptation and Few-Shot Learning
- Amphibian and Reptile Biology
- Zebrafish Biomedical Research Applications
- earthquake and tectonic studies
- Advanced Database Systems and Queries
- Fish Ecology and Management Studies
- Mathematics, Computing, and Information Processing
- Fish biology, ecology, and behavior
- Image and Object Detection Techniques
Freie Universität Berlin
2011-2025
The University of Tokyo
2021-2023
University of Tsukuba
2023
Vietnam National University Ho Chi Minh City
2023
RIKEN Center for Advanced Intelligence Project
2022
National Institute for Materials Science
2022
Kyoto University
2018-2021
Kyoto University Institute for Chemical Research
2018-2021
Leibniz Institute of Freshwater Ecology and Inland Fisheries
2016-2018
Tokyo University of Agriculture and Technology
2015-2016
Assistive mobile manipulators (AMMs) have the potential to one day serve as surrogates and helpers for people with disabilities, giving them freedom perform tasks such scratching an itch, picking up a cup, or socializing their families.
In recent years, simple biomimetic robots have been increasingly used in biological studies to investigate social behavior, for example collective movement. Nevertheless, a big challenge developing is the acceptance of robotic agents by live animals. this contribution, we describe our advances with regard RoboFish Trinidadian guppies (Poecilia reticulata). We provide detailed technical description system and show effect different appearance, motion patterns interaction modes on artificial...
We present a novel interface for human-robot interaction that enables human to intuitively and unambiguously select 3D location in the world communicate it mobile robot. The points at of interest illuminates (``clicks it'') with an unaltered, off-the-shelf, green laser pointer. robot detects resulting spot omnidirectional, catadioptric camera narrow-band filter. After detection, moves its stereo pan/tilt look this estimates location's position respect robot's frame reference.
We introduce ROS Commander (ROSCo), an open source system that enables expert users to construct, share, and deploy robot behaviors for home robots. A user builds a behavior in the form of Hierarchical Finite State Machine (HFSM) out generic, parameterized building blocks, with real develop test loop. Once constructed, save format direct use robots, or as parts new behaviors. When is deployed, can show where apply relative fiducial markers (AR Tags), which allows quickly become operational...
Responding towards the actions of others is one most important behavioural traits whenever animals same species interact. Mutual influences among interacting individuals may modulate social responsiveness seen and thus make it often difficult to study level individual variation in responsiveness. Here, open-loop biomimetic robots that provide standardized, non-interactive cues can be a useful tool. These are not affected by live animal's but assumed still represent valuable biologically...
Abstract This study aims to investigate the feasibility, usability, and effectiveness of a Retrieval-Augmented Generation (RAG)-powered Patient Information Assistant (PIA) chatbot for pre-CT information counseling compared standard physician consultation informed consent process. prospective comparative included 86 patients scheduled CT imaging between November December 2024. Patients were randomly assigned either PIA group ( n = 43), who received via chat app, or control with doctor-led...
In this work we present a set of integrated methods that enable an RFID-enabled mobile manipulator to approach and grasp object which self-adhesive passive (battery-free) UHF RFID tag has been affixed. Our primary contribution is new mode perception produces images the spatial distribution received signal strength indication (RSSI) for each tagged objects in environment. The intensity pixel 'RSSI image' measured RF particular corresponding direction. We construct these RSSI by panning...
The honeybee dance "language" is one of the most popular examples information transfer in animal world. Today, more than 60 years after its discovery it still remains unknown how follower bees decode contained dance. In order to build a robotic that allows deeper investigation communication process we have recorded hundreds videos waggle dances. this paper analyze statistics visually captured high-precision trajectories European honeybees (Apis mellifera carnica). were produced using novel...
Handheld manipulable objects can often be found on flat surfaces within human environments. Researchers have previously demonstrated that perceptually segmenting a surface from the resting it enable robots to pick and place objects. However, methods for performing this segmentation fail when applied scenes with natural clutter. For example, low-profile dense clutter obscures underlying complicate interpretation of scene. As first step towards characterizing statistics real-world in...
Biomimetic Robots (BRs) are becoming more common in behavioral research and, if they accepted as conspecifics, allow for new forms of experimental manipulations social interactions. Nevertheless, it is often not clear which cues emanating from a BR actually used communicative signals and how species or populations with different sensory make-ups react to specific types BRs. We herein present results experiments using two livebearing fishes that differ their capabilities. In the South Mexico,...
We present and examine a technique for estimating the ego-motion of mobile robot using memory-based learning monocular camera. Unlike other approaches that rely heavily on camera calibration geometry to compute trajectory, our method learns mapping from sparse optical flow platform velocity turn rate. also demonstrate an efficient computing high-quality flow, techniques this as input supervised method. employ voting scheme many learners use subsets cope with variable dimensionality reduce...
Passive UHF RFID tags are well matched to robots' needs. Unlike lowfrequency (LF) and high-frequency (HF) tags, passive readable from across a room, enabling mobile robot efficiently discover locate them. Using tags' unique IDs, semantic database, RF perception via actuated antennas, this paper shows how can reliably interact with people manipulate labeled objects.
This paper presents application of deep learning to recognize online handwritten mathematical symbols. Recently various architectures such as Convolution neural network (CNN), Deep (DNN) and Long short term memory (LSTM) RNN have been applied fields computer vision, speech recognition natural language processing where they shown produce state-of-the-art results on tasks. In this paper, we apply max-out-based CNN BLSTM image patterns created from the original patterns, respectively combine...
Abstract Motivation Recent success in metabolite identification from tandem mass spectra has been led by machine learning, which two stages: mapping to molecular fingerprint vectors and then retrieving candidate molecules the database. In first stage, i.e. prediction, spectrum peaks are features considering their interactions would be reasonable for more accurate of unknown metabolites. Existing approaches prediction based on only individual spectra, without explicitly peak interactions....
Metabolite identification is an important task in metabolomics to enhance the knowledge of biological systems. There have been a number machine learning-based methods proposed for this task, which predict chemical structure given spectrum through intermediate (chemical structure) representation called molecular fingerprints. They usually two steps: (i) predicting fingerprints from spectra; (ii) searching compounds (in database) corresponding predicted Fingerprints are feature vectors, very...
To develop useful drugs and materials, chemists synthesize diverse molecules by trying various reactants reaction routes. Toward automating this process, we propose a deep generative model, called cascaded variational autoencoder (casVAE), for synthesizable molecular design. It generates tree, where the are chosen from commercially available compounds synthesis route is constructed as tree of templates. The first part casVAE designed to generate molecule surrogate product, while second...
This paper presents deep learning to recognize online handwritten mathematical symbols. Recently various architectures such as Convolution neural networks (CNNs), Deep (DNNs), Recurrent (RNNs) and Long short-term memory (LSTM) RNNs have been applied fields computer vision, speech recognition natural language processing where they shown superior performance state-of-the-art methods on tasks. In this paper, max-out-based CNNs Bidirectional LSTM (BLSTM) are image patterns created from the...
Mormyrid weakly electric fish produce organ discharges (EODs) for active electrolocation and electrocommunication. These pulses are emitted with variable interdischarge intervals (IDIs) resulting in temporal discharge patterns interactive signaling episodes nearby conspecifics. However, unequivocal assignment of to a specific behavioral context has proven be challenging. Using an ethorobotical approach, we confronted single individuals Mormyrus rume proboscirostris mobile robot capable...