Optimization of Exoskeleton Trajectory Toward Minimizing Human Joint Torques.

Area of Science:

• Robotics
• Biomechanics
• Human-Robot Interaction

Background:

• Exoskeleton robots require precise kinematic guidance for effective operation.
• Reducing human power consumption is a key goal in exoskeleton design.

Purpose of the Study:

• To develop an optimal trajectory generation method for lower-limb exoskeletons.
• To minimize human joint torques and enhance walking efficiency.

Main Methods:

• Human joint angles computed using neighborhood field optimization (NFO).
• Inverse dynamic analysis with a seven-link human-exoskeleton model.
• Back propagation neural network (BPNN) for accelerated analysis.
• Fourier series perturbation and NFO for trajectory optimization.

Main Results:

• Achieved high accuracy in predicting human trajectory (RMSE < 5 mm) and ground reaction forces (RMSE < 3 kN).
• Generated trajectories preserve individual walking patterns and anticipate motion (mean leading value 4.6%).
• Significantly reduced joint torque across various gait phases.

Conclusions:

• The proposed method offers accurate and efficient trajectory generation for lower-limb exoskeletons.
• This approach is suitable for computational applications and designing energy-efficient assistive devices.
• Contributes valuable insights into human-exoskeleton interaction dynamics.