Explore the vast range of titles available.

This paper presents novel Neural Nanowire Field Effect Transistors (υ-NWFETs) based hardware-implementable neural network (HNN) approach for tactile data processing in electronic skin (e-skin). The viability of Si nanowires (NWs) as the active material for υ-NWFETs in HNN is explored through modeling and demonstrated by fabricating the first device. Using υ-NWFETs to realize HNNs is an interesting approach as by printing NWs on large area flexible substrates it will be possible to develop a bendable tactile skin with distributed neural elements (for local data processing, as in biological skin) in the backplane. The modeling and simulation of υ-NWFET based devices show that the overlapping areas between individual gates and the floating gate determines the initial synaptic weights of the neural network - thus validating the working of υ-NWFETs as the building block for HNN. The simulation has been further extended to υ-NWFET based circuits and neuronal computation system and this has been validated by interfacing it with a transparent tactile skin prototype (comprising of 6 × 6 ITO based capacitive tactile sensors array) integrated on the palm of a 3D printed robotic hand. In this regard, a tactile data coding system is presented to detect touch gesture and the direction of touch. Following these simulation studies, a four-gated υ-NWFET is fabricated with Pt/Ti metal stack for gates, source and drain, Ni floating gate, and Al O high-k dielectric layer. The current-voltage characteristics of fabricated υ-NWFET devices confirm the dependence of turn-off voltages on the (synaptic) weight of each gate. The presented υ-NWFET approach is promising for a neuro-robotic tactile sensory system with distributed computing as well as numerous futuristic applications such as prosthetics, and electroceuticals.

Understanding the world in terms of objects and the possible interactions with them is an important cognitive ability. However, current world models adopted in reinforcement learning typically lack this structure and represent the world state in a global latent vector. To address this, we propose FOCUS, a model-based agent that learns an object-centric world model. This novel representation also enables the design of an object-centric exploration mechanism, which encourages the agent to interact with objects and discover useful interactions. We benchmark FOCUS in several robotic manipulation settings, where we found that our method can be used to improve manipulation skills. The object-centric world model leads to more accurate predictions of the objects in the scene and it enables more efficient learning. The object-centric exploration strategy fosters interactions with the objects in the environment, such as reaching, moving, and rotating them, and it allows fast adaptation of the agent to sparse reward reinforcement learning tasks. Using a Franka Emika robot arm, we also showcase how FOCUS proves useful in real-world applications. Website: focus-manipulation.github.io .

State-of-the-art controllers for active back exosuits rely on body kinematics and state machines. These controllers do not continuously target the lumbosacral compression forces or adapt to unknown external loads. The use of additional contact or load detection could make such controllers more adaptive; however, it can be impractical for daily use. Here, we developed a novel neuro-mechanical model-based controller (NMBC) that uses a personalized electromyography (EMG)-driven musculoskeletal (MSK) model to estimate lumbosacral joint loading. NMBC provided adaptive, subject- and load-specific assistive forces proportional to estimates of the active part of biological joint moments through a soft back support exosuit. Without a priori information, the maximum assistive forces of the cable were modulated across weights. Simultaneously, we applied a non-adaptive, kinematic-dependent, trunk inclination-based controller (TIBC). Both NMBC and TIBC reduced the mean and peak biomechanical metrics, although not all reductions were significant. TIBC did not modulate assistance across weights. NMBC showed larger reductions of mean than peak values, significant reductions during the erect stance and the cumulative compressive loads by 21% over multiple cycles in a cohort of 10 participants. Overall, NMBC targeted mean lumbosacral compressive forces during lifting without a priori information of the load being carried. This may facilitate the adoption of non-hindering wearable robotics in real-life scenarios. As NMBC is informed by an EMG-driven MSK model, it is possible to tune the timing of NMBC-generated torque commands to the exosuit (delaying or anticipating commands with respect to biological torques) to target further reduction of peak or mean compressive forces and muscle fatigue.

Generalization ability in tactile sensing for robotic manipulation is a prerequisite to effectively perform tasks in ever-changing environments. In particular, performing dynamic tactile perception is currently beyond the ability of robotic devices. A biomimetic approach to achieve this dexterity is to develop machines combining compliant robotic manipulators with neuroinspired architectures displaying computational adaptation. Here we demonstrate the feasibility of this approach for dynamic touch tasks experimented by integrating our sensing apparatus in a 6 degrees of freedom robotic arm via a soft wrist. We embodied in the system a model of spike-based neuromorphic encoding of tactile stimuli, emulating the discrimination properties of cuneate nucleus neurons based on pathways with differential delay lines. These strategies allowed the system to correctly perform a dynamic touch protocol of edge orientation recognition (ridges from 0 to 40°, with a step of 5°). Crucially, the task was robust to contact noise and was performed with high performance irrespectively of sensing conditions (sensing forces and velocities). These results are a step forward toward the development of robotic arms able to physically interact in real-world environments with tactile sensing.

Numerical cognition is a fundamental component of human intelligence that has not been fully understood yet. Indeed, it is a subject of research in many disciplines, e.g., neuroscience, education, cognitive and developmental psychology, philosophy of mathematics, linguistics. In Artificial Intelligence, aspects of numerical cognition have been modelled through neural networks to replicate and analytically study children behaviours. However, artificial models need to incorporate realistic sensory-motor information from the body to fully mimic the children's learning behaviours, e.g., the use of fingers to learn and manipulate numbers. To this end, this article presents a database of images, focused on number representation with fingers using both human and robot hands, which can constitute the base for building new realistic models of numerical cognition in humanoid robots, enabling a grounded learning approach in developmental autonomous agents. The article provides a benchmark analysis of the datasets in the database that are used to train, validate, and test five state-of-the art deep neural networks, which are compared for classification accuracy together with an analysis of the computational requirements of each network. The discussion highlights the trade-off between speed and precision in the detection, which is required for realistic applications in robotics.

The possible control of axonal outgrowth during neural regeneration could be very useful not only from a neurobiological point of view, but also in the field of neural interfaces. In this manuscript, simulations are presented which investigate the possibility of guiding axons by using a hybrid approach based on the combined used of a chemical model and of a genetic algorithm. Microspheres embedding chemical cues on the basis of information provided by a genetic algorithm are placed to impose a desired trajectory on the axons. Two kinds of simulations were carried out: (i) tracking of linear trajectories; (ii) tracking of trajectories, which were reconstructed from real axonal extension. The results achieved during the simulations seem to confirm the possible use of this approach to guide axonal outgrowth, being the obtained trajectories congruent to possible actual situations. Moreover, the model can be easily extended to a three-dimensional environment.

Remarkable biological examples of molecular robots are the proteins kinesin-1 and dynein, which move and transport cargo down microtubule \"highways,\" e.g., of the axon, to final nerve nodes or along dendrites. They convert the energy of ATP hydrolysis into mechanical forces and can thereby push them forwards or backwards step by step. Such mechano-chemical cycles that generate conformal changes are essential for transport on all different types of substrate lanes. The step length of an individual molecular robot is a matter of nanometers but the dynamics of each individual step cannot be predicted with certainty (as it is a random process). Hence, our proposal is to involve the methods of quantum field theory (QFT) to describe an overall reliable, multi-robot system that is composed of a huge set of unreliable, local elements. The methods of QFT deliver techniques that are also computationally demanding to synchronize the motion of these molecular robots on one substrate lane as well as across lanes. Three different challenging types of solutions are elaborated. The impact solution reflects the particle point of view; the two remaining solutions are wave based. The second solution outlines coherent robot motions on different lanes. The third solution describes running waves. Experimental investigations are needed to clarify under which biological conditions such different solutions occur. Moreover, such a nano-chemical system can be stimulated by external signals, and this opens a new, hybrid approach to analyze and control the combined system of robots and microtubules externally. Such a method offers the chance to detect mal-functions of the biological system.

Neuro-robots are a class of autonomous machines that, in their architecture, mimic aspects of the human brain and cognition. As such, they represent unique artifacts created by humans based on human understanding of healthy human brains. European Union’s Convention on Roboethics 2025 states that the design of all robots (including neuro-robots) must include provisions for the complete traceability of the robots’ actions, analogous to an aircraft’s flight data recorder. At the same time, one can anticipate rising instances of neuro-robotic failure, as they operate on imperfect data in real environments, and the underlying AI behind such neuro-robots has yet to achieve explainability. This paper reviews the trajectory of the technology used in neuro-robots and accompanying failures. The failures demand an explanation. While drawing on existing explainable AI research, we argue explainability in AI limits the same in neuro-robots. In order to make robots more explainable, we suggest potential pathways for future research.

During surgery for foci-related epilepsy, neurosurgeons face significant difficulties in identifying and resecting MRI-negative or deep-seated epileptic foci. Here, we present a neuro-robotic navigation system that is specifically designed for resection of MRI negative epileptic foci. We recruited 52 epileptic patients, and randomly assigned them to treatment group with either neuro-robotic navigation or conventional neuronavigation system. For each patient, in the neuro-robotic navigation group, we integrated multimodality imaging including MRI and PET-CT into the robotic workstation and marked the boundary of foci from the fused image. During surgery, this boundary was delineated by the robotic laser device with high accuracy, guiding resection for the surgeon. For deeply seated foci, we exploited the neuro-robotic navigation system to localize the deepest point with biopsy needle insertion and methylene dye application to locate the boundary of the foci. Our results show that, compared with the conventional neuronavigation, the neuro-robotic navigation system performs equally well in MRI positive epilepsy patients (ENGEL I ratio: 71.4% vs 100%, p  = 0.255) systems and show better performance in patients with MRI-negative focal cortical dysplasia (ENGEL I ratio: 88.2% vs 50%, p  = 0.0439). At present, there are no documented neurosurgery robots with similar function and application in the field of epilepsy. Our research highlights the added value of using neuro-robotic navigation systems in resection surgery for epilepsy, particularly in cases that involve MRI-negative or deep-seated epileptic foci.