pith. sign in

arxiv: 1906.11753 · v1 · pith:VOYK4MIFnew · submitted 2019-06-27 · 💻 cs.HC

Dynamic Drawing Guidance via Electromagnetic Haptic Feedback

Thomas Langerak , Juan Zarate , Velko Vechev , Daniele Panozzo , Otmar Hilliges This is my paper

Pith reviewed 2026-05-25 14:38 UTC · model grok-4.3

classification 💻 cs.HC
keywords electromagnetic feedbackhaptic guidancedrawing assistancereceding horizon controloptimal controluser studytablet interaction
0
0 comments X

The pith

An electromagnet moving under a tablet delivers real-time haptic guidance that keeps drawing errors to 2.8 mm while letting users retain full control of pace and style.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a guidance system for drawing and sketching that uses an electromagnet positioned beneath a high-refresh-rate pressure-sensitive tablet. The system continuously tracks the pen and optimizes the magnet's position and power in a receding-horizon control loop to gently steer the pen toward a desired path. Because the control is time-free and error-correcting rather than setpoint-pushing, users keep their natural drawing rhythm. Experimental measurements show the pen stays within 2.8 mm dispersion of the target, and a user study finds that accuracy gains grow larger as the drawn shapes become more complex.

Core claim

The central discovery is that a novel approximate model of the electromagnet, combined with a receding-horizon optimal-control formulation, enables closed-loop haptic feedback that corrects drawing trajectories in real time. This approach produces a measured dispersion of 2.8 mm (+/-0.8 mm) and demonstrably raises user accuracy on shapes of varying complexity.

What carries the argument

The receding-horizon optimal control loop that repeatedly measures pen position and solves for the next magnet state using the fast approximate electromagnet model.

If this is right

  • The guidance remains effective even when the user draws at an arbitrary speed or pauses.
  • Accuracy improvement scales with task difficulty, offering more help on intricate curves.
  • The closed-loop nature produces corrective pulls back to the path rather than forcing the pen forward.
  • Real-time performance is sustained because the approximate model runs fast enough for iterative optimization.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar electromagnetic setups could be adapted to guide other manual tasks such as handwriting practice or surgical tool positioning.
  • Combining the haptic signal with visual overlays on the tablet might further amplify learning effects for novices.
  • The low dispersion suggests the method could support fine motor rehabilitation if the target trajectories are adjusted dynamically.

Load-bearing premise

The approximate electromagnet model must remain accurate enough during live operation to keep the closed-loop controller stable and effective.

What would settle it

Running the system on a new electromagnet or tablet where the model's force predictions deviate by more than a few percent from measured forces, causing the pen to drift beyond the reported 2.8 mm dispersion.

Figures

Figures reproduced from arXiv: 1906.11753 by Daniele Panozzo, Juan Zarate, Otmar Hilliges, Thomas Langerak, Velko Vechev.

Figure 1
Figure 1. Figure 1: Left-to-right: an electromagnet moving on a bi-axial linear stage underneath a high-speed pressure sensitive tablet delivers dynamically adjustable [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Hardware overview. can significantly increase the perceived force), ii) the fast pen mo￾tion compared to the speed of the linear stage, and iii) the hard to predict behavior of the user. These challenges enforce a very tight computational budget from pen motion to magnet actuation, that is, we need an efficient numerical solution to design a system with a low overall latency. To this end we propose an appr… view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of the model to compute the force [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 6
Figure 6. Figure 6: Overview of different control strategies and their theoretical behav [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 5
Figure 5. Figure 5: Illustration of actuation force Fa, desired force Fθ , and the force cost-term Cf associate with the difference between those two forces. separation: Fa = α F0 © ­ ­ ­ « d  4 − d 2 h 2  h  1 + d 2 h 2  7 2 ª ® ® ® ¬ ed , (10) where F0 is a constant force parameter given by the expression, F0 = 3 µ0 mp mm 4 π h 4 . (11) [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Illustration of lag- and contouring error decomposition. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Illustration of error correcting behavior. Left-to-right: (a) a simulated user is close to the desired path. (b) A sudden jump in pen-position causes the [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Overview of the different metrics for the preliminary user evaluation. [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Comparison of error over time for a single participant (P1). In 10a there is the inverse u-shape that illustrates that the [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Shapes used in our user tests. Note that the drawing surface only [PITH_FULL_IMAGE:figures/full_fig_p011_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Selected experimental results. Each shape drawn by different partic [PITH_FULL_IMAGE:figures/full_fig_p011_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Accuracy comparison for a single participant. a reference shape (dotted line) overlaid by path drawn by the same user with (orange) and without [PITH_FULL_IMAGE:figures/full_fig_p012_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Distribution of pen to reference distances. All subjects combined. [PITH_FULL_IMAGE:figures/full_fig_p012_14.png] view at source ↗
Figure 16
Figure 16. Figure 16: Curvature dependent error. Insets a-c show increasingly sharp [PITH_FULL_IMAGE:figures/full_fig_p013_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Different variants of the same dragon, drawn with identical system [PITH_FULL_IMAGE:figures/full_fig_p014_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: 3D overview of all data points. The x,y axis are position and the [PITH_FULL_IMAGE:figures/full_fig_p016_18.png] view at source ↗
read the original abstract

We propose a system to deliver dynamic guidance in drawing, sketching and handwriting tasks via an electromagnet moving underneath a high refresh rate pressure sensitive tablet. The system allows the user to move the pen at their own pace and style and does not take away control. The system continously and iteratively measures the pen motion and adjusts magnet position and power according to the user input in real-time via a receding horizon optimal control formulation. The optimization is based on a novel approximate electromagnet model that is fast enough for use in real-time methods, yet provides very good fit to experimental data. Using a closed-loop time-free approach allows for error-correcting behavior, gently pulling the user back to the desired trajectory rather than pushing or pulling the pen to a continuously advancing setpoint. Our experimental results show that the system can control the pen position with a very low dispersion of 2.8mm (+/-0.8mm). An initial user study indicates that it significantly increases accuracy of users drawing a variety of shapes and that this improvement increases with complexity of the shape.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper presents an electromagnetic haptic guidance system for drawing, sketching, and handwriting tasks. An electromagnet beneath a high-refresh-rate pressure-sensitive tablet is positioned and powered in real time via a receding-horizon optimal controller that uses a novel fast approximate electromagnet model. The approach is time-free and closed-loop, allowing users to proceed at their own pace while the system gently corrects trajectory errors. Experiments report a pen-position dispersion of 2.8 mm (±0.8 mm) under closed-loop control; an initial user study finds statistically significant accuracy gains that increase with shape complexity.

Significance. If the empirical claims hold, the work offers a practical contribution to HCI by demonstrating a non-intrusive, user-controlled haptic assistance method whose performance scales with task difficulty. The direct experimental measurement of dispersion and the user-study accuracy metrics provide independent support for the central performance claims, independent of any modeling circularity.

major comments (2)
  1. [Abstract / model description] The abstract states that the approximate electromagnet model 'provides very good fit to experimental data' and is 'fast enough for use in real-time methods,' yet no quantitative fit metrics (e.g., RMSE, R²) or timing benchmarks appear in the provided summary; these numbers are load-bearing for the claim that the receding-horizon controller can run closed-loop at tablet rates.
  2. [User study results] The user-study claim that 'improvement increases with complexity of the shape' requires the specific statistical test, effect sizes, and per-shape accuracy tables to be examined; without them it is unclear whether the interaction effect is robust or driven by a small number of complex shapes.
minor comments (2)
  1. [Control formulation] Clarify the exact formulation of the time-free cost function in the receding-horizon controller (e.g., how the reference trajectory is encoded without an explicit time index).
  2. [Experimental results] The reported dispersion of 2.8 mm (±0.8 mm) should be accompanied by the number of trials, the exact definition of 'dispersion' (e.g., RMS error or maximum deviation), and the baseline open-loop condition for comparison.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive review and the recommendation of minor revision. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract / model description] The abstract states that the approximate electromagnet model 'provides very good fit to experimental data' and is 'fast enough for use in real-time methods,' yet no quantitative fit metrics (e.g., RMSE, R²) or timing benchmarks appear in the provided summary; these numbers are load-bearing for the claim that the receding-horizon controller can run closed-loop at tablet rates.

    Authors: The model validation, including quantitative fit metrics and timing benchmarks, is reported in the body of the manuscript (Section 4). To ensure the abstract is self-contained and the real-time claim is immediately verifiable, we will revise the abstract to include a concise reference to these metrics. revision: yes

  2. Referee: [User study results] The user-study claim that 'improvement increases with complexity of the shape' requires the specific statistical test, effect sizes, and per-shape accuracy tables to be examined; without them it is unclear whether the interaction effect is robust or driven by a small number of complex shapes.

    Authors: The manuscript presents the user-study analysis, including the interaction test and per-shape data, in Section 5. We will revise to make the statistical details, effect sizes, and tables more explicitly referenced in the results summary so that the robustness of the complexity interaction is clear without requiring the reader to locate the full section. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper's central claims (2.8 mm pen dispersion under closed-loop control and accuracy gains in the user study) are established via direct experimental measurement of physical outcomes and participant performance, rather than any quantity defined circularly in terms of the electromagnet model's fitted parameters or receding-horizon optimization. The approximate model is described as enabling real-time control, but the reported dispersion and study results provide independent empirical support outside the model's internal equations. No load-bearing derivation step reduces by construction to a self-definition, fitted input renamed as prediction, or self-citation chain.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no explicit free parameters, axioms, or invented entities; the approximate electromagnet model is described as novel and fitted but details are unavailable.

pith-pipeline@v0.9.0 · 5717 in / 1000 out tokens · 23027 ms · 2026-05-25T14:38:45.807443+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

2 extracted references · 2 canonical work pages

  1. [1]

    In Proceedings of the 17th annual ACM symposium on User interface software and technology

    Haptic pen: a tactile feedback stylus for touch screens. In Proceedings of the 17th annual ACM symposium on User interface software and technology . ACM, 291–294. Yong Jae Lee, C Lawrence Zitnick, and Michael F Cohen. 2011. Shadowdraw: real-time user guidance for freehand drawing. InACM Transactions on Graphics (TOG), Vol. 30. ACM, 27. Alex Limpaecher, Ni...

    work page doi:10.1002/oca.2123 2011

  2. [2]

    Setting the electromagnet toα= 1 and moving it in a grid we attain multiple readings of the hall sensor for different electromagnet positions pm

    We use a hall sensor (Allegro A1324, sensitivity is 5 mV/G)2 to measure the z-magnetic flux at a fix heighthm, where the magnet of the pen would be. Setting the electromagnet toα= 1 and moving it in a grid we attain multiple readings of the hall sensor for different electromagnet positions pm. We present the obtained magnetic field plotted in Figure 18, t...

    work page arXiv 2019