Company News

Material Handling System (AMHS) is the "artery" of efficient operation in semiconductor fabs, and its operational efficiency directly determines the overall output rhythm. However, traditional AMHS scheduling modes are facing multiple challenges.

 

System barriers cause data fragmentation: data from MES (Manufacturing Execution System), MCS (Material Control System), and OHTC (Overhead Hoist Transport Control) cannot be efficiently integrated. Planning, allocation, and execution operate in silos, often contradicting the overall fab optimization goals.

Rigid scheduling struggles with complex conditions. Traditional algorithms based on static rules cannot flexibly adapt to dynamic constraints such as batch order deadlines, equipment resource contention, and track congestion, resulting in high OHT empty runs, long transport waiting times, and efficiency bottlenecks.

Operations and maintenance response relies heavily on experience. When anomalies occur, engineers need to manually query multiple systems for troubleshooting, taking hours to locate faults, missing the optimal recovery window.

The root cause of these problems is not insufficient equipment capability, but the lack of an integrated scheduling system that can integrate global business data, with intelligent judgment and autonomous decision-making capabilities. Targeting this industry dilemma, Wilson Optimus, a subsidiary of Gtrontec, precisely launched the AMHS Smart Logistics Solution, providing a full-process solution for semiconductor fab material handling challenges.

Breakthrough Method: AI+AMHS Smart Logistics Solution

Wilson Optimus AI+AMHS Smart Logistics Solution covers all material flow scenarios in semiconductor fabs horizontally, from warehouse storage, inter-line transport to machine loading/unloading. Vertically, it is supported by AI, multi-source data fusion, and Physical AI technology, creating an end-to-end closed-loop intelligent material handling decision system from "brain decision" to "limb execution".

The solution core consists of three layers, forming a complete closed loop of hardware and software coordination: Hardware Matrix: covering all material handling and storage scenarios, including OHT (Overhead Hoist Transport), Stocker (Automated Storage), Conveyor, and NTB (Near-Tool Buffer), building a stable and efficient hardware foundation. Intelligent Scheduling Hub: RTS/RTD (Smart Dispatching System), MCS (Material Control System), and OHTC (Overhead Hoist Transport Control) fully coordinated to achieve optimal matching and path planning for material handling tasks. Intelligent Agent Cluster: relying on multi-constraint reasoning agents for global optimal solution search, AI Agent scheduling operation agents for flexible adaptation to production demand changes, and Physical AI OHT autonomous control agents for millisecond-level local optimal execution decisions. The three work together to endow the system with core capabilities of autonomous judgment and flexible adaptation. This horizontal-to-edge, vertical-to-bottom hardware-software integrated solution successfully breaks the core bottlenecks of data silos and rigid scheduling in semiconductor fabs, truly achieving overall efficiency improvement in factory material handling. Deep Dive into Scenarios: In advanced wafer fabs where thousands of OHTs efficiently shuttle in a small area, OHTs need to execute tens of thousands of material handling tasks daily on a complex track network. Traditional scheduling models cannot quickly solve optimal solutions under production line fluctuations or multi-vehicle congestion, easily causing machine waiting, affecting overall OEE, and waste of capacity due to empty runs and idle waits. The advantages of Wilson Optimus's solution are concentrated in three core capabilities, precisely solving industry pain points: Global Optimization Engine: based on multi-constraint AI reasoning model, considering dozens of constraints such as batch deadlines, equipment resources, track flow, to search for optimal execution plans. Compared to traditional scheduling models, this engine integrates MES, MCS, OHTC multi-layer system data, anchors overall fab efficiency goals, and achieves four-dimensional collaborative optimization of process sequence, task assignment, path planning, and rhythm scheduling. Adaptive Operation Decision: AI Agent automatically adjusts operation parameters and task assignments based on real-time production demands and risk prediction, combined with large models and historical experience documents, to achieve system adaptive change and rescheduling, compressing emergency response from hours to minutes. Physical AI Local Autonomy: OHT is equipped with edge controllers, integrating multi-modal sensor information such as radar, cameras, gyroscopes, current and voltage. While receiving scheduling instructions, it makes local autonomous decisions based on environmental perception, including collision avoidance, path negotiation, and motion control adjustments, ensuring safe and smooth operation.

Practical Case: The Path to Production Increase and Efficiency Improvement for an 8-inch Wafer Fab

A top domestic 8-inch wafer manufacturer, whose production line heavily relies on manual material handling, urgently needed to build a high-tempo, fully automated logistics and production system to achieve a capacity ramp-up to 300,000 wafers per month with an 80% reduction in labor, i.e., the goal of "production increase and efficiency improvement".

Wilson Optimus integrated OHT+OHB+Stocker+Conveyor system layout for them, achieving path optimization and intelligent scheduling through RTD+MCS+OHTC system linkage, providing an automated material flow solution covering the entire production line.

Key Achievements

Yield and Capacity Dual Improvement: Eliminate vibration, contamination, and operational errors caused by manual handling, significantly improve production yield, and support capacity ramp-up goals. Full-stack Domestic Autonomous Control: Self-developed hardware and software ensure data security and real-time spare parts supply, providing controllable multi-party assurance. Unified Central Control Platform: Aggregate real-time hardware and software data from the entire fab, transform information into clear insights, enabling managers to achieve precise scheduling and efficient human-machine collaboration.

Next Episode Preview: The intelligent scheduling and operation system based on AI Agent will deeply deconstruct the core operation logic of the semiconductor logistics "brain", unlocking more efficient scheduling secrets. Stay tuned!