Terrain characterization and locomotion adaptation in a small-scale lizard-inspired robot

TL;DR

A principled framework for perception and control in small-scale locomotion is established and this work designs a simple linear feedback controller that modulates body phase and substantially improves locomotion performance on terrains with unknown depth.

Abstract

Unlike their large-scale counterparts, small-scale robots are largely confined to laboratory environments and are rarely deployed in real-world settings. As robot size decreases, robot-terrain interactions fundamentally change; however, there remains a lack of systematic understanding of what sensory information small-scale robots should acquire and how they should respond when traversing complex natural terrains. To address these challenges, we develop a Small-scale, Intelligent, Lizard-inspired, Adaptive Robot (SILA Bot) capable of adapting to diverse substrates. We use granular media of varying depths as a controlled yet representative terrain paradigm. We show that the optimal body movement pattern (ranging from standing-wave bending that assists limb retraction on flat ground to traveling-wave undulation that generates thrust in deep granular media) can be parameterized and approximated as a linear function of granular depth. Furthermore, proprioceptive signals, such as joint torque, provide sufficient information to estimate granular depth via a K-Nearest Neighbors classifier, achieving 95% accuracy. Leveraging these relationships, we design a simple linear feedback controller that modulates body phase and substantially improves locomotion performance on terrains with unknown depth. Together, these results establish a principled framework for perception and control in small-scale locomotion and enable effective terrain-adaptive locomotion while maintaining low computational complexity.

Figures (9)

• Figure 1: Small-scale, Intelligent, Lizard-inspired, Adaptive Robot (SILA Bot) a) Lizard robot with 7 actuated servos: 3 body servos and 4 leg servos. b) Wirelessly configured lizard robot walking through soil with unknown depth at a farm.
• Figure 2: Experimental setup and gait prescription. (a) Robot test arena consisting of an adjustable area containing granular media, surrounded by Vicon Vero motion-capture cameras for kinematic tracking. (b) Schematic of the robot traversing granular media. (c) Body joint angles $\alpha_1$–$\alpha_3$ defined along the body of the SILA Bot. L, R, F, and H stand for left, right, fore, and hind legs of the robot. (d) Cosine functions used to prescribe body joint trajectories; $A$ denotes the amplitude, and $\phi$ denotes the phase offset between consecutive joints. (e) Limb contact sequence for LH, LF, RF, and RH legs of the robot. Shaded regions indicate ground contact. The robot employs a 50% duty-cycle trotting gait in the current configuration.
• Figure 3: Deep granular media: body phase offset and locomotion speed. (a) Snapshots of SILA Bot executing a gait with body phase offset $\phi = -\pi/3$ on deep granular media. (b) Experimentally tracked displacement as a function of gait fraction over five cycles. Three gaits are compared (yellow: $\phi = 0$; red: $\phi = -\pi/3$; blue: $\phi = -\pi/2$). (c) Average forward speed (units: Body length per cycle, BL/C) as a function of body phase offset. Statistically significant differences (Paired T-Test) are observed between $\phi = -5\pi/12$ and $\phi = -\pi/3$ ($p < 0.01$), and between $\phi = -\pi/3$ and $\phi = -\pi/12$ ($p < 0.05$). The inset shows the robot traversing 40 mm-deep granular media.
• Figure 4: Shallow granular media: body phase offset and locomotion speed. (a) (left) Snapshots of the SILA Bot executing a gait with body phase offset $\phi = 0$ on flat ground. (right) Average forward speed (units: Body length per cycle BL/C) of the SILA Bot on flat ground. Inset: illustration of the robot traversing flat terrain. The red star corresponds to the best body phase, depicted left. (b) (left) Snapshots of the SILA Bot executing a gait with body phase offset $\phi = -\pi/6$ on shallow granular media (depth = 20 mm). (right) Average forward speed (units: BL/C) on shallow granular media. Inset: illustration of the robot traversing 20 mm-deep granular media. The orange star corresponds to the best body phase, depicted left.
• Figure 5: Model Prediction. (a) Left: Diagram of SILA Bot. Red shaded sections indicate contact with the ground. The magnitudes of Coulomb forces are indicated by magenta arrows, and Granular Media (GM) Resistive Force Theory (RFT) magnitudes are indicated by blue arrows. Right: Images of SILA Bot in GM. The bodies experiencing each type of friction at this time are highlighted in their respective colors. (b) An element experiencing GM RFT and another experiencing Coulomb Friction. The magnitudes of the forces are detailed in the plot. (c) Median torque, $\tilde{\tau}_m$, versus Ratio of GM RFT to Coulomb Friction across body joints at $\phi=-\pi/3$. (d) Median torque, $\tilde{\tau}_m$ versus Ratio of GM RFT to Coulomb Friction across body joints at $\phi=0$