Graph Neural Network-Based Predictive Modeling for Robotic Plaster Printing

TL;DR

This work tackles predicting wall thickness in a spray-based robotic plastering process from the robot’s trajectory and printing parameters. It introduces a graph neural network with an encoder-processor-decoder architecture that represents the wall as a graph of wall particles plus an end-effector node, with connectivity capturing end-effector influence and inter-particle relations. The model is trained on augmented time-series data and optimized via Bayesian methods, achieving higher accuracy than a benchmark across multiple metrics and displaying linear error growth with prediction horizon. This approach enables accurate, scalable predictive simulation for trajectory planning and parameter optimization in autonomous plastering workflows.

Abstract

This work proposes a Graph Neural Network (GNN) modeling approach to predict the resulting surface from a particle based fabrication process. The latter consists of spray-based printing of cementitious plaster on a wall and is facilitated with the use of a robotic arm. The predictions are computed using the robotic arm trajectory features, such as position, velocity and direction, as well as the printing process parameters. The proposed approach, based on a particle representation of the wall domain and the end effector, allows for the adoption of a graph-based solution. The GNN model consists of an encoder-processor-decoder architecture and is trained using data from laboratory tests, while the hyperparameters are optimized by means of a Bayesian scheme. The aim of this model is to act as a simulator of the printing process, and ultimately used for the generation of the robotic arm trajectory and the optimization of the printing parameters, towards the materialization of an autonomous plastering process. The performance of the proposed model is assessed in terms of the prediction error against unseen ground truth data, which shows its generality in varied scenarios, as well as in comparison with the performance of an existing benchmark model. The results demonstrate a significant improvement over the benchmark model, with notably better performance and enhanced error scaling across prediction steps.

Figures (16)

• Figure 1: Robotic Plaster Spraying (RPS), introduces an additive-only, spray-based printing technique that can be applied directly onto a building structure.
• Figure 2: Examples of standardized and bespoke (custom) surfaces created with the proposed plaster printing process.
• Figure 3: End-effector tools used by robotic_plastering; left: troweling tool used in conventional plastering processes to smoothen the surfaces; right: camera-based inspection of the quality.
• Figure 4: Graphical representation of the plastering process; the wall thickness at time step $t_k$ is represented by the point cloud $\mathbf{p}^{t_k}$; the operational point of the spraying gun is described by the working pressure $P^{t_k}$, the position $\mathbf{tp}^{t_k}$, the direction of spray-based printing $\mathbf{n}^{t_k}$ and the velocity $\mathbf{u}^{t_k}$.
• Figure 5: Illustration of a plaster pattern along with the spray-based printing trajectories, whose positions $\mathbf{tp}^{t_k}$ at each step are indicated by red points; the small vectors on each trajectory point represent the printing direction.