Bionic Control of Cheetah Bounding with a Segmented Spine.

Area of Science:

• Robotics
• Bio-inspired Engineering
• Biomechanics

Background:

• Cheetahs exhibit unique bounding gaits characterized by spinal and leg segmentation.
• Replicating these complex movements in robotic systems presents significant control challenges.
• Existing robotic models often lack the dynamic adaptability of biological locomotion.

Purpose of the Study:

• To develop a bio-inspired robotic cheetah model capable of mimicking natural bounding gait.
• To design and implement a state-machine-based controller inspired by cerebellar functions for locomotion and balance.
• To analyze the stability and dynamic properties of the realized bounding gait.

Main Methods:

• A cheetah robot was constructed with segmented spine and leg joints, using parameters from real cheetahs.
• A bio-inspired controller, mimicking the cerebellum, was developed using a state-machine approach.
• Haptic sensors, proprioception, and vestibular feedback were integrated for phase transition detection and trunk angle perception.
• A delay feedback control method was employed for leg joint motion planning to stabilize trunk pitching.

Main Results:

• The robotic cheetah successfully realized a cyclic bounding gait with biological properties.
• The bio-inspired controller effectively actuated and stabilized the bounding gait.
• Stability and dynamic characteristics of the cheetah bounding gait were elaborately analyzed.

Conclusions:

• The study demonstrates the feasibility of replicating cheetah bounding gait using a bio-inspired controller and a segmented robotic model.
• Cerebellum-inspired control strategies are effective for achieving stable and dynamic locomotion in legged robots.
• The developed model provides a platform for further research into animal locomotion and robotic control.