Learning motor control for muscle-driven musculoskeletal models is hindered by the computational cost of biomechanically accurate simulation and the scarcity of validated, open full-body models. Here we present MuscleMimic, an open-source framework for scalable motion imitation learning with physiologically realistic, muscle-actuated humanoids. MuscleMimic provides two validated musculoskeletal embodiments - a fixed-root upper-body model (126 muscles) for bimanual manipulation and a full-body model (416 muscles) for locomotion - together with a retargeting pipeline that maps SMPL-format motion capture data onto musculoskeletal structures while preserving kinematic and dynamic consistency. Leveraging massively parallel GPU simulation, the framework achieves order-of-magnitude training speedups over prior CPU-based approaches while maintaining comprehensive collision handling, enabling a single generalist policy to be trained on hundreds of diverse motions within days. The resulting policy faithfully reproduces a broad repertoire of human movements under full muscular control and can be fine-tuned to novel motions within hours. Biomechanical validation against experimental walking and running data demonstrates strong agreement in joint kinematics (mean correlation r = 0.90), while muscle activation analysis reveals both the promise and fundamental challenges of achieving physiological fidelity through kinematic imitation alone. By lowering the computational and data barriers to musculoskeletal simulation, MuscleMimic enables systematic model validation across diverse dynamic movements and broader participation in neuromuscular control research. Code, models, checkpoints, and retargeted datasets are available at: https://github.com/amathislab/musclemimic
Xingxin He, Zachary E Stewart, Nikitha Crasta
et al.
Abstract Background Large language models (LLMs) have shown promising abilities in text-based clinical tasks but they do not inherently interpret medical images such as knee radiographs. Purpose To develop a human-artificial intelligence interactive diagnostic approach, named radiology generative pretrained transformer (RadGPT), aimed at assisting and synergizing with human users for the interpretation of knee radiological images. Materials and Methods A total of 22 512 knee roentgen ray images and reports were retrieved from Massachusetts General Hospital; 80% of these were used for model training and 10% were used for model testing and validation, respectively. Fifteen diagnostic imaging features (eg, osteoarthritis, effusion, joint space narrowing, osteophyte) were selected to label images based on their high frequency and clinical relevance in the retrieved official reports. Area under the curve scores were calculated for each feature to assess the diagnostic performance. To evaluate the quality of the generated medical text, historical clinical reports were used as the reference text. Several metrics for text generation tasks are applied, including BiLingual Evaluation Understudy, Recall-Oriented Understudy for Gisting Evaluation, Metric for Evaluation of Translation with Explicit Ordering, and Semantic Propositional Image Caption Evaluation. Results RadGPT, in collaboration with human users, achieved area under the curve scores ranging from 0.76 for osteonecrosis to 0.91 for arthroplasty across 15 diagnostic categories for knee conditions. Compared with the baseline LLM method, RadGPT achieved higher scores, specifically 0.18 in BiLingual Evaluation Understudy score, 0.30 in Recall-Oriented Understudy for Gisting Evaluation-L, 0.10 in Metric for Evaluation of Translation with Explicit Ordering, and 0.15 in Semantic Propositional Image Caption Evaluation, which is significantly higher than the baseline LLM method, demonstrating good linguistic overlap and clinical consistency with the reference reports. Conclusion RadGPT has achieved advanced results in knee roentgen ray image feature recognition, illustrating the potential of LLMs in medical image interpretation. The study establishes a training protocol for developing artificial intelligence-assisted tools specifically focusing on the diagnosis and interpretation of knee radiological images.
Abstract Osteosarcoma (OS) is the most common primary bone malignancy because of its extra high tendency of metastasis. In-depth research is needed to uncover the pathogenesis of patients with OS cells. We collected 74 tissue samples from patients with OS cells and measured the expression levels of ghrelin by immunohistochemistry. Ghrelin was added into OS cell lines in CCK8 assays, JC-1 staining and Western blot analysis were performed to explore its effect on the aggressiveness of OS cells and drug resistance. To determine its function, ghrelin was overexpressed or knocked down in OS cells and then detect cell proliferation in the xenograft mouse model and orthotopic model. Western blot analysis was performed to explore ghrelin-regulated signal pathways. In this work, we identified the relation between the level of ghrelin expression and poor prognosis of OS patients. As well as promoting proliferation, migration, and invation, ghrelin promotes the survival of OS in vitro as well as in vivo, and reduces the apoptosis of OS cells. What’s more, ghrelin increases the resistance of cis-platinum by changing mitochondrial function and decreases the expression of MDR-1. Above all, these results demonstrated ghrelin exerts tumorigenic and metastatic effects and may be a potential therapeutic target.
Orthopedic surgery, Diseases of the musculoskeletal system
Eduardo Borges Ferreira Jr, Daniel Yiteh Lin, Maria Mascarenhas
et al.
Introduction:
Injuries to the brachial triceps (BT) tendon are rare, accounting for about 1% of all tendon injuries. The complexity of these injuries, especially in their proximal portion, and the scarcity of data in the literature make each case essential to expand knowledge. This report demonstrates an effective treatment method for extensive muscle loss around the shoulder using a homologous graft, offering new perspectives for similar cases.
Case Report:
A 65-year-old female patient presented with a history of desmoid tumor treated surgically through an extensive surgical resection of the proximal portion of the BT muscle, followed by a third-degree burn in the left scapular region. The patient exhibited sequelae, manifesting complaints related to esthetics and reduced active extension of the elbow. Hypotrophy of the deltoid muscle was also observed, associated with paresthesia in the corresponding C5 dermatome. The patient underwent surgical treatment for reconstruction of the proximal portion of the BT muscle using a homologous graft from the tensor fascia lata muscle, which resulted in esthetic and functional improvements without additional neurological deficits.
Conclusion:
The case report demonstrates that reconstruction of the proximal portion of the BT muscle with a homologous graft is a promising approach for treating this type of injury, bringing significant improvements in muscle function and esthetics. Furthermore, a multidisciplinary approach, postoperative surveillance, and the continuous pursuit of technical advancements are essential to optimize results and minimize complications, broadening therapeutic options in complex cases.
Orthopedic surgery, Diseases of the musculoskeletal system
Panos Christodoulou, Donald Osarumwense, Antonios Papadopoulos
Introduction:
Acute carpal tunnel syndrome (CTS) is a rare but serious condition requiring prompt diagnosis and intervention to prevent permanent neurological deficits. This case series presents our department’s experience in managing this condition, highlighting diagnostic challenges, treatment strategies, and patient outcomes.
Materials & Methods:
This retrospective study analyses medical records of patients diagnosed with acute carpal tunnel syndrome from 2021 to 2023 within our coverage area. Information was extracted from medical records, including diagnostic codes and clinical assessments for all cases of acute carpal tunnel syndrome, diagnosed during the period from 2021 to 2023. Data analysed, include age, sex, medical history, symptoms, mechanism of injury, clinical features, imaging, management, and follow-up. Six native Scottish patients, aged 43 to 65, developed acute carpal tunnel syndrome due to various causes, including flexor sheath infection, wrist soft tissue injury, distal radius fractures (with and without plate fixation), and scaphoid fracture.
Discussion and Results:
Our case series highlights the heterogeneous presentation and aetiology of acute carpal tunnel syndrome, with causes ranging from trauma to infection. Early recognition remained challenging due to overlapping symptoms with other wrist pathologies. Surgical decompression was the mainstay of treatment, performed with variable urgency depending on clinical progression. Despite intervention outcomes were mixed. Some patients experienced full recovery while others had lingering symptoms.
Conclusion:
These findings underscore the importance of high clinical suspicion and prompt surgical intervention to optimise outcomes.
Orthopedic surgery, Diseases of the musculoskeletal system
Abstract Background Metacarpal neck fractures are common and there are numerous surgical methods available, but each has certain disadvantages and limitations. We modified the conventional Ilizarov external mini-fixator and this study is designed to compare the biomechanical stability of a modified Ilizarov external mini-fixator with conventional fixation methods for metacarpal neck fractures and to provide a basis for its clinical application. Methods Forty fresh porcine metacarpal specimens were used to create metacarpal neck fracture models. The specimens were randomly assigned to four fixation groups (n = 10) as follows: (1) modified Ilizarov external mini-fixator (IEF), (2) retrograde crossed Kirschner wires (KW), (3) antegrade intramedullary Kirschner wires (IK), and (4) locking plate fixation (LP). In the IEF group, the modified design involved crossing two Kirschner wires (K-wires) through the fracture line, with their tails bent twice and connected to the external fixator frame. Biomechanical testing was performed using a modified three-point bending test. Maximum fracture force and bending stiffness were calculated from the force-displacement curves. Kruskal–Wallis test was used to compare statistical differences in maximum fracture force and stiffness among the groups, followed by post hoc pairwise comparisons adjusted with Bonferroni corrections. Results The median maximum fracture force values (± interquartile range, IQR) for each group were as follows: IEF 160.3 ± 55.6 N, LP 173.5 ± 42.6 N, KW 91.1 ± 23.1 N, and IK 79.8 ± 37.8 N. The corresponding stiffness values were as follows: IEF 29.5 ± 10.4 N/mm, LP 32.9 ± 10.4 N/mm, KW 17.2 ± 11.3 N/mm, and IK 18.2 ± 13.7 N/mm. The IEF group demonstrated significantly higher maximum fracture force and stiffness than the KW and IK groups; however, no statistically significant differences were observed in the IEF group compared with the LP group. Conclusion The modified Ilizarov external mini-fixator provided significantly greater biomechanical stability for metacarpal neck fractures than retrograde crossed K-wires and antegrade intramedullary K-wires, achieving comparable performance to the locking plate system. This modified design combines the simplicity and minimally invasive advantages of K-wire fixation with enhanced stability, potentially facilitating early joint mobilization and minimizing the risk of complication.
Orthopedic surgery, Diseases of the musculoskeletal system
Introduction:
Metallosis is a pathological condition associated with the release of metal debris from joint implants, particularly metal-on-metal (MoM) articulations in total hip arthroplasty (THA). The metal debris triggers adverse local tissue reactions, including aseptic lymphocyte-vasculitis-associated lesions (ALVALs), pseudotumor formation, and progressive implant loosening.
Case Report:
Case 1 - A 64-year-old male presented with hip pain and limping 15 years after an uncemented MoM THA. Investigations revealed femoral stem loosening. Intraoperatively, dark synovial fluid, necrotic tissue, elevated cobalt, and chromium levels were consistent with metallosis were identified. Revision surgery replaced the components with alternative bearings, and the patient returned to normal activities within a year. Case 2 - A 63-year-old female with bilateral MoM THAs reported similar symptoms 15-year post-surgery. Imaging identified a pseudotumor and femoral loosening. Elevated serum metal ion levels confirmed the diagnosis. Revision surgery revealed ALVAL, and components were replaced with ceramic-on-polyethylene implants, resolving symptoms.
Conclusion:
Metallosis poses a significant risk in MoM THA, caused by wear and corrosion, leading to systemic and local tissue damage. Timely diagnosis through clinical evaluation, imaging, and serum metal ion levels is critical for intervention. Revision surgeries are effective in managing metallosis and restoring function. These cases highlight the importance of long-term monitoring and the consideration of alternative bearing surfaces in hip arthroplasty.
Orthopedic surgery, Diseases of the musculoskeletal system
Saraswati Soedarmadji, Yunyue Wei, Chen Zhang
et al.
Discovering effective reward functions remains a fundamental challenge in motor control of high-dimensional musculoskeletal systems. While humans can describe movement goals explicitly such as "walking forward with an upright posture," the underlying control strategies that realize these goals are largely implicit, making it difficult to directly design rewards from high-level goals and natural language descriptions. We introduce Motion from Vision-Language Representation (MoVLR), a framework that leverages vision-language models (VLMs) to bridge the gap between goal specification and movement control. Rather than relying on handcrafted rewards, MoVLR iteratively explores the reward space through iterative interaction between control optimization and VLM feedback, aligning control policies with physically coordinated behaviors. Our approach transforms language and vision-based assessments into structured guidance for embodied learning, enabling the discovery and refinement of reward functions for high-dimensional musculoskeletal locomotion and manipulation. These results suggest that VLMs can effectively ground abstract motion descriptions in the implicit principles governing physiological motor control.
Shaheer Ahmad Khan, Muhammad Usamah Shahid, Ahmad Abdullah
et al.
This study addresses a critical gap in the healthcare system by developing a clinically meaningful, practical, and explainable disease surveillance system for multiple chronic diseases, utilizing routine EHR data from multiple U.S. practices integrated with CureMD's EMR/EHR system. Unlike traditional systems--using AI models that rely on features from patients' labs--our approach focuses on routinely available data, such as medical history, vitals, diagnoses, and medications, to preemptively assess the risks of chronic diseases in the next year. We trained three distinct models for each chronic disease: prediction models that forecast the risk of a disease 3, 6, and 12 months before a potential diagnosis. We developed Random Forest models, which were internally validated using F1 scores and AUROC as performance metrics and further evaluated by a panel of expert physicians for clinical relevance based on inferences grounded in medical knowledge. Additionally, we discuss our implementation of integrating these models into a practical EMR system. Beyond using Shapley attributes and surrogate models for explainability, we also introduce a new rule-engineering framework to enhance the intrinsic explainability of Random Forests.
Alberto Pliego Marugán, Jesús M. Pinar-Pérez, Fausto Pedro García Márquez
Efficient maintenance has always been essential for the successful application of engineering systems. However, the challenges to be overcome in the implementation of Industry 4.0 necessitate new paradigms of maintenance optimization. Machine learning techniques are becoming increasingly used in engineering and maintenance, with reinforcement learning being one of the most promising. In this paper, we propose a gamma degradation process together with a novel maintenance model in which repairs are increasingly imperfect, i.e., the beneficial effect of system repairs decreases as more repairs are performed, reflecting the degradational behavior of real-world systems. To generate maintenance policies for this system, we developed a reinforcement-learning-based agent using a Double Deep Q-Network architecture. This agent presents two important advantages: it works without a predefined preventive threshold, and it can operate in a continuous degradation state space. Our agent learns to behave in different scenarios, showing great flexibility. In addition, we performed an analysis of how changes in the main parameters of the environment affect the maintenance policy proposed by the agent. The proposed approach is demonstrated to be appropriate and to significatively improve long-run cost as compared with other common maintenance strategies.
Franco Oberti, Stefano Di Carlo, Alessandro Savino
The Controller Area Network (CAN) protocol, essential for automotive embedded systems, lacks inherent security features, making it vulnerable to cyber threats, especially with the rise of autonomous vehicles. Traditional security measures offer limited protection, such as payload encryption and message authentication. This paper presents a novel Intrusion Detection System (IDS) designed for the CAN environment, utilizing Hardware Performance Counters (HPCs) to detect anomalies indicative of cyber attacks. A RISC-V-based CAN receiver is simulated using the gem5 simulator, processing CAN frame payloads with AES-128 encryption as FreeRTOS tasks, which trigger distinct HPC responses. Key HPC features are optimized through data extraction and correlation analysis to enhance classification efficiency. Results indicate that this approach could significantly improve CAN security and address emerging challenges in automotive cybersecurity.
Kento Kawaharazuka, Manabu Nishiura, Shinsuke Nakashima
et al.
Although robots with flexible bodies are superior in terms of the contact and adaptability, it is difficult to control them precisely. On the other hand, human beings make use of the surrounding environments to stabilize their bodies and control their movements. In this study, we propose a method for the bracing motion and extension of the range of motion using the environment for the musculoskeletal humanoid. Here, it is necessary to recognize the stability of the body when contacting the environment, and we develop a method to measure it by using the change in sensor values of the body when actively vibrating a part of the body. Experiments are conducted using the musculoskeletal humanoid Musashi, and the effectiveness of this method is confirmed.
Andrea Fava, Catriona A. Wagner, Carla J. Guthridge
et al.
ObjectiveAutoantibodies are a hallmark of lupus nephritis (LN), but their association with LN classes and treatment response are not adequately known. In this study, we quantified circulating autoantibodies in the Accelerating Medicines Partnership LN longitudinal cohort to identify serological biomarkers of LN histologic classification and treatment response and how these biomarkers change over time based on treatment response.MethodsPeripheral blood samples were collected from 279 patients with systemic lupus erythematosus undergoing diagnostic kidney biopsy based on proteinuria. Of these, 268 were diagnosed with LN. Thirteen autoantibody specificities were measured by bead‐based assays (Bio‐Rad Bioplex 2200) and anti‐C1q by enzyme‐linked immunosorbent assay at the time of biopsy (baseline) and at 3, 6, and 12 months after biopsy. Clinical response was determined at 12 months.ResultsProliferative LN (International Society of Nephrology/Renal Pathology Society class III/IV±V, n = 160) was associated with higher concentrations of anti‐C1q, anti‐chromatin, anti–double‐stranded DNA (dsDNA), and anti–ribosomal P autoantibodies compared to nonproliferative LN (classes I/II/V/VI, n = 108). Anti‐C1q and‐dsDNA were independently associated with proliferative LN. In proliferative LN, higher baseline anti‐C1q levels predicted complete response (area under the curve [AUC] 0.72; P = 0.002) better than baseline proteinuria (AUC 0.59; P = 0.21). Furthermore, all autoantibody levels except for anti‐La/SSB decreased over 12 months in patients with proliferative, but not membranous, LN with a complete response.ConclusionBaseline levels of anti‐C1q and anti‐dsDNA may serve as noninvasive biomarkers of proliferative LN, and anti‐C1q may predict complete response at the time of kidney biopsy. In addition, tracking autoantibodies over time may provide further insights into treatment response and pathogenic mechanisms in patients with proliferative LN.image
Abstract Objectives This study aims to identify predictors of knee osteoarthritis (KOA) risk in middle-aged population, construct and validate a nomogram for KOA in this demographic. Methods From June to December 2020, we conducted a cross-sectional survey on 5,527 middle-aged individuals from Changsha and Zhangjiajie cities in Hunan Province, selected using a stratified multi-stage random sampling method. Data collection involved a structured questionnaire encompassing general demographic, physical condition, and lifestyle behaviors dimensions. The dataset was randomly split into a training set (n = 3868) and a validation set (n = 1659) at a 7:3 ratio via computerized randomization. We analyzed the prevalence of self-reported KOA and identified its influencing factors using logistic regression. A nomogram was constructed based on these "three-dimensional" factors. Subsequent validation was conducted, and the nomogram's performance was further evaluated through ROC curves, C-index, Hosmer–Lemeshow test, and calibration curves. Results The self-reported prevalence of KOA in the middle-aged population was 11.4% (632/5527). The risk factor with the greatest impact is: diagnosed with osteoporosis(95% CI 2.269–3.568, OR = 2.845), followed by age between 51 to 60 years (95% CI 2.176–3.151, OR = 2.619), diagnosed with hypertension(95% CI 1.633–2.499, OR = 2.02), diagnosed with diabetes (OR = 1.689), ethnic Han Chinese (OR = 1.673), exercise according to physical condition (OR = 1.643), pay attention to keeping the knee joint warm (OR = 1.535), eating habits are mainly light vegetables (OR = 1.374), male gender (OR = 1.343), drink occasionally in small amounts (OR = 1.286); a higher level of education (OR = 0.477) and frequently or always apply an external or plaster to relieve symptoms after knee discomfort (OR = 0.377; OR = 0.385) are protective factors. The C-index of the training set model was 0.8107 (95% CI: 0.8102–0.8111), with a statistically significant area under the ROC curve (AUC = 0.818), and the calibration curve showed a good fit. The C-index for the validation set was 0.8124 (95% CI: 0.8109–0.8140), with an AUC of 0.812. The Hosmer–Lemeshow test resulted in a P-value of 0.46 (P ≥ 0.05)indicating good calibration of the model. Conclusion The three dimensions nomogram generated in this study was a valid and easy-to-use tool for assessing the risk of KOA in middle-aged population, and helped healthcare professionals to screen the high-risk population.
Orthopedic surgery, Diseases of the musculoskeletal system
Abstract Background The present study evaluated whether the lack of histone deacetylase 4 (HDAC4) increases endoplasmic reticulum stress-induced chondrocyte apoptosis by releasing activating transcription factor 4 (ATF4) in human osteoarthritis (OA) cartilage degeneration. Methods Articular cartilage from the tibial plateau was obtained from patients with OA during total knee replacement. Cartilage extracted from severely damaged regions was classified as degraded cartilage, and cartilage extracted from a relatively smooth region was classified as preserved cartilage. Terminal deoxynucleotidyl transferase dUTP nick end labeling staining was used to detect chondrocyte apoptosis. HDAC4, ATF4, and C/EBP homologous protein (CHOP) expression levels were measured using immunohistochemistry staining and real-time quantitative PCR. Chondrocytes were transfected with HDAC4 or HDAC4 siRNA for 24 h and stimulated with 300 µM H2O2 for 12 h. The chondrocyte apoptosis was measured using flow cytometry. ATF4, CHOP, and caspase 12 expression levels were measured using real-time quantitative PCR and western blotting. Male Sprague-Dawley rats (n = 15) were randomly divided into three groups and transduced with different vectors: ACLT + Ad-GFP, ACLT + Ad-HDAC4-GFP, and sham + Ad-GFP. All rats received intra-articular injections 48 h after the operation and every three weeks thereafter. Cartilage damage was assessed using Safranin O staining and quantified using the Osteoarthritis Research Society International score. ATF4, CHOP, and collagen II expression were detected using immunohistochemistry, and chondrocyte apoptosis was detected using terminal deoxynucleotidyl transferase dUTP nick end labeling staining. Results The chondrocyte apoptosis was higher in degraded cartilage than in preserved cartilage. HDAC4 expression was lower in degraded cartilage than in preserved cartilage. ATF4 and CHOP expression was increased in degraded cartilage. Upregulation of HDAC4 in chondrocytes decreased the expression of ATF4, while the expression of ATF4 was increased after downregulation of HDAC4. Upregulation of HDAC4 decreased the chondrocyte apoptosis under endoplasmic reticulum stress, and chondrocyte apoptosis was increased after downregulation of HDAC4. In a rat anterior cruciate ligament transection OA model, adenovirus-mediated transduction of HDAC4 was administered by intra-articular injection. We detected a stronger Safranin O staining with lower Osteoarthritis Research Society International scores, lower ATF4 and CHOP production, stronger collagen II expression, and lower chondrocyte apoptosis in rats treated with Ad-HDAC4. Conclusion The lack of HDAC4 expression partially contributes to increased ATF4, CHOP, and endoplasmic reticulum stress-induced chondrocyte apoptosis in OA pathogenesis. HDAC4 attenuates cartilage damage by repressing ATF4-CHOP signaling-induced chondrocyte apoptosis in a rat model of OA.
Abstract Background The relationship between obesity and type 2 diabetes with bone health has always been a topic of debate. The weight-adjusted waist index has become a commonly used indicator for assessing central obesity, fat, and muscle mass. However, currently there is no research reporting the association between weight-adjusted waist index and risk of osteoporosis in populations of type 2 diabetes. Therefore, this study aims to provide new information on the association between weight-adjusted waist index and risk of osteoporosis in type 2 diabetes. Methods This cross-sectional study involved 963 patients with type 2 diabetes who were admitted to the Department of Endocrinology of Cangzhou Central Hospital. Multivariate logistic regression models were used to assess the association between weight-adjusted waist index and osteoporosis. The potential nonlinear association was evaluated. The effects of interaction between subgroups were assessed using the likelihood ratio test. Results Weight-adjusted waist index was positively associated with the risk of osteoporosis, regardless of traditional confounding factors. For each 1 unit increased in weight-adjusted waist index, the risk of osteoporosis increased by 67%. Furthermore, there was a nonlinear relationship between weight-adjusted waist index and osteoporosis. The subgroup analysis did not reveal any significant interactions. Conclusions Our study indicated a positive association between weight-adjusted waist index and the risk of osteoporosis in adult Chinese type 2 diabetes patients, and this relationship was nonlinear.
Orthopedic surgery, Diseases of the musculoskeletal system
Kento Kawaharazuka, Masaya Kawamura, Shogo Makino
et al.
The body structure of an anatomically correct tendon-driven musculoskeletal humanoid is complex, and the difference between its geometric model and the actual robot is very large because expressing the complex routes of tendon wires in a geometric model is very difficult. If we move a tendon-driven musculoskeletal humanoid by the tendon wire lengths of the geometric model, unintended muscle tension and slack will emerge. In some cases, this can lead to the wreckage of the actual robot. To solve this problem, we focused on reciprocal innervation in the human nervous system, and then implemented antagonist inhibition control (AIC) based on the reflex. This control makes it possible to avoid unnecessary internal muscle tension and slack of tendon wires caused by model error, and to perform wide range motion safely for a long time. To verify its effectiveness, we applied AIC to the upper limb of the tendon-driven musculoskeletal humanoid, Kengoro, and succeeded in dangling for 14 minutes and doing pull-ups.
Kento Kawaharazuka, Kei Tsuzuki, Moritaka Onitsuka
et al.
The flexible musculoskeletal hand is difficult to modelize, and its model can change constantly due to deterioration over time, irreproducibility of initialization, etc. Also, for object recognition, contact detection, and contact control using the hand, it is desirable not to use a neural network trained for each task, but to use only one integrated network. Therefore, we develop a method to acquire a sensor state equation of the musculoskeletal hand using a recurrent neural network with parametric bias. By using this network, the hand can realize recognition of the grasped object, contact simulation, detection, and control, and can cope with deterioration over time, irreproducibility of initialization, etc. by updating parametric bias. We apply this study to the hand of the musculoskeletal humanoid Musashi and show its effectiveness.
Kento Kawaharazuka, Tasuku Makabe, Shogo Makino
et al.
By the recent spread of machine learning in the robotics field, a humanoid that can act, perceive, and learn in the real world through contact with the environment needs to be developed. In this study, as one of the choices, we propose a novel humanoid TWIMP, which combines a human mimetic musculoskeletal upper limb with a two-wheel inverted pendulum. By combining the benefit of a musculoskeletal humanoid, which can achieve soft contact with the external environment, and the benefit of a two-wheel inverted pendulum with a small footprint and high mobility, we can easily investigate learning control systems in environments with contact and sudden impact. We reveal our whole concept and system details of TWIMP, and execute several preliminary experiments to show its potential ability.