An efficient processor-based online generation of time-optimal trajectories

We present a novel method to online generation of time-optimal acceleration, velocity, and position trajectory values for servo-control systems using low-cost fixed-point processors with no on-chip multipliers. As in the case of many product designs, reducing system cost is a key constraint in our method. We address this constraint by optimizing required memory size, computational time, and introducing multiplication-free recursive algorithms most suitable for real-time implementation on low-cost microcontrollers. An efficient search algorithm is presented optimizing the trajectories with respect to total move time for any distance subject to physical constraints of the system. We also introduce a normalized set of first-order difference equations using only integer arithmetic and additions to compute the trajectory values. Complexity of the proposed method is compared with multiplication-based and lookup table methods using three popular microcontrollers with and without an on-chip hardware multiplier