From CAD to POMDP: Probabilistic Planning for Robotic Disassembly of End-Of-Life Products
Jan Baumgärtner, Malte Hansjosten, David Hald, Adrian Hauptmannl, Alexander Puchta, Jürgen Fleischer
AI summary
Problem
Existing robotic disassembly planners assume deterministic, fully observable product models, failing to handle the uncertainty caused by wear, corrosion, and undocumented repairs in end-of-life products.
Approach
The authors formulate robotic disassembly as a Partially Observable Markov Decision Process (POMDP) where hidden variables represent uncertain structural conditions, then automatically generate instance-specific POMDP models from CAD data, robot capabilities, and inspection priors, solved via reinforcement learning and Bayesian filtering.
Key results
- Automatically derives POMDP models from CAD and inspection data
- Outperforms deterministic baselines in average disassembly time and variance
- Generalizes across different robotic setups and product types
- Successfully adapts to CAD deviations like missing or stuck parts
Why it matters
Enables scalable, robust robotic disassembly for circular economy initiatives by handling real-world product degradation without manual reprogramming.
Abstract
To support the circular economy, robotic systems must not only assemble new products but also disassemble end- of-life (EOL) ones for reuse, recycling, or safe disposal. Existing approaches to disassembly sequence planning often assume deterministic and fully observable product models, yet real EOL products frequently deviate from their initial designs due to wear, corrosion, or undocumented repairs. We argue that disas- sembly should therefore be formulated as a Partially Observable Markov Decision Process (POMDP), which naturally captures uncertainty about the product’s internal state. We present a mathematical formulation of disassembly as a POMDP, in which hidden variables represent uncertain structural or physical properties. Building on this formulation, we propose a task and motion planning framework that automatically derives specific POMDP models from CAD data, robot capabilities, and inspection results. To obtain tractable policies, we approximate this formulation with a reinforcement-learning approach that operates on stochastic action outcomes informed by inspection priors, while a Bayesian filter continuously maintains beliefs over latent EOL conditions during execution. Using three products on two robotic systems, we demonstrate that this probabilistic planning framework outperforms deterministic baselines in terms of average disassembly time and variance, generalizes across different robot setups, and successfully adapts to deviations from the CAD model, such as missing or stuck parts.