Seyed Mehdi Rezaei

Continuous prediction of joint parameters is replacing discrete gait phase control to be the mainstream in rehabilitation robot control field. The sensor-based methods which use inertial measurement units (IMU), surface electromyography (sEMG) and so on are widely used in gaining joint parameters for robot control. However, those methods introduce many sensors attached to patients and affect the walking during training. To reduce the number of sensors needed, a method is proposed to use centre of mass (CoM) horizontal position to predict angles, angular velocities and accelerations of ankle, knee, and hip joints. Long short-term memory (LSTM) is a kind of recurrent neural network (RNN) widely used in predicting time series data. To gain the most suitable model to predict joint parameters of each joint, the performances of Autoencoding-LSTM (combining encoder-decoder with LSTM), CNN-LSTM (combining convolutional neural network with LSTM) and stacked LSTM with different input window sizes on predicting joint parameters of each joint are compared, and the models with optimal performances are selected as a model pack to achieve high quality prediction of joint parameters. The number of sensors needed is reduced by 50% with the accuracy equal to those methods using IMU or sEMG sensors. And the results additionally show that different models perform variously on predicting different joint parameters of different joints.

Grinding is one of the most complex and precise machining operations, with the performance of the grinding wheel being heavily reliant on its surface conditions. Removed chips tend to be forced back to the wheel surface due to the intense heat generated in the grinding zone. The loaded chips block the wheel's grinding capability, resulting in increased grinding force and temperature, wheel chatter, and workpiece thermal degradation. To solve these problems, this research examines the effectiveness of an auxiliary compressed air jet system during the grinding of spk tool steel with an alumina grinding wheel to clean the wheel surface of lodged chips and then the percentage of wheel loading for different tests will also be evaluated. The compressed air nozzle was examined from a variety of perspectives (0, 30, 45, and 60). Grinding with and without a cleaning system, as well as conventional cutting fluid, were also employed as comparisons. In certain output results, input parameters such as grinding wheel speed (20 m/s), cutting depth (20 m), and feed rate (100 mm/s) were maintained constant. When compared to other experiments with different cleaning nozzles, the optimal inclination angle of 45°, which allows the maximum air to reach the interface between the wheel surface and chips, removed chips by pushing them off the wheel surface. Other cleaning nozzles, such as 0°, 30°, and 60°, were less successful. The incidence angle of 45° produced the best results in terms of surface roughness, diametrical grinding wheel wear, G-ratio, tangential forces, specific energy, and wheel loading analysis. Image processing was used to detect and identify chips accumulated between abrasive grains on the wheel surface. To better understand the structure of the wheel and workpiece surface, optical and scanning electron microscopy pictures were taken and analyzed. This research adds to our understanding of the grinding wheel loading mechanism and its dynamics.