PCBRouteAgent: How Close Are LLM Agents to Becoming PCB Routing Engineers?
Paper · https://arxiv.org/abs/2611.08175
Automating printed circuit board (PCB) routing requires coordinating board-level connectivity decisions with geometric execution under strict design rules. Although recent agents can interact with native routing tools, reliable routing on complex boards remains challenging. We present PCBRouteAgent, a multimodal PCB routing agent inspired by how human engineers coordinate global strategy with local execution. A Global Routing Planner combines board images and structured electrical and geometric information to specify connection topology, routing order, corridors, and layer transitions. A Detailed Routing Executor realizes these guides through KiCad's native routing engine and performs bounded local recovery. Unresolved failures are returned as structured feedback, enabling targeted guide revisions while preserving accepted routes. We evaluate PCBRouteAgent on 247 synthetic and real-world boards from four PCBWorld-Bench test splits. It achieves clean-pass rates of 100%, 94.9%, 100%, and 40% on D2, D3-A, D3-B, and D3-C, respectively. On the three real-world subsets, PCBRouteAgent outperforms the best evaluated baseline in clean-pass rate by 13.9, 36, and 10 percentage points, respectively.
First seen by Oriel on .
AI summary
PCBRouteAgent is a multimodal PCB routing agent that combines global planning with native KiCad execution, substantially outperforming evaluated baselines on real-world boards.
- The Global Routing Planner uses board images plus structured electrical and geometric data to define topology, routing order, corridors, and layer transitions.
- The Detailed Routing Executor follows these guides through KiCad's native routing engine, performs bounded local recovery, and returns structured failure feedback for targeted revisions.
- Evaluation covered 247 synthetic and real-world boards across four PCBWorld-Bench test splits.
- Clean-pass rates were 100%, 94.9%, 100%, and 40% on D2, D3-A, D3-B, and D3-C; real-world gains were 13.9, 36, and 10 percentage points.
#llm agents #pcb routing #multimodal ai #kicad
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hugging Face Papers: 42, latest observation ; +8 over observed period since
Comments
Comment count reported by the source.
- Hugging Face Papers: 8, latest observation ; 0 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hugging Face Papers | 2611.08175 | https://arxiv.org/abs/2611.08175 Read discussion |