Adaptive Path Following of the Biomorphic Hyper-Redundant Snake Robot in Unconstructed Environment
Authors
Yang, Weixin
Issue Date
2019
Type
Dissertation
Language
Keywords
Adaptive Controller , Bio-inspired , Model-based Control , Pathfinding , Path-following , Snake Robot
Alternative Title
Abstract
The efficient movement of biological snakes in various environments is the result of a long evolution. It is a challenge for researchers to make snake robots gain high efficiency, athletic agility, and environmental adaptability of snakes due to the complicated motion control issues caused by under-actuation, high-dimension nonlinearity, and uncertainties in kinematics, dynamics and interactions with complex environments. This work relies on a custom-built snake robot that can mimic the locomotion capabilities of snakes, such as serpentine, sidewinding, and rectilinear motions, which stems from our prior eorts on its modular mechanical design, advanced electronic system, and efficient driving capabilities. Based on this robot system, we moved forward to study the closed-loop control strategy for a class of snake robots and implemented pathfinding and following, which is the challenging and practical problem in control of snake robots. In this dissertation, we first investigate straight line path-following problems for a class of planar underactuated snake robots. For this purpose, an adaptive controller that can autonomously drive the snake robot moving on ground with unknown and varied friction coefficients is designed and validated. Combining with time-varying Line of Sight (LOS) and Parametric Cubic Spline Interpolation (PCSI) path planning methods, snake robots tracking arbitrary paths is further explored and the comprehensive curve path following is realized. Last, the perception-aware pathfinding and following for snake robots in unmodeled and unknown environments are also studied, and a simple but efficient control methodology is developed. The proposed method is validated by extensive simulations and experimental results. More importantly, in this dissertation, we addressed to achieve a generic adaptive controller for motion control of a class of snake robots with uncertain dynamics. The advanced controller is proved to be stable and also it does not require a high gain for reaching ideal control performance. With the proposed controller and path following methods, snake robots are capable of improving mobility in different environments, which promises the potential of using snake robots in diverse real-world applications, like the earthquake rescue, narrow space surveillance and even overwater or underwater explorations.