Design and adaptive control of a kinematically redundant robot with enhanced trajectory tracking for climbing in tight spaces

This work investigates the design and adaptive control of a miniature robot with multi-modal locomotion which has the ability to climb inside train bogies for inspection purposes. We propose and analyse a kinematically redundant mechanism with six 2-DOF couple joints. The robot can squeeze through narrow spaces and also climb on surfaces with transitions, irregularities and discontinuities. The unique design allows desirable self-motion close to obstacles but imposes strict requirements in motion control and precise path following. This paper applies such redundancy and self-motion by constructing an adaptive controller with time-varying safety constraints for all twelve joints of the mechanism. The control strategy relies on the time-varying Barrier Lyapunov Function to bound the trajectory error. It also deploys an adaptive radial basis function neural network to estimate the system parameters of the robot. Various simulation experiments show that the proposed controller satisfies all safety and physical joint constraints. It also minimises trajectory tracking error irrespective of initial conditions, disturbances, and unmodelled dynamic effects. Finally, we compare the tracking results with those obtained by a Feedback Linearisation controller and a Quadratic Lyapunov Function-based controller. Results demonstrate enhanced locomotion and trajectory tracking for collision-free manoeuvring in tight spaces.