Optimizing Manufacturing Assembly with Pick-to-Light
Our client is a leading automotive manufacturer in Southeast Asia, renowned for its state-of-the-art production facilities and rapid global expansion. Operating one of the most advanced and large-scale assembly plants in the region, they required a cutting-edge solution to ensure flawless accuracy and speed in their complex parts-picking process.
1. Requirement
1.1. Purpose
The project's primary objective was to design and implement a central software control system to fully automate and optimize the parts-picking process for the client's high-volume assembly line. This software was engineered to integrate seamlessly with the factory's Manufacturing Execution System (MES), automatically converting production orders into intuitive visual signals on the Pick-to-Light system. The ultimate goal was to replace the manual process, thereby eliminating errors, dramatically increasing picking speed and productivity, and ensuring an accurate, efficient, and timely material flow while providing management with real-time operational data.
1.2. Detail Requirement
- MES Integration and Device Control
The system must feature a robust API for seamless, two-way communication with the factory's MES to receive and parse production orders. It is required to map part data to physical shelf locations and transmit real-time control signals to activate specific Pick-to-Light indicators, display quantities, and register worker confirmations. - Real-Time Order Management
A centralized, real-time dashboard is required to provide supervisors with a complete overview of all orders. This interface must display live order statuses (e.g., pending, in-progress, complete), track individual picking progress, and automatically update as tasks are completed. The system must also capture key performance data, such as cycle times. - Error Handling and Custom Business Logic
The software must include robust error-handling workflows for exceptions like stockouts or short picks, with automated alerts for supervisors. It must enforce custom business logic, such as optimized picking sequences to minimize worker travel time. The system also needs to support different user roles (worker, supervisor) and allow for basic operator functions, like canceling the last action.
2. Customer Problems
- Outdated Manual Process: The client relied entirely on a paper-based picking process unsuitable for a modern, high-volume production environment.
- Low Productivity & Wasted Time: Significant time was lost as workers manually searched for parts and visually verified picks, drastically reducing labor productivity.
- High Error Rate: The manual method led to a high risk of picking errors (wrong parts or quantities), which compromised product quality and wasted resources.
- Production Bottlenecks: Cumulatively, these issues created a critical bottleneck in the material flow, causing delays and jeopardizing the factory's ability to meet its strict production schedule.
1. Programming Language
- Typescript
- Node.js
2. Framework
- NestJS
- ReactJS
3. Third Party
Microsoft Azure Authentication
4. Database
- MySQL (version 8)
- MySQL Router
1. Challenge
- Complex Business Logic: The system needed to support the factory's complex and strict workflows, including precise picking sequences for the assembly line, which demanded a highly flexible and customizable software architecture.
- Minimizing Human Error: With each order containing numerous items, reducing the high risk of operator error was critical. The solution had to significantly lower worker cognitive load by providing intuitive, visual guidance for every action.
- Real-Time Data Synchronization: Maintaining ultra-low-latency, two-way communication between the central server and a large number of devices was a critical technical challenge, as instantaneous updates were essential for immediate error detection in the fast-paced environment.
2. How to resolve these challenges
To comprehensively address the client's problems and overcome the stated challenges, we developed a multi-layered, custom-designed, and intelligent software solution focused on three core pillars:
- Custom Hardware SDK: We built a custom Software Development Kit (SDK) to provide direct, low-latency control over all Pick-to-Light hardware. This ensured instantaneous and reliable communication between the software and hundreds of devices, solving the real-time synchronization challenge.
- Robust & Validated Business Logic: We developed robust business logic to automate the factory's complex workflows and optimized picking sequences. This logic was rigorously validated through pre-deployment simulations on a physical model to guarantee accuracy and flawless operation.
- Proactive Monitoring & Error Handling: An intelligent monitoring system was established, featuring a real-time dashboard and automated alerts. This empowered supervisors to proactively track progress and resolve operational errors, drastically reducing the impact of human error on the assembly line.
Other Works
Bioequivalence AI Platform — From Clinical PDFs to Regulatory-Grade Assessment
An enterprise-grade bioequivalence AI platform delivering FDA-aligned assessments from unstructured clinical data - production-ready, fully client-owned.
From Prototype to Enterprise-Ready: A macOS Security Agent That Actually Ships
From proof-of-concept to enterprise distribution - a macOS compliance agent hardened, signed, and production-ready.
From Reactive to Predictive - Revolutionizing US B2B Services
Revolutionizing US B2B Services from Reactive to Predictive using AI/ML and Custom ERP; achieving 15% cost savings and 30% process optimization.