Pcbrouteagent automates complex printed circuit board routing tasks
PCBRouteAgent: How Close Are LLM Agents to Becoming PCB Routing Engineers?
Artificial Intelligence
Summary
Printed circuit boards (PCBs) need carefully planned wiring to work properly, but routing these connections is hard and requires following strict rules. The authors created PCBRouteAgent, a system that mimics how human engineers plan and execute PCB wiring by combining overall strategy with detailed adjustments. It uses images and electrical info to plan routing paths and then works with routing software to draw wires, fixing errors as it goes. In tests on hundreds of PCB designs, PCBRouteAgent routed boards more accurately than previous methods.
What this means in practice
- •For pcb designers: Automate complex routing decisions on PCB layouts while respecting design rules, improving board design efficiency.
- •For electronics manufacturing teams: Use improved routing plans to reduce manual correction and speed up PCB production workflows.
Authors
Yue Zhang, Jilei Sun, Liqiang Jing, Yijie Wei, Ruohan Yang
Abstract
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.