Locomotion Stability Classifier
Objective
Memprediksi apakah pergerakan humanoid stabil atau tidak.
Input Features
- step_length_cm
- stride_frequency_hz
- center_of_mass_shift_cm
- ground_reaction_force_n
- terrain_type_index
- imu_balance_variance
Output
Architecture
- Dense(128) + ReLU
- Dense(64) + ReLU
- Dropout(0.3)
- Dense(1) + Sigmoid
Loss
Binary Crossentropy
Metrics
Accuracy, Precision, Recall, F1
