Horse-like walking, trotting, and galloping derived from kinematic Motion Primitives (kMPs) and their application to walk/trot transitions in a compliant quadruped robot.

This research paper discusses a method to mimic horse-like movements such as walking, trotting, and galloping in a four-legged robot. The researchers utilized a mathematical technique known as principal component analysis on a horse’s joint movement data to get a set of four kinematic Motion Primitives (kMPs). These kMPs were then used to create stable movements in a robot. The study also explores the optimal walking and trotting frequencies and presents a method to transition between different gaits successfully.

Principal Component Analysis and Kinematic Motion Primitives (kMPs)

• The study first focuses on understanding the natural movement of horses. The researchers gathered data on joint trajectories of horses during their movement.
• Using principal component analysis, a statistical procedure that utilizes orthogonal transformation to convert a set of observations into a set of values of linearly uncorrelated variables, they were able to derive four kinematic Motion Primitives (kMPs).
• These kMPs represent the fundamental motion patterns of horse-like walking, trotting, and galloping.

Application of kMPs in Quadruped Robot

• After establishing the kMPs, the study then applies these to a four-legged (quadruped) robot. The goal is to emulate horse-like movement in the robot.
• The application of the derived kMPs resulted in valid, stable gaits in the compliant quadruped robot, thus validating the efficacy of the extracted kMPs.

Gait Frequency Scaling

• The researchers also conducted tests to determine the optimal gait frequency for the robot’s movements.
• The results point towards a speed optimal walking frequency around 3.4 Hz and an optimal trotting frequency around 4 Hz.

Walk/Trot Transitions

• In addition to individual gaits, the paper also delves into transitions between different gaits, particularly the transition from walking to trotting.
• The researchers propose a criterion to synthesize these transitions smoothly and then successfully test this on the robot.

Performance Evaluation

• The final part of the research evaluates the performance of the robot as transitions are scaled in frequency.
• This is carried out by analyzing roll and pitch angle phase plots, which show the robot’s orientation or attitude at various phases of the gait transition.