From Continuous Design to Delay-Aware Discrete Synthesis: Guaranteed High-Bandwidth Joint Control for PMSM Drives

Edmundo Pozo Fortunić; Mehmet C. Yildirim; Sami Haddadin

cs.ROAug 14, 2026

From Continuous Design to Delay-Aware Discrete Synthesis: Guaranteed High-Bandwidth Joint Control for PMSM Drives

Authors: Edmundo Pozo Fortunić, Mehmet C. Yildirim, Sami Haddadin

Organizations: Chair of Robotics and Systems Intelligence and the Munich Institute of Robotics and Machine Intelligence (MIRMI), Technical University of Munich, Germany · Mohamed Bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates

Abstract

The increasing dynamic demands of modern robotic joints require current controllers to achieve high bandwidth over wide operating ranges of speed, acceleration, and torque, where communication, computation, and discrete-time effects can no longer be neglected. Conventional PMSM current controllers are typically designed in continuous time and subsequently discretized, leaving the sampling frequency and the impact of implementation delays largely to heuristic selection and iterative validation. This paper introduces a task-aware, delay-extended discrete-time joint model that explicitly accounts for physical communication and computation delays and enables direct synthesis of a discrete PI current controller with prescribed bandwidth and delay guarantees throughout the operating envelope. The framework analytically determines the minimum required sampling frequency, controller gains, and DC-link voltage needed to satisfy the specified motor and joint performance. Simulations across a range of dynamic requirements validate the methodology and demonstrate substantially reduced sampling-frequency and DC-link-voltage requirements compared with conventional continuous-time-based design. Experiments on a newly developed custom robotic joint further validate the proposed framework under real embedded implementation conditions.

Figures & tables

Fig. 1: Design Synthesis for a PMSM embedded discrete controller based on its joint-level specifications.

Fig. 2: Magnitudes (top) and Phases (bottom) of the closed loop current controllers 26 with different $Kp$ parameters. Each $\Omega_{b}w$ is highlighted and marked on the top axis. The bottom axis has the equivalent $f_{bw}/f_{s}$ ratio. Each corresponding phase is also marked.

Fig. 3: Boundaries required to select the correct minimum sampling frequency

f_{s}

that ensure the proper operation of two motors (see Tab. I ) for the required conditions of bandwidth, maximum speed and current ripple.

Parameter	Symbol	Unit	$M_{1}$	$M_{2}$
Phase Resistance	$R$	$\Omega$	12.65e-2	7.55e-2
Phase Inductance	$L$	$\mathrm{H}$	3.21e-5	5.95e-5
Pole Pairs	$n_{p}$	$-$	2	10
Torque Constant Back-EMF Constant	$k_{t}$	$\mathrm{N\cdot m/A}$ $(\mathrm{V\cdot s/rad})$	2.81e-2	3.1e-2
Rated Current	$i_{\mathrm{max}}$	$\mathrm{A}$	6.6	10
Rated Speed	$\omega_{\mathrm{max}}$	$\mathrm{rad/s}$	924	717

TABLE I: Motor parameters of the robotic joint prototype.

Appendix figures & tables1 asset

Supplementary material from the paper’s appendix.

Appendix

$\begin{gathered}\textbf{Actuator}\\ \textbf{Model}\end{gathered}$	Closed-Loop Transfer Function $\boldsymbol{H^{A,t}_{cl}(s)}$	$\boldsymbol{\omega_{n}^{2}}$	$\boldsymbol{\zeta}$	Additional pole(s)
$\begin{gathered}\text{Ideal}\\[2.84526pt] 1\end{gathered}$	$\dfrac{1}{b_{\theta}}\dfrac{k+ds}{s^{2}+2\zeta\omega_{n}s+\omega_{n}^{2}}$	$\dfrac{k}{J_{\theta}}$	$\dfrac{d_{\theta}}{2\omega_{n}}$	–
$\begin{gathered}\text{Delay}\\[2.84526pt] e^{-sT_{d}}\end{gathered}$	$\dfrac{\frac{T_{d}}{2}s+1}{\frac{1}{p}s+1}\!\left(\dfrac{2}{T_{d}p\,b_{\theta}}\right)\dfrac{(k+ds)e^{-sT_{d}}}{s^{2}+2\zeta\omega_{n}s+\omega_{n}^{2}}$	$\dfrac{2}{T_{d}p}\dfrac{k}{J_{\theta}}$	$\dfrac{2\frac{d}{J}-d_{\theta}+\frac{2}{T_{d}}-p}{2\omega_{n}}$	$p=\dfrac{2}{T_{d}}-2\!\left(d_{\theta}-\dfrac{d}{J}\right)+\mathcal{O}(T_{d})$
$\begin{gathered}\text{Finite}\\[2.84526pt] \dfrac{\omega_{bw}}{s+\omega_{bw}}\end{gathered}$	$\dfrac{\omega_{0}}{s+\omega_{0}}\!\left(\frac{\omega_{bw}}{\omega_{0}b_{\theta}}\right)\dfrac{k+ds}{s^{2}+2\zeta\omega_{n}s+\omega_{n}^{2}}$	$\dfrac{\omega_{bw}}{\omega_{0}}\dfrac{k}{J_{\theta}}$	$\dfrac{\frac{d}{J}+\omega_{bw}-\omega_{0}}{2\omega_{n}}$	$\omega_{0}=\omega_{bw}-\left(d_{\theta}-\dfrac{d}{J}\right)+\mathcal{O}(\omega_{bw}^{-1})$
$\begin{gathered}\text{Finite +}\\ \text{Delay}\\[2.84526pt] \dfrac{\omega_{bw}e^{-sT_{d}}}{s+\omega_{bw}}\end{gathered}$	$\dfrac{\omega_{0}}{s+\omega_{0}}\dfrac{\frac{T_{d}}{2}s+1}{\frac{1}{p}s+1}\!\left(\dfrac{2\omega_{bw}}{T_{d}p\,\omega_{0}b_{\theta}}\right)\dfrac{(k+ds)e^{-sT_{d}}}{s^{2}+2\zeta\omega_{n}s+\omega_{n}^{2}}$	$\dfrac{2\omega_{bw}}{T_{d}p\omega_{0}}\dfrac{k}{J_{\theta}}$	$\!\!\dfrac{\frac{d}{J}\!+\!\frac{2}{T_{d}}\!\!-\!p\!+\!\omega_{bw}\!-\!\omega_{0}\!}{2\omega_{n}}$	$\begin{gathered}p=\dfrac{2}{T_{d}}-\!\ \frac{\omega_{bw}\left(2d_{\theta}-\dfrac{d}{J}\right)}{\frac{2}{T_{d}}-\omega_{bw}}+\mathcal{O}\left(\!T_{d},\omega_{bw}^{-1}\!\right)\!\\[2.84526pt] \!\omega_{0}=\omega_{bw}\!-\!\frac{\frac{2}{T_{d}}\!\left(\!d_{\theta}\!-\!\dfrac{d}{J}\!\right)\!+\!\omega_{bw}d_{\theta}}{\frac{2}{T_{d}}\!-\!\omega_{bw}}\!+\!\mathcal{O}\!\left(\!T_{d},\omega_{bw}^{-1}\!\right)\!\end{gathered}$

Appendix

TABLE II: Equivalent actuator models and resulting closed-loop transfer functions.

Feedback Control

Feedback control uses ongoing measurements to adjust actions, stabilize dynamical systems, track goals, and respond to disturbances and constraints.

See all 255 papers

Jul 27, 2026

Input Shaping for Point-to-Point Motion with a Continuum Robot Arm

A cable-driven continuum robot arm is an underactuated mechanism and may suffer residual vibration at the end of a rest-to-rest maneuver. In this work, a time-delay filter is applied as an input shaper to the system to eliminate the excitation of vibratory modes. A non-robust and a robust time-delay filter are designed based on a linear system model and demonstrate improved response compared to a velocity-driven pulse input. Experimental results using the continuum robot validate the application of the input shaper, with reduced overshoot and settling time exemplifying the reduction in oscillation at the end of the maneuver. It is also shown that utilizing the robust shaper further improves the response of the arm in comparison to applying the non-robust shaper. These results are significant towards the precise and robust implementation of continuum robots in applications involving arbitrary end-effector trajectories.

Rodolfo Hdz. Ibarra, Karan Baker, Parsa Molaei +2

Department of Mechanical and Industrial Engineering, Louisiana State University, LA 70803, USA

Robust Control Continuum Robot Control

Ask PDF

May 11, 2026

VRA: Grounding Discrete-Time Joint Acceleration in Voltage-Constrained Actuation

Discrete-time joint acceleration constraints are widely used to enforce position and velocity limits. However, under voltage-constrained electric actuators, kinematically admissible accelerations may be physically unrealizable, exposing a missing execution-level abstraction. We propose Voltage-Realizable Acceleration (VRA), a joint-level acceleration interface that grounds kinematic acceleration in voltage-constrained actuator physics by restricting commanded accelerations to voltage-realizable constraints. Hardware experiments on electric actuators and a wheel-legged quadruped show that VRA removes unrealizable accelerations, restores consistent near-constraint execution, and reduces constraint-induced oscillations.

Lingwei Zhang, Jiaming Wang, Tianlin Zhang +5

Department of Mechanical and Automation Engineering; Hong Kong Embodied AI Lab, The Chinese University of Hong Kong

Robotic Control

Ask PDF

Jul 10, 2026

Residual Physics-Informed Neural Networks for High-Fidelity BLDC Motor Modeling

Accurate dynamics modeling of Brushless DC (BLDC) motors is fundamental to high-performance robotic joint control. This paper presents a Physics-Informed Neural Network (PINN) with a deep residual (ResNet) backbone that learns a continuous-time surrogate of the full six-state BLDC motor dynamics. Given simulation time, applied three-phase voltages, and excitation parameters as inputs, the network directly predicts all motor state variables -- rotor angle, angular velocity, three-phase currents, and winding temperature -- while simultaneously satisfying the governing electromechanical and thermal ODEs through a composite physics-data loss. A curriculum scheduling strategy gradually activates the physics penalty to prevent premature convergence. Training runs are completed in under two minutes on a standard CPU. Crucially, once trained, PINN inference achieves latencies of 0.1--22, mu s per query, up to 118x faster than conventional ODE solvers, making it suitable for real-time observer and control applications.

Haitham El-Hussieny

Department of Mechatronics and Robotics Engineering, Egypt-Japan University of Science and Technology (E-JUST), New Burg El-Arab City, Alexandria, Egypt

Neural Surrogate Modeling Physics-Informed ML

Ask PDF

Abstract

Feedback Control

Explore similar work

Input Shaping for Point-to-Point Motion with a Continuum Robot Arm

VRA: Grounding Discrete-Time Joint Acceleration in Voltage-Constrained Actuation

Residual Physics-Informed Neural Networks for High-Fidelity BLDC Motor Modeling