arXiv subjects

Annika Raatz

Publications and source records attributed to Annika Raatz.

3 recordsLinked to original sources

arXiv2026

Towards Track Height Control in Additive Manufacturing on Porous Substrates: An Approach for System Identification and Controller Selection

Additive manufacturing is transforming the creation of materials and parts, enabling the design and manufacturing of multifunctional parts with tailored properties. One compelling example is the combination of a porous core and a non-porous surface, which offers significant advantages in terms of weight reduction, material efficiency, strength and wear resistance. Directed Energy Deposition (DED) processes enable the transition from porous core to non-porous part surface by closing the pores. However, the surface-layer of porous substrates presents a substantial challenge to achieving consistent track heights during DED, directly impacting the final part geometry and performance. Therefore, precise control of the deposition process is crucial for realising the full potential of the multifunctional parts. In order to address this challenge, this work presents a conceptual approach for track height control during Laser-based Directed Energy Deposition (DED-LB) on porous substrates. First, the DED-LB process and the associated measurement setup are explained, along with a discussion of their limitations. The focus then shifts to the control problem, beginning with system identification to characterise the process dynamics. Subsequently, a selection of control strategies is investigated, as well as their compatibility with the process. This research provides a foundational step toward reliable additive manufacturing on porous substrates, and thus toward producing multifunctional parts.

eess.SY↗

arXiv2026

Accurate Open-Loop Control of a Soft Continuum Robot Using Visually Learned Latent Dynamics

This work addresses open-loop control of a soft continuum robot (SCR) from video-learned latent dynamics. Visual Oscillator Networks (VONs) from previous work are used, which provide mechanically interpretable 2D oscillator latents through an attention broadcast decoder (ABCD). Open-loop, single-shooting optimal control is performed in latent space to track image-specified waypoints without camera feedback. An interactive SCR live simulator enables design of static, dynamic, and extrapolated targets and maps them to model-specific latent waypoints. On a two-segment pneumatic SCR, Koopman, MLP, and oscillator dynamics, each with and without ABCD, are evaluated on setpoint and dynamic trajectories. ABCD-based models consistently reduce image-space tracking error. The VON and ABCD-based Koopman models attain the lowest mean squared errors (MSEs). Ablation and simulation stress tests confirm that training and architecture choices support open-loop performance, including static holding, stable extrapolated equilibria, and relaxation to rest. To the best of our knowledge, this is the first demonstration of reliable long-horizon open-loop control of a physical soft robot using interpretable latent dynamics models learned solely from video.

cs.RO↗

arXiv2026

Learning Visually Interpretable Oscillator Networks for Soft Continuum Robots from Video

Learning soft continuum robot (SCR) dynamics from video offers flexibility but existing methods lack interpretability or rely on prior assumptions. Model-based approaches require prior knowledge and manual design. We bridge this gap by introducing: (1) The Attention Broadcast Decoder (ABCD), a plug-and-play module for autoencoder-based latent dynamics learning that generates pixel-accurate attention maps localizing each latent dimension's contribution while filtering static backgrounds, enabling visual interpretability via spatially grounded latents and on-image overlays. (2) Visual Oscillator Networks (VONs), a 2D latent oscillator network coupled to ABCD attention maps for on-image visualization of learned masses, coupling stiffness, and forces, thereby enabling mechanical interpretability. We validate our approach on single- and double-segment SCRs, demonstrating that ABCD-based models significantly improve multi-step prediction accuracy with 5.8x error reduction for Koopman operators and 3.5x for oscillator networks on a two-segment robot. VONs autonomously discover a chain structure of oscillators. This fully data-driven approach yields compact, mechanically interpretable models with potential relevance for future control applications.

cs.RO↗