Safety-Aware Nonlinear Model Predictive Control for Physical Human-Robot Interaction
Artemiy Oleinikov, Sanzhar Kusdavletov, Almas Shintemirov, Matteo Rubagotti
Abstract
This letter proposes a nonlinear model predictive control (NMPC) approach for real-time planning of point-to-point motions of serial robot manipulators that share their workspace with a human. The NMPC law solves a nonlinear program online, based on a kinematic model, and guarantees safety by constraining the robot speed within the time-varying bounds determined by the speed-and-separation-monitoring (SSM) principle. Closed-loop stability is proven in detail, and the performance (in terms of productivity) of the proposed method is tested against standard SSM schemes via experiments on a Kinova Gen3 robot.
Topics & Concepts
WorkspaceModel predictive controlKinematicsControl theory (sociology)Nonlinear systemRobotComputer scienceNonlinear modelStability (learning theory)Control engineeringControl (management)EngineeringArtificial intelligenceMachine learningPhysicsQuantum mechanicsClassical mechanicsProsthetics and Rehabilitation RoboticsAdvanced Control Systems OptimizationRobot Manipulation and Learning