Optimized and Routed Wiring Harness Based on Zonal Clustering Concept Using AI in the Automotive Industry

This paper presents an AI-driven approach for multi-zonal clustering and harness routing optimization in automotive electrical/electronic (E/E) systems. A methodology integrating K-means clustering with dynamic grid-based routing algorithms (A* and Bresenham’s line algorithm) is applied to optimize the wiring harness layout across the vehicle’s zones, covering six key domains (comfort, chassis, drive assist, infotainment, system, and powertrain). The proposed method accounts for realistic vehicle constraints, including restricted zones and diverse wire types, which are often overlooked in prior work. A 12-Zone architecture under a high application-load scenario is examined, demonstrating that this configuration achieves an optimal trade-off between harness length, complexity, and cost efficiency. Compared to manual design approaches and existing zonal architectures in the literature, the AI-based method reduced design time, produced a leaner harness layout, and achieved measurable material and cost savings. Specifically, the 12-Zone cluster arrangement achieved a total wiring distance of 241.056 m and a total wire mass of 22.92 kg for 417 ECUs, outperforming both lower and higher zone counts in overall efficiency. Scalability tests from 1 to 14 zones (under Low, Mid, and High load variants) confirmed the robustness of the approach and identified the 12-Zone configuration as the best-balanced solution. These findings highlight the potential of AI-optimized zonal E/E architectures to significantly reduce wiring weight, complexity, and design effort while maintaining system performance and cost competitiveness, providing valuable insights for next-generation automotive E/E design.