Abstract
This paper presents an algorithm for generating the feedrate profile for Computerized Numerically Controlled (CNC) machines with non-Cartesian H-Bot kinematics. Feedrate profile defines the relationship between the end-effector or tool velocity tangent to the toolpath and time or the toolpath's parameter. To maximize machining effectiveness the feedrate should be as high as possible. However, the feedrate, acceleration and jerk of the end-effector and individual machine axes should be within their respective limits. The toolpath is defined as a Non-Uniform Rational B-Spline (NURBS) polynomial curve to ensure smooth motion. This makes the problem more difficult for non-Cartesian machines due to non-linear dependencies between kinematic parameters of the end effector and the machine's axes. In this paper Particle Swarm Optimization (PSO) gradient free algorithm is used to determine the optimal shape of the feedrate profile to achieve shortest travel time within the velocity, acceleration and jerk constraints imposed by the machine's axes. Compared to more common approaches the profile is initialized with a near-optimal shape determined by a feedrate limit curve and then optimized to the final shape. The method is experimentally verified on an actual H-Bot gantry plotter. Presented results show that the proposed method effectively generates a feedrate profile within the imposed limits.