Duration: 2013 – 2016
ID: 2013111901
Status: Completed
Domain: Industrial Robotics, Advanced Manufacturing, Software Engineering
An advanced software system for intelligent management of palletization patterns and hardware configurations in flexible robotic palletizing cells. The system provides automated pattern generation using proprietary heuristic algorithms, 3D visualization, multi-language support, and automatic robot program generation, serving both sales representatives and cell operators with an intuitive tablet-optimized interface.
Project Overview
Efficient palletization represents a critical challenge in modern manufacturing and logistics operations. The optimal arrangement of boxes on pallets directly impacts storage space utilization, transportation costs, handling efficiency, and product safety during shipping. However, finding optimal packing solutions is computationally complex (an NP-complete problem), and traditional approaches often rely on manual planning or rigid pre-defined patterns that cannot adapt to varying product dimensions and operational requirements.
The Expert System for Smart Robots project began in 2013 as a research collaboration between our research group and CSi Industries B.V. Netherlands. Initially conceived to develop an algorithm for optimal box arrangement on pallets, the project scope expanded significantly through 2016, evolving into a comprehensive software solution that became the “companion” system for CSi’s robotic palletizing cells. The system addresses the needs of two distinct user groups: sales representatives requiring technical configuration capabilities, and cell operators needing straightforward operational control with minimal technical knowledge.
The resulting application provides intelligent pattern generation using proprietary heuristic algorithms, real-time 3D visualization of pallets and robotic cells, intuitive touch-optimized interfaces for tablet operation, automatic robot program code generation, and comprehensive hardware configuration management. This dynamically reconfigurable solution delivers the exact optimal palletization pattern precisely when needed, transforming how CSi Industries configures, sells, and operates their robotic palletizing systems.
Partners & Acknowledgments
Technical University of Cluj-Napoca
Romania
Research and Development Lead
Department of Manufacturing Engineering
Project Director: Assoc. Prof. Mircea Fulea, PhD
CSi Industries B.V.
Netherlands
Industry Partner and Beneficiary
Robotic Palletizing Systems Manufacturer
This collaborative research and development project was initiated in 2013 and extended through multiple addenda until 2016 due to significant scope expansion. The partnership exemplifies successful university-industry collaboration delivering practical industrial solutions grounded in advanced computer science and robotics research.
Project Objectives
Main Objective: To design and develop an advanced software application for managing palletization patterns and hardware configurations in flexible robotic palletizing cells, providing both optimal packing solutions and comprehensive system configuration capabilities.
- Develop intelligent pattern generation algorithms – To create proprietary heuristic-based algorithms that generate near-optimal box arrangements on pallets, addressing the NP-complete nature of the packing problem while providing practical solutions that maximize pallet surface utilization, minimize robot movements, and respect operational constraints.
- Create intuitive user interfaces for diverse users – To design and implement tablet-optimized, touch-friendly interfaces serving both technically proficient sales representatives configuring custom palletizing solutions and operators with minimal technical knowledge running day-to-day cell operations.
- Implement comprehensive 3D visualization – To provide real-time 3D visualization of palletization patterns, loaded pallets, and complete robotic cell configurations with interactive controls for rotation, zooming, and detailed inspection to support both sales presentations and operational validation.
- Enable automatic robot program generation – To develop seamless integration with robot control systems through automatic generation of robot movement programs based on palletization patterns, including coordinate transformation and interfacing with PLC controllers.
- Provide flexible hardware configuration management – To create a system for managing modular robotic cell configurations including robot selection, end-effector types, conveyors, safety systems, and other components with automatic compatibility verification and performance calculation.
Development Methodology
Agile Development Approach
The project followed an agile development methodology with 4-6 week sprints coordinated with regular progress meetings with CSi Industries. This iterative approach allowed for continuous refinement based on user feedback and evolving requirements as the project scope expanded.
Focus on Usability and User Experience
Special emphasis was placed on usability and UX design, particularly information architecture. Recognizing that even the most compelling content and graphics become inefficient without proper information architecture, the team carefully designed the structural framework governing visual elements, functionality, interaction mechanisms, and navigation. Poor content organization can make navigation difficult and unclear, leading to user confusion, inefficiency, increased process times, and higher costs from internal non-conformities in industrial environments.
Information architecture principles guided the organization of content to enable users to quickly adapt to application functionality and identify necessary information with minimal effort. Key design considerations included the organization system, labeling (naming) system, navigation system, and search capabilities.
The application running on a Tablet PC, optimized for touch input in industrial environments.
Results & Key Features
Core System Capabilities
- Project-based organization: Complete project management system where each project represents a palletization pattern for a product batch, with support for unlimited projects and single customizable hardware configuration
- Comprehensive input management: Detailed specification of product name, pattern number, box dimensions and weight, pallet dimensions, maximum stack height, inter-box spacing (with 3D visualization), overhang/underhang options, packaging structure selection (one-block, two-block, three-block), multi-grabbing settings (up to 20 boxes), and feeding direction
- Intelligent pattern generation: Proprietary heuristic-based algorithm generating numerous patterns by dividing pallets into blocks and sub-blocks, then filling with similarly-oriented boxes, exploring all reasonable combinations up to 5-block complexity, with automatic calculation of surface utilization percentage
- Pattern customization: Manual pattern creation, fine-tuning of generated patterns including box coordinate adjustment with automatic collision detection, 180-degree rotation for label positioning, drag-and-drop box manipulation, and group movement options
- Layer-based management: Patterns managed at layer level with flexible layer stacking to build complete pallets, cardboard sheet insertion between layers, and maximum height enforcement
Advanced Features
- 3D visualization: Real-time 3D rendering of pallets and patterns with interactive rotation, zoom, top/bottom/side views, and pick-point visualization for validation
- Robot palletizing order: Automatic calculation of collision-free palletizing sequences with visualization and manual override capability
- Multi-language menu system: Internationalized interface supporting multiple languages for global deployment
- Customizable visualization: Color selection for pattern display and configurable visual preferences
- Data export: Pattern structure export as data files for PLC interface and robot program generation
- Performance metrics: Automatic calculation of total packages per pallet, number of blocks per layer, block dimensions, cycles per layer, volume utilization, area utilization, optimum pattern recommendations, palletizing time, throughput (boxes/hour, pallets/shift), and productivity indicators
Hardware Configuration Management
The system includes comprehensive hardware configuration management with modular standard assembly lists, 3D cell layout visualization with module highlighting, automatic compatibility checking between components (robots, end-effectors, bases, conveyors, safety systems), and management of future options and incompatibilities as user-defined information.
Technical Implementation
Pattern Generation Algorithm
The core pattern generation algorithm addresses the NP-complete nature of optimal box packing through a proprietary heuristic approach. The algorithm divides the pallet into blocks and sub-blocks, exploring all reasonable combinations of block arrangements and box orientations. Testing demonstrated that the algorithm consistently identified optimal patterns (maximum boxes per pallet) across all test cases.
Each generated pattern is classified by surface area coverage percentage, number of robot movements required, and number of rectangular blocks. The algorithm accounts for multi-grabbing capabilities (robot end-effector simultaneously picking multiple boxes) and label/design orientation requirements for aesthetic pallet appearance.
Automatic Robot Program Generation
The application exports palletization patterns as structured data files containing box coordinates and rotation information. A software agent (codeblock processor in the reconfigurable cell control architecture) receives these coordinates in a JavaScript-like format specifying X, Y, Z positions, rotation angles, and placement indices for each box across all layers.
Integration with RobotStudio enabled offline development of coordinate transformation algorithms. By establishing correlation between pallet-relative coordinates from the application and robot workspace coordinates, the system generates movement instructions automatically. The robot driver translates these instructions into robot-specific motion commands, completing the seamless integration from pattern design to robot execution.
The application interface showcasing project management, pattern generation, 3D visualization, and hardware configuration capabilities.
Impact & Industrial Application
The Expert System for Smart Robots transformed CSi Industries’ approach to robotic palletizing systems, providing significant competitive advantages in both sales and operations. The system became an integral tool for sales representatives, enabling them to rapidly configure custom palletizing solutions during customer meetings, visualize complete cell layouts in 3D, and calculate precise performance metrics including throughput, cycle times, and productivity measures. This capability dramatically shortened the sales cycle and improved proposal quality.
For end users operating CSi palletizing cells, the application provided unprecedented flexibility through dynamically reconfigurable palletization patterns. Operators could easily adapt to new products or box dimensions without requiring external engineering support or lengthy reconfiguration processes. The intuitive tablet-based interface made the system accessible to operators with minimal technical training, reducing operational complexity and improving efficiency.
The automatic robot program generation capability eliminated manual programming efforts, significantly reduced setup times for new products, and virtually eliminated human errors in coordinate specification. The seamless integration between pattern design and robot execution created a truly plug-and-play experience for changing palletization configurations.
From a technical perspective, the project successfully addressed an NP-complete computational problem with practical heuristic solutions that consistently delivered optimal or near-optimal results. The pattern generation algorithm’s ability to explore complex multi-block arrangements while maintaining computational efficiency represented a significant achievement. Testing validated that the system identified maximum-density packing solutions across diverse box and pallet dimensions.
The project exemplifies successful university-industry collaboration, demonstrating how academic research in algorithms, computer graphics, and human-computer interaction can be translated into commercially valuable industrial software. The three-year project duration reflected both the initial technical challenges and the recognition of expanding opportunities as the system’s capabilities grew, ultimately delivering a comprehensive solution far exceeding the original scope of simple pattern optimization.
Last updated: January 2025 | Project Duration: 3 years (2013-2016) | Project Code: 2013111901