Bionic Energy-Efficient Inverse Kinematics Method Based on Neural Networks for the Legs of Hydraulic Legged Robots

Area of Science:

• Robotics
• Artificial Intelligence
• Biomimetics

Background:

• Hydraulic legged robots offer high load capacity and power density, crucial for intelligent mobile platforms.
• High energy consumption in these robots limits their endurance and operational efficiency.
• Mammalian autonomous energy efficiency inspired a new approach to robot motion planning.

Purpose of the Study:

• To propose a bionic energy-efficient inverse kinematics method for hydraulic legged robots with redundant degrees of freedom (RDOFs).
• To enhance the energy efficiency and operational performance of hydraulic legged robots.
• To reduce energy dissipation and improve the long-term viability of these robotic systems.

Main Methods:

• Utilized dynamic programming (DP) to determine optimal joint configurations minimizing energy loss for generating training data.
• Developed an energy-efficient inverse kinematics model using neural networks (NNs) to simulate mammalian energy efficiency.
• Employed extensive comparative experiments to validate the proposed method's effectiveness.

Main Results:

• The proposed bionic energy-efficient inverse kinematics method (EIKNN) significantly reduces energy dissipation in hydraulic legged robots.
• Demonstrated superior performance compared to existing methods in energy-efficient motion planning.
• Validated the method's effectiveness through extensive comparative experiments.

Conclusions:

• The EIKNN method offers a significant reduction in energy consumption for hydraulic legged robots.
• This approach provides a foundational advancement for developing highly efficient and sustainable hydraulic legged robots.
• The study paves the way for more viable long-duration operations of intelligent mobile platforms.