As mechanical design guidance, the AI learns knowledge from mechanical design databases by using neural networks. The AI can find unknown parameters for mechanical design and evaluate the performance of mechanical design. We are researching a design guidance algorithm of the AI and developing a framework of mechanical design guidance.

Each person has a different pattern of body movements, so tuning control parameters for each person are needed to maximize the robot's performance. Human-in-the-loop (HIL) optimization showed the possibility of obtaining the best control parameters for each individual in a well-refined environment (e.g., laboratory). For the extension of HIL in real life, we are studying advanced HIL to optimize the control parameters of wearable robots in mobile situations using human biomechanical signals acquired in real-time.