This study presents the development of a hybrid active disturbance rejection control (H-ADRC) implementing a homogeneous extended state observer (ESO) to solve the trajectory tracking problem of an inchworm bioinspired flexible robotic device (IBFRD) with partially unknown dynamics. The proposed controller is implemented on a self-designed IBFRD instrumented with shape memory alloys (SMAs) operating as actuators in each joint of the robotic mobile device. The set of SMAs constitutes an agonist–antagonist mechanism together with regular springs that counteract the SMA displacement. The characteristic two-anchor crawling of an inchworm is modeled as a hybrid system and exerted by a bistable pneumatic circuit that switches suction cups at the terminal links of the IBFRD to exert the inchworm walking. The corresponding reference trajectories emulate an inchworm in a regular gait cycle. Experimental results compare the H-ADRC with a proportional derivative and a proportional–integral–derivative (PID) controller. The obtained results show that the proposed H-ADRC enforces a reduced tracking error than the other controllers used for validation.
18 March 2021 in IEEE Transactions on Pattern Analysis and Machine Intelligence
This manuscript presents the design of a deep differential neural network (DDNN) for pattern classification. First, we proposed a DDNN topology with three layers, whose learning laws are derived from a Lyapunov analysis, justifying local asymptotic convergence of the classification error and the weights of the DDNN. Then, an extension to include an arbitrary number of hidden layers in the DDNN is analyzed. The learning laws for this general form of the DDNN offer a contribution to the deep learning framework for signal classification with biological nature and dynamic structures. The DDNN is used to classify electroencephalographic signals from volunteers that perform an identification graphical test. The classification results show exponential growth in the signal classification accuracy from 82 with one layer to 100 with three hidden layers. Working with DDNN instead of static deep neural networks (SDNN) represents a set of advantages, such as processing time and training period reduction up to almost 100 times, and the increment of the classification accuracy while working with less hidden layers than working with SDNN, which are highly dependent on their topology and the number of neurons in each layer. The DDNN employed fewer neurons due to the induced feedback characteristic.
This work presents the MuscleTracker Hand Movement dataset, containing Surface Electromyography (sEMG) data from the right arm of 49 healthy subjects without neuromuscular or cardiovascular issues. Subjects performed five hand movements—pronation with extended fingers, flexion, extension, pronation with flexed fingers, and relaxation—while standing, with one hand palm-down. Data was recorded from two sEMG channels using Biopac MP36 (1000 Hz) and MuscleTracker (512 Hz), with three and four repetitions per device, respectively, for each movement. The dataset includes 825 samples, along with subject details such as gender, age, physical condition, and, for MuscleTracker subjects, anthropometric measurements. This data supports machine-learning development for classifying hand gestures in sEMG signals, with applications in prosthetics control and human-computer interaction. In addition, validation experiments were performed to validate the database and stablish a comparison baseline.