Computational Design and Single-Wire Sensing of 3D Printed Objects with Integrated Capacitive Touchpoints

TL;DR

This work tackles the labor-intensive process of making interactive 3D printed objects by introducing a computational design pipeline that embeds multiple capacitive touchpoints inside any closed, non-self-intersecting 3D mesh using multi-material printing. By engineering internal circuitry whose traces produce distinct RC delays, the system enables sensing of all touchpoints with either a single wire or two wires, dramatically reducing instrumentation. The approach combines a graph-based pathfinding routing, a space-filling serpentine trace for resistance, and a resistance-optimization strategy to maximize delay differences, yielding mean recognition accuracies around $93.35\%$ (single-wire) and $89.49\%$ (double-wire) across diverse geometries. Applications span from MIDI drumpads to cultural artifacts, demonstrating scalable, robust, and fabricable interactive objects ready to use straight off the printer, with future work targeting smaller objects, more signals, real-time calibration, and multi-touch support.

Abstract

Producing interactive 3D printed objects currently requires laborious 3D design and post-instrumentation with off-the-shelf electronics. Multi-material 3D printing using conductive PLA presents opportunities to mitigate these challenges. We present a computational design pipeline that embeds multiple capacitive touchpoints into any 3D model that has a closed mesh without self-intersection. With our pipeline, users define touchpoints on the 3D object's surface to indicate interactive regions. Our pipeline then automatically generates a conductive path to connect the touch regions. This path is optimized to output unique resistor-capacitor delays when each region is touched, resulting in all regions being able to be sensed through a double-wire or single-wire connection. We illustrate our approach's utility with five computational and sensing performance evaluations (achieving 93.35% mean accuracy for single-wire) and six application examples. Our sensing technique supports existing uses (e.g., prototyping) and highlights the growing promise to produce interactive devices entirely with 3D printing. Project website: https://github.com/d-rep-lab/3dp-singlewire-sensing

Figures (13)

• Figure 1: Capacitive sensing for the Stanford Bunny: (a) overall schematic on how the five touchpoints are connected via conductive traces (colored blue) and connected in series to a microcontroller's circuit using two wires; (b) the circuit diagram corresponding to (a) with the representative resistance and capacitance measurements. Each colored dashed wire corresponds to a case when a point is touched (e.g., orange: tail, green: foot); and (c) the voltage change measured at the microcontroller's receive pin when a point is touched. (d--f) correspond to (a--c) but the Stanford Bunny is connected in parallel to a microcontroller's circuit using one wire.
• Figure 2: A computational design pipeline to create a freeform interface with embedded capacitive touchpoints. Rounded rectangles represent output data or physical objects. Arrows represent different processes. Each image at the bottom is the corresponding output data (i.e., rounded rectangles) from the computational pipeline.
• Figure 3: 3D model of the Stanford Bunny with five selected touchpoints (foot, nose, left ear, right ear, and tail) and two wiring connection points. The right ear and the two wiring points are hidden from the viewpoint. The orange meshes indicate the user's lassoed selections. The dashed lines in (a) show where users can download the selected mesh coordinates.
• Figure 4: Automatic circuit design. (a) Our automatic circuit design first generates a circuit template, which outlines how we will embed the touchpoints and conductive traces inside a freeform interface. (b) During circuit embedding, serpentine trace patterns are generated inside low conductivity conduits by using a space-filling algorithm. (c) The output is fabrication data (STL files), which will be used for multi-material printing.
• Figure 5: Representations involved in pathfinding: (a) input 3D model for pathfinding; (b) voxel representation of the input model; (c) voxel representation after trimming voxels close to the model's surface; (d) graph representation of the trimmed voxel representation, where a vertex closest to each touchpoint is colored red; (e) a close-up look of the graph representation; and (f) identified paths using Dijkstra's pathfinding algorithm. Except for (e), all figures share the same camera position and angle. For presentation purposes, the voxel and graph representations have lower resolution (i.e., smaller numbers of voxels and vertices) than implemented.