A Predictive Technique for the Real-TimeTrajectory Scaling under High-Order Constraints

Modern robotic systems must be able to react tounexpected environmental events. To this purpose, planningtechniques for the real-time generation/modification of trajec-tories have been developed in recent times. In the frequent caseof applications which require following a predefined path, theassigned timing law must be inspected in real time so as to verifywhether it satisfies the system constraints or, conversely, if it mustbe scaled in order to obtain a feasible trajectory. The problemhas been addressed in several ways in the literature. One of theknown approaches, based on the use of nonlinear filters, is revisedin this paper in order to return feasible solutions under anycircumstances. Differently from alternative strategies, it managesconstraints up to the torque derivatives and has evaluation timescompatible with the ones required by modern control systems.The proposed technique is validated through simulations and realexperiments. Comparisons are proposed with an algorithm basedon a model predictive technique and with an alternative scalingsystem.