Optimal Trajectory Planning in a Vertically Undulating Snake Locomotion using Contact-implicit Optimization

TL;DR

This paper develops a simplified, mathematically rigorous model for snake robot locomotion that incorporates contact dynamics, validated through simulations and experiments, to improve trajectory planning in complex contact-rich environments.

Contribution

It introduces a reduced-order model based on differential inclusion mathematics for snake locomotion, bridging the gap between simple and complex contact interaction models.

Findings

Model accurately predicts snake robot movement.

Validated model through simulations and experiments.

Enables optimized trajectory planning in contact-rich scenarios.

Abstract

Contact-rich problems, such as snake robot locomotion, offer unexplored yet rich opportunities for optimization-based trajectory and acyclic contact planning. So far, a substantial body of control research has focused on emulating snake locomotion and replicating its distinctive movement patterns using shape functions that either ignore the complexity of interactions or focus on complex interactions with matter (e.g., burrowing movements). However, models and control frameworks that lie in between these two paradigms and are based on simple, fundamental rigid body dynamics, which alleviate the challenging contact and control allocation problems in snake locomotion, remain absent. This work makes meaningful contributions, substantiated by simulations and experiments, in the following directions: 1) introducing a reduced-order model based on Moreau's stepping-forward approach from differential inclusion mathematics, 2) verifying model accuracy, 3) experimental validation.