Ti-6Al-4V titanium alloy is widely utilized in aerospace components, such as torque tubes and turbine blades, due to its outstanding strength-to-weight ratio and corrosion resistance. However, controlling surface roughness during machining is challenging because the alloy's low thermal conductivity and high chemical reactivity result in unpredictable variations in Ra. Traditional contact-based measurement methods are not only time-consuming but also incompatible with in-process monitoring, creating a disconnect between production and quality control. This study introduces a CNN-LSTM hybrid model for predicting surface roughness in Ti-6Al-4V shape machining, utilizing multi-sensor CNC data. The model effectively captures spatial correlations among nine sensors and temporal dependencies in sequential operations. We implement a stratified time-series split validation that maintains chronological order while ensuring a representative distribution of Ra values, reflecting realistic deployment conditions. Data were collected from machining tests on features of a torque tube part, comprising 4,154 samples with Ra values ranging from 0.57 to 0.74 μm. The CNN-LSTM model achieved R² = 0.8512, RMSE = 0.0199 μm, and MAE = 0.0096 μm, outperforming Random Forest, XGBoost, and standalone neural networks. These results demonstrate the feasibility of non-contact, in-process surface roughness prediction in aerospace manufacturing, facilitating proactive quality control without interrupting operations.

Falls are common among older people. Age-related changes in toe strength and force steadiness may increase fall risk. This study aimed to evaluate the performance of a fall risk prediction model using toe strength and force steadiness data as input variables. Participants were four healthy adults (25.5±1.7 yrs). To indirectly reproduce physical conditions of older adults, an experiment was conducted by adding conditions for weight and fatigue increase. The maximal strength (MVIC) was measured for 5 s using a custom toe dynamometer. For force steadiness, toe flexion was measured for 10 s according to the target line, which was 40% of the MVIC. A one-leg-standing test was performed for 10 s with eyes-opened using a force plate. Deep learning experiments were performed with seven conditions using long short-term memory (LSTM) algorithms. Results of the deep learning model were randomly mixed and expressed through a confusion matrix. Results showed potential of the model"s fall risk prediction with force steadiness data as input variables. However, experiments were conducted on young adults. Additional experiments should be conducted on older adults to evaluate the predictive model.