InverseMuscleNET: Alternative Machine Learning Solution to Static Optimization and Inverse Muscle Modeling

InverseMuscleNET, a machine learning model, is proposed as an alternative to static optimization for resolving the redundancy issue in inverse muscle models. A recurrent neural network (RNN) was optimally configured, trained, and tested to estimate the pattern of muscle activation signals. Five biomechanical variables (joint angle, joint velocity, joint acceleration, joint torque, and activation torque) were used as inputs to the RNN. A set of surface electromyography (EMG) signals, experimentally measured around the shoulder joint for flexion/extension, were used to train and validate the RNN model. The obtained machine learning model yields a normalized regression in the range of 88–91% between experimental data and estimated muscle activation. A sequential backward selection algorithm was used as a sensitivity analysis to discover the less dominant inputs. The order of most essential signals to least dominant ones was as follows: joint angle, activation torque, joint torque, joint velocity, and joint acceleration. The RNN model required 0.06 s of the previous biomechanical input signals and 0.01 s of the predicted feedback EMG signals, demonstrating the dynamic temporal relationships of the muscle activation profiles. The proposed approach permits a fast and direct estimation ability instead of iterative solutions for the inverse muscle model. It raises the possibility of integrating such a model in a real-time device for functional rehabilitation and sports evaluation devices with real-time estimation and tracking. This method provides clinicians with a means of estimating EMG activity without an invasive electrode setup.

MuscleNET: Mapping Electromyography to kinematic and dynamic biomechanical variables by machine learning

Objective: This paper proposes machine learning models for mapping surface electromyography (sEMG) signals to regression of joint angle, joint velocity, joint acceleration, joint torque, and activation torque. Approach: The regression models, collectively known as MuscleNET, take one of four forms: ANN (Forward Artificial Neural Network), RNN (Recurrent Neural Network), CNN (Convolutional Neural Network), and RCNN (Recurrent Convolutional Neural Network). Inspired by conventional biomechanical muscle models, delayed kinematic signals were used along with sEMG signals as the machine learning model's input; specifically, the CNN and RCNN were modeled with novel configurations for these input conditions. The models' inputs contain either raw or filtered sEMG signals, which allowed evaluation of the filtering capabilities of the models. The models were trained using human experimental data and evaluated with different individual data. Main results: Results were compared in terms of regression error (using the root-mean-square) and model computation delay. The results indicate that the RNN (with filtered sEMG signals) and RCNN (with raw sEMG signals) models, both with delayed kinematic data, can extract underlying motor control information (such as joint activation torque or joint angle) from sEMG signals in pick-and-place tasks. The CNNs and RCNNs were able to filter raw sEMG signals. Significance: All forms of MuscleNET were found to map sEMG signals within 2 ms, fast enough for real-time applications such as the control of exoskeletons or active prostheses. The RNN model with filtered sEMG and delayed kinematic signals is particularly appropriate for applications in musculoskeletal simulation and biomechatronic device control.

Simulation of Upward Jump Control for One-Legged Robot Based on QP Optimization

An optimization framework for upward jumping motion based on quadratic programming (QP) is proposed in this paper, which can simultaneously consider constraints such as the zero moment point (ZMP), limitation of angular accelerations, and anti-slippage. Our approach comprises two parts: the trajectory generation and real-time control. In the trajectory generation for the launch phase, we discretize the continuous trajectories and assume that the accelerations between the two sampling intervals are constant and transcribe the problem into a nonlinear optimization problem. In the real-time control of the stance phase, the over-constrained control objectives such as the tracking of the center of moment (CoM), angle, and angular momentum, and constraints such as the anti-slippage, ZMP, and limitation of joint acceleration are unified within a framework based on QP optimization. Input angles of the actuated joints are thus obtained through a simple iteration. The simulation result reveals that a successful upward jump to a height of 16.4 cm was achieved, which confirms that the controller fully satisfies all constraints and achieves the control objectives.

MATLAB Simulation for Trajectory/Path Efficiency Comparison Between Robotic Manipulators

This paper examines results of MATLAB simulation for three robotic arm configurations to determine their abilities in trajectory/path tracing. The trajectories tested were a spline curve, a circle, and square. The joint position vs. time, joint velocity vs. time, joint acceleration vs. time, and the cartesian position vs. time were plotted to compare their agility.

Online optimization based adaptive tracking control for redundant manipulators with model uncertainties

Tracking control of robot manipulators is always a fundamental problem in robot control, especially for redundant manipulators with higher DOFs. This problem may become more complicated when there exist uncertainties in the robot model. In this paper, we propose an adaptive tracking controller considering the uncertain physical parameters. Based on the coordinate feedback, a Jacobian adaption strategy is firstly built by updating kinematic parameters online, in which neither cartesian velocity nor joint acceleration is required, making the controller much easier to built. Using the Pseudo-inverse method of Jacobian, the optimal repeatability solution is achieved as the secondary task. Using Lyapunov theory, we have proved that the tracking errors of the end-effector asymptotically converge to zero. Numerical simulations are provided to validate the effectiveness of the proposed tracking method.

Tri-Criteria Optimization Motion Planning at Acceleration-Level of Dual Redundant Manipulators

SUMMARYIn order to solve joint-angle drift problem of dual redundant manipulators at acceleration-level, an acceleration-level tri-criteria optimization motion planning (ALTC-OMP) scheme is proposed, which combines the minimum acceleration norm, repetitive motion planning, and infinity-norm acceleration minimization solutions via weighting factor. This scheme can resolve the joint-angle drift problem of dual redundant manipulators which will arise in single criteria or bi-criteria scheme. In addition, the proposed scheme considers joint-velocity joint-acceleration physical limits. The proposed scheme can not only guarantee joint-velocity and joint-acceleration within their physical limits, but also ensure that final joint-velocity and joint-acceleration are near to zero. This scheme is realized by dual redundant manipulators which consist of left and right manipulators. In order to ensure the coordinated operation of manipulators, two motion planning problems are reformulated as two general quadratic program (QP) problems and further unified into one standard QP problem, which is solved by a simplified linear-variational-inequalities-based primal-dual neural network at the acceleration-level. Computer-simulation results based on dual PUMA560 redundant manipulators further demonstrate the effectiveness and feasibility of the proposed ALTC-OMP scheme to resolve joint-angle drift problem arising in the dual redundant manipulators.

Study and Analysis of Pounding Effect between Adjacent RC Buildings

Pounding occurs when the adjacent buildings start vibration out of phase during the seismic activity which causes the collision between the adjacent structures. Due to higher cost of land in cities people have tendency to attach the buildings at property line. Earthquakes can cause pounding when adjacent buildings have little gap or no gap providing separation. Due to pounding effect structural and non – structural damage may occur in the adjacent buildings. The main objective of this research is to assess the seismic response of common residential RC buildings that has been constructed with no gap with the adjacent structures and to find the minimum gap requirement for the commonly constructed buildings of Nepal. For this study two different cases with varying separation distance between adjacent buildings have been considered. First case is the adjacent buildings having equal storey height but different number of stories. It includes models having 4 and 2 stories and 4 and 3 stories. Second case is the adjacent buildings having unequal storey height but same number of stories. It includes models having 3 and 3 stories and 4 and 4 stories. In both cases adjacent buildings have same material & sectional properties. Non-linear dynamic analysis is performed using El-centro earthquake data as ground motion. Gap element has been used to simulate the pounding force between buildings. Adjacent buildings having different overall height are modelled in SAP 2000 v 15 using gap element for pounding study. The seismic responses in terms of joint displacement, joint acceleration, pounding force are presented. Joint displacement and joint acceleration comparison for both pounding and no pounding cases are presented. Gap calculation from NBC and IS code, ABS and SRSS method was compared with gap required to avoid pounding force between adjacent structures and appropriate gap was recommended.