Rest-to-Rest Trajectory Planning for Underactuated Cable-Driven Parallel Robots Using Input Shaping: Preliminary Results
摘要
Underactuated robots have fewer actuators than degrees of freedom (DoF) of the end-effector. This limits the number of configuration variables that can be controlled and results in possibly large oscillations during motion. Rest-to-rest planners are used to generate a trajectory of a subset of the configuration variables, and allow the system to transition between static poses. Such planners solve a boundary value problem with pose and twist constraints in the start and end states. The problem is solvable only if a nominal trajectory is modified by adding some free parameters, which are found by the solver so that constraints are satisfied. In this work, input shaping is used to modify the nominal trajectory. The problem of finding the shaper parameters is formulated as a feasibility problem, as this formulation allows nonlinear constraints to be enforced on variables. The proposed approach is validated through experimental tests on a planar 3-DoF robot suspended by two cables.