Publications
Journals & Conference Proceedings
Vlachos, K., & Doulgeri, Z. (2026). A novel interconnection of Dynamic Movement Primitives (DMP) with a modified low impedance controlled robot for accurate tracking and compliant robot reactions. 2026 European Control Conference (ECC), Reykjavik, Iceland, 7–10 July 2026
Abstract: In this paper, a novel control scheme is proposed interconnecting a Dynamic Movement Primitives (DMP) system with a modified low impedance controlled robot to achieve compliance to unexpected contact events and high tracking accuracy under the presence of model and task uncertainties. The proposed control scheme is formulated in the joint space and theoretically shown to accurately track the desired trajectory under the presence of uncertainties. Simulations of 2-DOF manipulator model carrying an unknown load under the proposed scheme demonstrate its high tracking accuracy as compared to the low impedance controlled robot while exhibiting highly compliant reactions to external contact forces.
Yagüe, M. P., García, J. E. S., & Penas, M. S. (2026). Human-intelligent trajectory optimization for robotic manipulators with hybrid PSO-PS algorithm. Advanced Engineering Informatics, 69, 103941.
Abstract: Industry 5.0 is driving a new era in industrial automation, where the collaboration between artificial intelligence (AI) and human supervision enables the development of smarter, more adaptive, and more efficient systems. Robotic trajectory generation is a clear example of this new paradigm. Metaheuristic techniques help automatically generate optimized trajectories, thereby improving operational efficiency. However, optimizing trajectories using AI alone also presents limitations. Starting from random trajectories, the optimization process becomes computationally expensive, especially in complex environments. In this context, initial input from human experts plays a crucial role: expert-defined trajectories provide structured, feasible, and contextual starting points that guide AI more effectively toward high-quality solutions. Therefore, this work proposes a novel human-guided trajectory optimization algorithm. In this way, human knowledge, Particle Swarm Optimization (PSO), and Pattern Search (PS) are efficiently combined. The results demonstrate that this approach significantly improves robotic system performance, achieving cycle time reductions of up to 16.69% compared to expert-defined trajectories. This approach establishes a solid framework for intelligent automation in Industry 5.0, promoting the development of more efficient, sustainable, and adaptive robotic systems.
Peñacoba-Yagüe, M., & Sierra-García, J. E. (2026). Generalized Design Methodology for Dual-Arm Robotic Platforms: From Conceptualization to Experimental Validation Within the MANiBOT Framework. Machines, 14(1), 74.
Abstract: This work proposes a general methodology for the design and experimental validation of dual-arm robotic platforms intended for intelligent manipulation tasks in real-world environments. The proposed framework formalizes the complete engineering process, from the definition of functional requirements to the structural validation of the final prototype, ensuring reproducibility and adaptability across different applications. The methodology is organized into five main stages: (i) requirement analysis and context characterization; (ii) conceptual architecture definition; (iii) detailed mechanical design and structural analysis; (iv) prototype construction and integration; and (v) experimental validation and iterative refinement. Each stage defines its expected deliverables, evaluation metrics, and decision criteria to support systematic design progression. The approach is demonstrated through its implementation within the European project MANiBOT, where the framework guided the development of a modular bimanual robotic platform capable of integrating collaborative manipulators and conveyor subsystems for dual-arm manipulation. Structural testing, deflection measurements, and stability analyses confirmed the robustness and safety of the resulting design. Beyond this specific case, the proposed methodology provides a replicable and extensible design reference for research and industrial teams developing modular robotic structures, supporting the standardization of engineering practices in bimanual mobile robotics.
M. Kiatos, L. Koutras, I. Sarantopoulos and Z. Doulgeri, “Learning a Pre-Grasp Manipulation Policy to Effectively Retrieve a Target in Dense Clutter,” 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Abu Dhabi, United Arab Emirates, 2024, pp. 7543-7549.
Abstract: Robotic grasping of a target object in cluttered environments poses considerable challenges, often due to limited collision-free grasp affordances caused by the close proximity of other objects. To overcome this limitation, non-prehensile actions like pushing can be strategically employed to manipulate the environment and improve the chances of successful grasps. In this paper, we introduce a novel pre-grasp manipulation policy designed to efficiently retrieve a target object from dense clutter by leveraging pushing actions and considering the gripper’s kinematic capabilities to strategically position the target object within the gripper’s closing region for a secure grasp. Unlike conventional approaches, our policy incorporates sequential pushing, allowing the robot to make decisions while within the camera’s field of view without retracting to a home position, leading to significantly reduced execution time per action. Our policy, trained in simulation, seamlessly transfers to real-world scenarios. Extensive experimental evaluation demonstrates superior performance, faster completion times, and robust generalization to unseen objects compared to existing baselines.
L. Koutras, I. Ntoliou and Z. Doulgeri, “Towards Passivity Based Nonprehensile Bimanual Manipulation of Large Objects,” 2023 IEEE-RAS 22nd International Conference on Humanoid Robots (Humanoids), Austin, TX, USA, 2023, pp. 1-8.
Abstract: In this work, a passivity-based control scheme to manipulate large objects supported on planar surfaces by bimanual robots is introduced. A control input is designed to accomplish the rotation and displacement of the object to a desired pose without assuming any knowledge of the object dynamics and friction dynamics between the object and the supporting surface. Passivity arguments are used to rigorously prove its stability. The proposed method is evaluated in simulations with friction forces simulated using the elastoplastic model. Experimental results performed on a bimanual robot with two 7-dof KUKA LBR iiwa 7 arms manipulating a variety of large objects verify the theoretical results.
L. Koutras, S. Stavridis, C. Papakonstantinou and Z. Doulgeri, “Robotic Shelf Replenishment by Combining Non-Prehensile Object Manipulation with Simple Grasping,” 2025 IEEE-RAS 24th International Conference on Humanoid Robots (Humanoids), Seoul, Korea, Republic of, 2025, pp. 1-8.
Abstract: In this work, the problem of robotic shelf replenishment is being studied. Such tasks involve a variety of object types and geometries, which should be picked from boxes where they are tightly packed and placed on shelves in tight formations with the appropriate orientation. We consider a bimanual robotic set up with a parallel finger gripper and a 3Dprinted end-effector and propose to combine simple grasping with a set of non-prehensile manipulations to achieve such a replenishment task. This work reports on the details of the implementation of the proposed strategy and on the initial investigation of the feasibility and effectiveness of the proposed solution for a representative set of super-market products, demonstrated in a number of experiments in the lab.
S. Jauhri, I. Lunawat, and G. Chalvatzaki, “Learning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Rendering,” Robotics: Science and Systems 2024, Delft, Netherlands, July 15-July 19, 2024.
Abstract: A significant challenge for real-world robotic manipulation is the effective 6DoF grasping of objects in cluttered scenes from any single viewpoint without needing additional scene exploration. This work re-interprets grasping as rendering and introduces NeuGraspNet, a novel method for 6DoF grasp detection that leverages advances in neural volumetric representations and surface rendering. We encode the interaction between a robot’s end-effector and an object’s surface by jointly learning to render the local object surface and learning grasping functions in a shared feature space. Our approach uses global (scene-level) features for grasp generation and local (grasp-level) neural surface features for grasp evaluation. This enables effective, fully implicit 6DoF grasp quality prediction, even in partially observed scenes. NeuGraspNet operates on random viewpoints, common in mobile manipulation scenarios, and outperforms existing implicit and semi-implicit grasping methods. We demonstrate the real-world applicability of the method with a mobile manipulator robot, grasping in open cluttered spaces.
L. N. Bikas and G. A. Rovithakis, “Discrete-time Prescribed Performance Control and Maximum Allowable Transmission Interval,” 2024 European Control Conference (ECC), Stockholm, Sweden, 2024, pp. 1159-1164.
Abstract: In this paper, we consider a discrete-time implementation of prescribed performance control (PPC), focusing on its robustness and operability. Specifically, given a prescribed performance controller that guarantees prescribed performance attributes, in terms of maximum overshoot, minimum convergence rate and maximum steady-state error, when operating in continuous-time, the task is to derive sufficient conditions on the maximum allowable transmission interval to enable PPC to preserve its performance characteristics. Interestingly, the maximum allowable transmission interval is directly related with the performance achieved at steady-state. Simulations clarify and verify the theoretical findings.
T. A. Aforozi and G. A. Rovithakis, “Event-Triggered Prescribed Performance Control for SISO Uncertain Nonlinear Systems in Brunovsky Canonical Form,” 2024 European Control Conference (ECC), Stockholm, Sweden, 2024, pp. 1177-1182.
Abstract: In this work, we consider the problem of designing tracking controllers for SISO uncertain high relative degree systems in Brunovsky canonical form in the presence of non-periodic communication. The proposed control scheme is static, and requires no hard calculations, analytic or numerical, to produce the control signal. Event-triggered mechanisms are considered in both sensor-to-controller and controller-to-actuator channels, yet the enforcement of prescribed performance bounds in terms of steady-state accuracy and convergence rate is ensured. No prior knowledge or estimation structure regarding system nonlinearities are required and no high-order derivatives of the desired output trajectories are incorporated in the controller design. Simulation results clarify and verify the theoretical findings.
V. Prasad, A. Kshirsagar, D. Koert, R. Stock-Homburg, J. Peters and G. Chalvatzaki, “MoVEInt: Mixture of Variational Experts for Learning Human–Robot Interactions From Demonstrations,” in IEEE Robotics and Automation Letters, vol. 9, no. 7, pp. 6043-6050, July 2024.
Abstract: Shared dynamics models are important for capturing the complexity and variability inherent in Human-Robot Interaction (HRI). Therefore, learning such shared dynamics models can enhance coordination and adaptability to enable successful reactive interactions with a human partner. In this work, we propose a novel approach for learning a shared latent space representation for HRIs from demonstrations in a Mixture of Experts fashion for reactively generating robot actions from human observations. We train a Variational Autoencoder (VAE) to learn robot motions regularized using an informative latent space prior that captures the multimodality of the human observations via a Mixture Density Network (MDN). We show how our formulation derives from a Gaussian Mixture Regression formulation that is typically used approaches for learning HRI from demonstrations such as using an HMM/GMM for learning a joint distribution over the actions of the human and the robot. We further incorporate an additional regularization to prevent “mode collapse”, a common phenomenon when using latent space mixture models with VAEs. We find that our approach of using an informative MDN prior from human observations for a VAE generates more accurate robot motions compared to previous HMM-based or recurrent approaches of learning shared latent representations, which we validate on various HRI datasets involving interactions such as handshakes, fistbumps, waving, and handovers. Further experiments in a real-world human-to-robot handover scenario show the efficacy of our approach for generating successful interactions with four different human interaction partners.
Abstract: This work introduces BioTacTip, a soft biomimetic optical tactile sensor inspired by the interlocking structure of the epidermal–dermal boundary in human skin. The sensor uses a structured fingertip design and an internal camera to capture tactile images that encode contact position, depth, contact geometry, and force direction. The approach enables direct use of tactile images without a separate marker-detection step, supporting efficient reconstruction of 3D external forces and contact information. Indentation and press-and-shear experiments confirm accurate localization and force estimation, making the sensor suitable for dexterous robotic manipulation tasks.
Abstract: This review provides a structured classification of vision-based tactile sensors, a rapidly growing class of sensors used in robotic hands, grippers, and prosthetic devices. It distinguishes between marker-based and intensity-based transduction principles, showing how different sensor designs convert physical contact into tactile images. The paper compares the hardware characteristics, sensing mechanisms, multimodal combinations, and common interpretation methods used in these sensors. It also highlights current challenges and future research directions, offering a reference framework for the design and selection of vision-based tactile sensing technologies.
Abstract: This work addresses the need for reliable, real-time human detection in collaborative robotics, where robots must operate safely near people. It proposes a computationally efficient transformer-based deep learning model that uses capacitive proximity sensing to distinguish human presence from other nearby objects. By combining proximity sensors with a lightweight transformer architecture, the method supports safer and more responsive human–robot interaction. The approach is designed for real-time deployment in collaborative environments where fast and accurate human proximity detection is essential.
Abstract: This paper investigates how the definition of a fitness function affects the offline optimization of trajectories for automated guided vehicles and industrial robots operating in environments with static obstacles. The work focuses on industrial contexts where trajectory stability, predictability, and timing are critical for reliable operation. The main contribution is an integrated process for defining effective fitness functions that guide optimization toward feasible and high-quality trajectories. The results support better trajectory generation strategies for automated guided vehicle applications by balancing geometric, physical, and operational requirements.
Abstract: This paper presents a tactile control scheme for grasping and manipulating delicate objects with an underactuated anthropomorphic SoftHand equipped with biomimetic tactile sensors on all five fingertips. These sensors estimate contact geometry and force, enabling the hand to regulate its grasp while interacting with objects. The framework processes tactile images from multiple fingertips and uses force and pose feedback to respond to disturbances during manipulation. Experiments involving flexible objects, pouring, and tactile leader–follower manipulation demonstrate more responsive, human-like grasp control for underactuated robotic hands.
Abstract: This work introduces the Tactile SoftHand-A, a low-cost, 3D-printed, highly underactuated anthropomorphic robotic hand with tactile sensing and an antagonistic tendon mechanism. The design enables active opening and closing while keeping the number of actuators low, supporting adaptable and human-like grasping. The paper also integrates 3D-printed vision-based tactile sensors into the fingers and uses tactile feedback to detect contact location and slip. Experiments show improved reactivity, load-bearing capacity, and manipulation ability, while the open-source design provides a reusable basis for accessible tactile robotic hands.
Abstract: This paper introduces ActPerMoMa, an active-perception motion generation pipeline for mobile manipulators operating in unknown and cluttered scenes. The method plans robot movements that improve visual understanding of the environment while also supporting manipulation tasks such as grasping. The approach samples candidate paths in a receding-horizon manner and evaluates them using utilities that balance visual information gain, scene reconstruction, and grasp reachability. Simulated experiments with a dual-arm TIAGo++ robot and real-world transfer tests show the potential of active-perceptive planning for mobile manipulation.
Abstract: This paper presents the mechanical design and electromechanical integration of a mobile bimanual robotic platform developed within the MANiBOT project. The platform is intended for real and unstructured environments such as supermarkets and airports, where stable, safe, and adaptive manipulation is required. The work describes the functional requirements, component selection, structural frame design, and modular architecture needed to support two collaborative robotic arms on an autonomous mobile base. It also outlines planned and ongoing functional tests to validate the platform’s ability to perform complex manipulation tasks in realistic operating conditions.
Abstract: This paper proposes a reinforcement learning approach for solving the inverse kinematic model of collaborative industrial manipulators. It uses a Deep Deterministic Policy Gradient agent with replay memory, Hindsight Experience Replay, and input normalization in a simulated environment based on the ABB GOFA 5_95 collaborative robot. The reward function combines position and angular error penalties, and the agent is trained across 27 workspace subregions. Evaluation on spherical and spheroidal surfaces shows fast inference times, high success rates, and moderate positioning errors, suggesting a real-time alternative to iterative Jacobian-based methods where sub-centimetre accuracy is not required.
Abstract: This paper proposes a framework for generating human-like bimanual robot-to-human handovers, especially for large, deformable, or delicate objects. The method uses a Hidden Semi-Markov Model to reactively generate robot response trajectories based on the observed motion of the human partner. The generated trajectories are adapted with task-space constraints to ensure accurate handover execution. A pilot study indicates that the resulting motions are perceived as more human-like than a baseline inverse-kinematics approach, supporting smoother and more natural human–robot object transfer.
Abstract: This paper presents an initial intelligent methodology for optimizing industrial manipulator trajectories using Particle Swarm Optimization. The evaluation is carried out in a MATLAB simulation environment with a dynamic robot model that includes mass, volume, inertia, obstacles, and collision detection. Three cost functions are compared: total trajectory time, joint effort, and a weighted combination of both. The optimization process first removes colliding trajectories to ensure geometric feasibility, then refines physically meaningful metrics, producing more efficient and realistic trajectories for demanding industrial environments.
Abstract: This paper extends trajectory optimization for industrial manipulators to a multi-objective formulation based on Multi-Objective Particle Swarm Optimization. The method is implemented in MATLAB using a dynamic model of the ABB CRB15000-95 collaborative robot and accounts for link mass, volume, inertia, environmental obstacles, and self-collisions. The strategy optimizes execution time and joint effort simultaneously through staged cost functions that first ensure collision-free feasibility and then improve physical performance. The resulting Pareto front captures efficient trade-offs between time and effort, enabling the selection of smooth and feasible trajectories adapted to industrial constraints.
Abstract: This paper addresses the segmentation of human–robot interaction demonstrations for learning from human and robot motion data. It builds on Hidden Markov Models with Gaussian Mixture structure, which are commonly used for modelling and predicting interactive tasks. The proposed method learns additional transition-state clusters at the boundaries between interaction phases, reducing misclassification in ambiguous states. Experiments on tasks such as handshaking and fist bumps show that transition-aware segmentation improves the predictive performance of Gaussian-mixture-based interaction learning.
Abstract: This paper introduces ZS6D, a zero-shot method for estimating the 6D pose of novel objects without object-specific training. The approach uses visual descriptors from pre-trained Vision Transformers to match rendered templates to query images and establish local correspondences. These correspondences are converted into geometric matches and used with RANSAC-based perspective-n-point estimation to recover object pose. Experiments on LMO, YCBV, and TLESS show improvements over existing novel-object pose estimation methods, highlighting the value of general-purpose Vision Transformer features for robotic perception.
Magazines, Press, & Media Coverage
Fraport Greece: New increase in passenger traffic in 2024 – Robotic services to be launched
Fraport Greece: The innovations entering the airports and changing the passenger experience