Multi-robot connective collaboration toward collective obstacle field traversal

TL;DR

The paper addresses collective locomotion for a small swarm of connectable robots navigating obstacle fields with height variations comparable to leg length. It combines experiments with two simple, magnetically or mechanically connected robots and an energy-landscape model to understand how inter-robot connection length $C$ modulates flow versus jam dynamics, predicting optimal ranges such as $C \in [0.86,0.90]\mathrm{UBL}$ for traversability. The key contributions include empirical evidence that modest changes in $C$ drastically alter mobility, a quantitative link between the sign of $\hat{\vec{v}}_1 \cdot \hat{\vec{v}}_2$ and traversal outcomes, and a model that explains these effects via a system energy $E(X,Y,Z,\alpha,\beta,\theta)$ minimized with $E=mgZ$. The authors further demonstrate a model-guided adaptation strategy to switch connection lengths across a multi-segment obstacle field, enabling successful traversal with a simple control paradigm, and discuss implications for scalable, ant-like swarms negotiating diverse environments.

Abstract

Environments with large terrain height variations present great challenges for legged robot locomotion. Drawing inspiration from fire ants' collective assembly behavior, we study strategies that can enable two ``connectable'' robots to collectively navigate over bumpy terrains with height variations larger than robot leg length. Each robot was designed to be extremely simple, with a cubical body and one rotary motor actuating four vertical peg legs that move in pairs. Two or more robots could physically connect to one another to enhance collective mobility. We performed locomotion experiments with a two-robot group, across an obstacle field filled with uniformly-distributed semi-spherical ``boulders''. Experimentally-measured robot speed suggested that the connection length between the robots has a significant effect on collective mobility: connection length C in [0.86, 0.9] robot unit body length (UBL) were able to produce sustainable movements across the obstacle field, whereas connection length C in [0.63, 0.84] and [0.92, 1.1] UBL resulted in low traversability. An energy landscape based model revealed the underlying mechanism of how connection length modulated collective mobility through the system's potential energy landscape, and informed adaptation strategies for the two-robot system to adapt their connection length for traversing obstacle fields with varying spatial frequencies. Our results demonstrated that by varying the connection configuration between the robots, the two-robot system could leverage mechanical intelligence to better utilize obstacle interaction forces and produce improved locomotion. Going forward, we envision that generalized principles of robot-environment coupling can inform design and control strategies for a large group of small robots to achieve ant-like collective environment negotiation.

Figures (9)

• Figure 1: Animals and robots can utilize physical connections to navigate challenging environments. (a) Ants collectively overcome a large gap by physically connecting with one another. (b, c) Multiple robots form different physical connection configurations to negotiate complex terrains.
• Figure 2: Robot and Experiment setup. (a) The design of each individual robot. (b) A two-robot system connected to each other via an electrical magnet. (c) Locomotion experiment setup, where the two-robot system move across an bumpy terrain comprising of uniformly distributed semispherical boulders. $LF$, $RF$, $RH$, $LH$ denote the left front, right front, right hind, and left hind leg of each robot, respectively. Leg $LF_1$, $RH_1$, $LF_2$, $RH_2$ ("leg group 1") move synchronously, and alternate with the other four legs, $RF_1$, $LH_1$, $RF_2$, $LH_2$ ("leg group 2"). $\theta$ denotes the orientation angle of the two-robot system in the yaw direction.
• Figure 3: Experimentally-measured robot stride length ( i.e., displacement during each cycle) with different robot connection lengths. Markers represent robot stride-wise displacement averaged from all steps from the same connection length. Error bars represent one standard deviation.
• Figure 4: Experimentally-measured CoM velocity of the two-robot system during one stride period, for (a) connection lengths 5.0cm, (b) 5.5cm, and (c) 6.0cm. The red and green shaded color regions represent the collective flowing phase and collective jamming phase, respectively.
• Figure 5: Experimentally-measured robot legs contact positions when the connected robot were during the jamming phase, for 3 representative connection lengths: (top) $C$ = 5.0cm, (middle) 5.5cm, and (bottom) 6.0cm. The robot states in the left diagram illustrated the experimentally-observed robot jamming states. Red, black, green, brown markers represent the contact positions of $LF$, $RH$, $RF$, $LH$ legs on the semispherical boulder (represented as blue circles) from top view. The leg contact positions plotted were measured from the last 5 strides of the 3 trials for each connection length.