Effect of Gait Speed on Sensing During Locomotion Performance of a Quadrupedal Legged Robot

Robot-Assisted exploration can advance our knowledge of Earth and other Planetary bodies. Historically, terrain measurements require dedicated instruments and were time consuming. Recent advances show that legged robots with low gear ratio joints are capable of measuring terrain properties from their legs during locomotion, providing unprecedented sampling density. Yet, how gait choices impact sensing accuracy remains unclear. Here we quantified the terrain sensing accuracy with 3 different quadrupedal gaits from the least to most dynamic: stand-n-sense, crawl-n-sense, and trot. For each gait, we quantify its ability to measure substrate penetration resistance and heterogeneity, as the robot locomoted over a laboratory transect consisting of rigid surface, soft sand, and sand with synthetic surface crusts. Results shown that while the ability to measure relative substrate penetration resistance was relatively insensitive to gait speed, the absolute error increased with gait speed due to the increased inertial effect. Our findings provided important insights for legged robots to adapt its gait speed to balance sensing accuracy and efficiency, allowing enhanced scientific outcomes in terrestrial and planetary exploration.