Adaptive Locomotion of Planar Snake Robots: Body Shape Control and Path-Following

Snake robots constitute a class of hyper-redundant, articulated systems inspired by their biological counterpart. Their many degrees of freedom make them challenging to control, but provide unique locomotion capabilities in irregular and challenging environments. However, the inherent complexity of full order models constrains the design of high-performance model-based controllers, especially for advanced tasks such as path-following, which is often based on a simplified control-oriented model. While this model allows for tractable controller design, it suffers from a structural mismatch with the full system and requires precise parameter knowledge – such as rotational drag coefficients – that is not available in practice. This thesis proposes two adaptive control schemes to address these challenges. The first focuses on body-shape control when friction coefficients are partially known, completely unknown, or vary across the environment. The second addresses path-following with adaptation of the rotational drag coefficients of the simplified model. Combined, they form a unified adaptive approach to snake robot locomotion in practical applications. The organization of this work is deliberately structured to mirror the hierarchical design process of the control architecture, starting from the inner building blocks and progressively adding more sophisticated layers in a cascaded structure. Each scheme is formally analyzed, and numerical simulation studies demonstrate the tracking performance and effective compensation despite parametric uncertainty.