AIMS-MOS Industrial IoT Platform: Building an Intelligent Operating Foundation for the Manufacturing Floor

On one factory floor, the four-way shuttles come from one supplier and the AGVs from another, while APS, MES and WMS are each built by a different team. Every piece of equipment runs automatically, yet the plant as a whole does not necessarily run in sync.

The AIMStek AIMS-MOS Industrial IoT Platform sets out to give the manufacturing floor a unified, intelligent operating foundation. Equipment connects to it, data converges on it, business applications run on it, and AI continuously optimizes it.

Guided by the philosophy of “Software-Defined Logistics, Data-Driven Manufacturing,” this foundation uses a six-layer architecture spanning device connectivity, task scheduling, business applications and AI optimization. It serves three major scenarios: smart intralogistics, manufacturing operations, and process and energy optimization.

1. Three Shop-Floor Challenges Call for a Unified Foundation

Automated equipment does not guarantee an efficient production system. Machines must work in coordination, plans must align with what happens on the floor, material delivery must keep pace with production takt, and process parameters and energy use must adapt to operating conditions. Manufacturers pushing efficiency further typically have to clear three hurdles. Adding more equipment or deploying another standalone system cannot clear them. They call for a unified foundation, and AIMS-MOS addresses each hurdle with one core capability.

The platform’s intelligence revolves around perception, decision-making and coordination. Perception captures equipment and business operating data. Decision-making analyzes and optimizes against business objectives. Coordination drives task and business execution. Together they form a closed loop from shop-floor sensing to business execution.

2. Six-Layer Architecture: From Device Connectivity to Intelligent Applications

AIMS-MOS adopts a six-layer architecture. From the bottom up, the layers are the Device Layer, MOS Kernel Layer, MOS Engine Layer, MOS Development Toolkit Layer, APPs Layer and Insight AI Layer. The lower layers connect to the shop floor, the middle layers provide runtime and development capabilities, and the upper layers host business logic and intelligence. Together they form an intelligent operating foundation built for the manufacturing floor.

2.1 Device Layer: Cross-Brand, Cross-Protocol Unified Scheduling

The Device Layer serves a wide range of industrial equipment: material handling and conveying systems, special-purpose machining equipment, production and manufacturing equipment, data acquisition devices, environmental sensing systems and on-site HMI devices. It establishes an open, standardized and extensible framework for device connectivity and capability abstraction. This removes the technical barriers between brands, manufacturers and communication protocols, and lays the groundwork for unified connectivity, management and scheduling of heterogeneous equipment.

The platform covers logistics equipment, special-purpose machining equipment, data acquisition and inspection devices, environmental sensing systems, production and processing equipment, and on-site interaction devices. It supports conversion across more than 20 industrial protocols, and integration with equipment from mainstream logistics brands is already in place. Interfaces, commands and status information from different manufacturers, brands and protocols are mapped into standardized device capabilities (data acquisition, special-purpose machining, transport, output, pick-up, put-down and so on), forming a unified capability model decoupled from the underlying implementation. As a result, scheduling logic is decoupled from equipment brand, manufacturer, protocol and driver. The upper layers orchestrate tasks, allocate resources and coordinate control against standardized capabilities, shifting the approach from customized control of individual devices to generalized scheduling based on device capabilities.

Thanks to a pluggable driver extension mechanism, integrating a new device only requires extending the corresponding driver. The existing scheduling framework and business workflows can then be reused as they are, which lowers the cost of integration, expansion and operations and maintenance. The platform further links equipment status, job tasks and on-site resources, providing support for IoT data acquisition, the uRMS scheduling system and upper-layer business coordination, and supplying the data foundation for process applications such as intelligent feeding optimization.

2.2 Kernel Layer and Engine Layer: One Manages “What You Have,” the Other “How to Use It”

The MOS Kernel Layer comprises task management, resource management, data management, data ELT and device drivers. It organizes equipment, tasks, resources and data within a single framework, rather than having each upper-layer application maintain its own disconnected set of master data.

The MOS Engine Layer comprises the intelligence engine, data engine, business engine and event engine. It lets scheduling rules and business logic run on the foundation.

Consider an urgent task entering the system. The Kernel Layer records the task and checks the status of available equipment and resources. The Engine Layer then determines how to handle it based on business rules and scheduling strategies, and combines this with optimization capabilities to produce an execution plan. The Kernel Layer manages what you have, and the Engine Layer manages how to use it. Common operating logic is built into the Engine Layer once, so it does not have to be rebuilt in every project.

2.3 MOS Development Toolkit Layer: Adapt to Change Through Configuration, Not Code Rewrites

Manufacturing projects vary widely in their scenarios. Even for the same logistics scheduling task, each plant has its own equipment mix, task rules and operating procedures. Within a single plant, workflows must also change when capacity or processes are adjusted.

The Development Toolkit Layer provides two types of tools:

• Visual development tools for configuring business workflows and building applications.

• Debugging and simulation tools for validating scheduling logic and operating plans.

Workflows are configured in a graphical interface and run through in a simulation environment before going live. Projects are adapted from the platform’s existing capabilities rather than custom-developed requirement by requirement, which means faster delivery and lighter-weight changes afterward.

2.4 APPs Layer: From Scheduling to Warehousing, One Platform for the Full Business

The APPs Layer turns platform capabilities into concrete business processes:

APS: production planning and scheduling optimization

MES: shop-floor manufacturing execution management

WMS: warehouse operations and inventory management

LES: production logistics execution and material delivery

uRMS: unified management and scheduling of multiple types of robots and related resources

IOCS Hot Blast Stove and IOCS Polysilicon Reduction Furnace: process optimization and control applications for process industries

The platform supplies common capabilities for devices, data, engines and development, while each application carries its own business logic. The platform also exposes open APIs, SDKs and low-code development tools, so customers, partners and third-party developers can build, integrate and extend applications on it.

2.5 Insight AI Layer: Models That Keep Optimizing from Operating Data

The Insight AI Layer covers model training, optimization and deployment. Its data comes from the lower layers: equipment, platform operations and business applications.

Using accumulated operating data, the platform can analyze and optimize models against specific business objectives, then apply validated models or optimization results to the relevant operations. New operating data in turn provides the basis for further model evaluation and refinement, giving the platform a foundation for continuous optimization. Applications such as the IOCS Hot Blast Stove and Polysilicon Reduction Furnace systems show where combining AI with process mechanism knowledge can lead.

2.6 How the Foundation Operates as a Whole

The six layers form one operating mechanism through two interconnected pathways.

Data converges from the bottom up. Data generated by equipment, platform operations and business applications is managed and processed, then serves as the basis for business analysis and for model training and optimization. This connects data across business processes.

Decisions are executed from the top down. Decisions come from business rules, scheduling algorithms and validated AI models. Business applications process them in combination with business rules, and the platform’s engines, task management and resource management turn them into execution plans, which the relevant equipment and control systems then carry out.

The new data generated by execution flows back into the platform, completing a closed loop of “perception, decision-making, coordination and feedback.”

3. An Open Foundation: Why AIMS-MOS

No lock-in to a single equipment brand. The platform is compatible with equipment from multiple brands. Enterprises can select equipment based on site conditions, existing assets and investment plans, then bring it into a unified business framework through the platform’s adaptation and management capabilities.

Existing systems can be retained and integrated. The platform does not assume that all business processes must run on one set of applications. Enterprises can choose the deployment approach that suits their situation:

• New plants: deploy the platform together with the relevant applications as business needs require.

• Enterprises with existing group-level systems: retain those systems and integrate through interfaces.

• Enterprises with complex software landscapes: define the collaboration between the platform and specialized business systems along business boundaries.

Less redundant development, more room for business change. Standardized modules, platform engines and development tools make common capabilities reusable. Visual configuration and adjustable scheduling strategies leave room for changes in production organization, logistics processes and business rules.

Room for intelligent applications. A unified data and business foundation, together with Insight AI’s training, optimization and deployment capabilities, means enterprises can later introduce data analytics and AI applications without building a separate system.

4. Three Business Solutions Built on the Foundation

Smart intralogistics, manufacturing operations, and process and energy optimization are three solution categories built on the same foundation. They share platform capabilities such as device connectivity, data and engines.

4.1 Intralogistics

With AIMS-MOS as the platform, this solution serves warehouses, workshops and industrial parks. It combines automated storage and retrieval systems (AS/RS), AGVs, four-way shuttles, loading and unloading systems and other equipment with applications such as WMS, uRMS and LES. Warehouse operations, material handling and production delivery are orchestrated as one, so logistics keeps pace with production takt.

4.2 Manufacturing Operations

Centered on production planning, scheduling, manufacturing execution and materials management, this solution combines applications such as APS, MES and WMS. It keeps plans, shop-floor execution and material supply in step, and allows timely adjustment when orders change or production fluctuates. The application mix and division of responsibilities are determined by each enterprise’s existing systems and business needs.

4.3 Process and Energy Optimization

Serving steel, thermal power, non-ferrous metals and other process industries, this solution combines AI with process mechanism knowledge to build applications for process parameter optimization and energy efficiency. Current products include the IOCS Hot Blast Stove Intelligent Optimization and Control System and the IOCS Polysilicon Reduction Furnace Intelligent Optimization and Control System.

Conclusion

Digitalization on the manufacturing floor cannot stop at equipment automation or the rollout of a single system. AIMS-MOS uses a six-layer architecture to build an intelligent operating foundation for the shop floor. Equipment connects to it, data converges on it, business applications run on it, and AI continuously optimizes it. The foundation is open to multi-brand equipment and third-party applications, and it supports three major scenarios: smart intralogistics, manufacturing operations, and process and energy optimization. It leaves manufacturers free to choose their equipment, their software mix and their own long-term path.

Guided by “Software-Defined Logistics, Data-Driven Manufacturing,” AIMStek will continue to respond to real shop-floor needs. It will advance platform capabilities, intelligent equipment and industry solutions in step with one another, and keep strengthening the intelligent operating foundation for the manufacturing floor.

Disclaimer: Actual delivery results depend on equipment conditions, business complexity, data readiness and project scope. Integrating equipment and applications requires meeting the relevant interface, protocol, data, permission and runtime environment requirements. Actual process optimization results are subject to project operating outcomes.