In recent years, flexible sensors for data gloves have been developed that aim to achieve excellent wearability, but they are associated with difficulties due to the complicated manufacturing and embedding into the glove. This study proposes a knitted glove integrated with strain sensors for pattern recognition of hand postures. The proposed sensing glove is fabricated at all once by a knitting technique without sewing and bonding, which is composed of strain sensors knitted with conductive yarn and a glove body with non-conductive yarn. To verify the performance of the developed glove, electrical resistance variations were measured according to the flexed angle and speed. These data showed different values depending on the speed or angle of movements. We carried out experiments on hand postures pattern recognition for the practicability verification of the knitted sensing glove. For this purpose, 10 able-bodied subjects participated in the recognition experiments on 10 target hand postures. The average classification accuracy of 10 subjects reached 94.17% when their own data were used. The accuracy of up to 97.1% was achieved in the case of grasp posture among 10 target postures. When all mixed data from 10 subjects were utilized for pattern recognition, the average classification expressed by the confusion matrix arrived at 89.5%. Therefore, the comprehensive experimental results demonstrated the effectiveness of the knitted sensing gloves. In addition, it is expected to reduce the cost through a simple manufacturing process of the knitted sensing glove.

This paper proposes a kinematic modeling method for robotic manipulators by extracting the modified Denavit–Hartenberg (MDH) parameters using line geometry. For single-branched manipulators, various joint axes can be represented as lines using Plücker coordinates. The forward kinematics is derived by performing the product of matrices which are the exponential maps lifted from two kinds of exponential coordinates using the MDH parameters. For extracting MDH parameters, line geometry systematically analyzes the following: (1) the closest point between a point and line, (2) the closest distance and twist angle between two lines, (3) the common perpendicular line and its intersection points, and (4) classifies line relationships into collinear, distant parallel, intersected, and skewed cases. For each case, five parameters including twist angle, closest distance, common perpendicular direction vector, and both feet on a common perpendicular line are sequentially computed as results of the line geometry block. Finally, the aforementioned line geometry blocks are utilized to extract the four MDH parameters according to their definitions. The effectiveness of the proposed algorithm is verified by four examples including a typical Selective Compliance Assembly Robot Arm (SCARA) robot and three different commercial manipulators.

The degrees-of-freedom of robotic prosthetic hands have recently increased, but high-level amputees such as those with shoulder disarticulation and trans-humeral amputation do not have enough muscular areas on their upper limbs upon which to base surface electromyogram (sEMG) signals. In this paper, a wearable fabric sensor is proposed to measure the sEMG on the lower limb and to classify the foot postures by using the proposed convolutional neural network (CNN), ultimately, for the application to high-level upper limb amputees. First, we determined that sEMG signals of the lower limb can be classified into levels in a manner similar to those of the upper limb for eight postures. Second, a multilayer perceptron (MLP) and the proposed CNN was used to compare the pattern recognition accuracy for classifying eight postures. Finally, the wearable fabric sensor and the proposed CNN network were demonstrated by the trans-radial amputees. These results showed that the wearable fabric sensor verified different eight patterns based on similar motions of both limbs ( p < 0.001). In addition, the classification accuracy (91.3%) of the proposed CNN was much higher than that (79%) of MLP ( p < 0.05). The wearable fabric sensor allowed the measurement location to change from the upper limb to the lower limb and allowed the number of the classifiable patterns to increase thanks to the sixteen-channel sEMG signals acquired from 32 fabric electrodes. The high classification accuracy of the proposed CNN will be useful for various users who have to wear myoelectric prosthesis every day.