The human hand’s hybrid structure combines soft and rigid anatomy to provide strength and compliance for versatile object grasping. Tactile sensing by skin mechanoreceptors enables precise and dynamic manipulation. Attempts to replicate the human hand have fallen short of a true biomimetic hybrid robotic hand with tactile sensing. We introduce a natural prosthetic hand composed of soft robotic joints and a rigid endoskeleton with three independent neuromorphic tactile sensing layers inspired by human physiology. Our innovative design capitalizes on the strengths of both soft and rigid robots, enabling the hybrid robotic hand to compliantly grasp numerous everyday objects of varying surface textures, weight, and compliance while differentiating them with 99.69% average classification accuracy. The hybrid robotic hand with multilayered tactile sensing achieved 98.38% average classification accuracy in a texture discrimination task, surpassing soft robotic and rigid prosthetic fingers. Controlled via electromyography, our transformative prosthetic hand allows individuals with upper-limb loss to grasp compliant objects with precise surface texture detection.
As higher spatiotemporal resolution tactile sensing systems are being developed for prosthetics, wearables, and other biomedical applications, they demand faster sampling rates and generate larger data streams. Sparsifying transformations can alleviate these requirements by enabling compressive sampling and efficient data storage through compression. However, research on the best sparsifying transforms for tactile interactions is lagging. In this work we construct a library of orthogonal and biorthogonal wavelet transforms as sparsifying transforms for tactile interactions and compare their tradeoffs in compression and sparsity. We tested the sparsifying transforms on a publicly available high-density tactile object grasping dataset (548 sensor tactile glove, grasping 26 objects). In addition, we investigated which dimension wavelet transform—1D, 2D, or 3D—would best compress these tactile interactions. Our results show that wavelet transforms are highly efficient at compressing tactile data and can lead to very sparse and compact tactile representations. Additionally, our results show that 1D transforms achieve the sparsest representations, followed by 3D, and lastly 2D. Overall, the best wavelet for coarse approximation is Symlets 4 evaluated temporally which can sparsify to 0.5% sparsity and compress 10-bit tactile data to an average of 0.04 bits per pixel. Future studies can leverage the results of this paper to assist in the compressive sampling of large tactile arrays and free up computational resources for real-time processing on computationally constrained mobile platforms like neuroprosthetics.
Cardiac arrest (CA) remains the leading cause of coma, and early arousal recovery indicators are needed to allocate critical care resources properly. High-frequency oscillations (HFOs) of somatosensory evoked potentials (SSEPs) have been shown to indicate responsive wakefulness days following CA. Nonetheless, their potential in the acute recovery phase, where the injury is reversible, has not been tested. We hypothesize that time-frequency (TF) analysis of HFOs can determine arousal recovery in the acute recovery phase. To test our hypothesis, eleven adult male Wistar rats were subjected to asphyxial CA (five with 3-min mild and six with 7-min moderate to severe CA) and SSEPs were recorded for 60 min post-resuscitation. Arousal level was quantified by the neurological deficit scale (NDS) at 4 h. Our results demonstrated that continuous wavelet transform (CWT) of SSEPs localizes HFOs in the TF domain under baseline conditions. The energy dispersed immediately after injury and gradually recovered. We proposed a novel TF-domain measure of HFO: the total power in the normal time-frequency space (NTFS) of HFO. We found that the NTFS power significantly separated the favorable and unfavorable outcome groups. We conclude that the NTFS power of HFOs provides earlier and objective determination of arousal recovery after CA.
Humans possess an exquisite sense of touch, which robotic and prosthetic systems aim to replicate. Algorithms are developed to create neuron‐like (neuromorphic) spiking representations of texture that are invariant to the scanning speed and contact force applied in the sensing process. These spiking representations mimic the activity of mechanoreceptors in human skin and subsequent processing up to the brain. The algorithms are tested on a tactile texture dataset collected under 15 speed–force conditions. An offline texture classification system based on the invariant representations demonstrates higher classification accuracy, improved computational efficiency, and enhanced capability to identify textures explored in novel speed–force conditions. The speed‐invariance algorithm is adapted for a real‐time human‐operated texture classification system. In this system, invariant representations again improve classification accuracy, computational efficiency, and the ability to identify textures encountered under novel conditions. The invariant representation is particularly critical in this context, as human imprecision can be perceived as a novel condition by the classification system. These results show that invariant neuromorphic representations enable superior performance in neurorobotic sensing systems. Additionally, because the neuromorphic representations are grounded in biological processing, this work can serve as the basis for naturalistic sensory feedback for upper limb amputees.
Whole-body, large-area tactile skin is well-cited as a requirement for emerging humanoids, prosthetics, and other robotic devices operating in unstructured environments. However, developing effective large-area tactile sensing skin has been elusive due to many complexities, particularly challenged by wiring and big data management, and remains a persisting issue. This review describes the challenges of large-area tactile skin and summarizes the recent progress in addressing the wiring and data management problems. This review summarizes tactile wiring solutions, such as time, frequency, spike/code multiplexing, tomographic approaches, wireless communication, and tactile data management solutions, such as send-on-delta, hardware compression, and compressed sensing. Several notable scalable architectures are also highlighted, and possible future directions are discussed. Unlike previous reviews focused on power management, manufacturing techniques, and neural encoding, this review concentrates specifically on the scalable communication methods developed for large tactile skins, which have not been previously reviewed. Future research can use this review as a guide to build on the current state of the art to develop advanced large-area tactile skin technologies that support the next generation of environment-aware robotic devices.