HiRoCast (High-Speed Robot Casting) is an advanced robotics optimization system designed to solve a critical industrial challenge: transporting hot liquid materials from point A to point B as quickly as possible without spillage. This project combines computational fluid dynamics (CFD), optimal control theory, and robotic path planning to achieve time-optimal trajectories while preventing liquid sloshing.
Traditional robotic liquid transport faces a fundamental trade-off:
- Fast movement → Risk of spillage due to fluid sloshing
- Slow movement → Safe but inefficient
This project solves this problem through intelligent trajectory optimization that accounts for fluid dynamics in real-time.
- Optimal Control Algorithm: Time-optimal trajectory generation using constrained nonlinear optimization (fmincon)
- CFD Integration: Real-time fluid dynamics simulation to predict and prevent sloshing
- Multi-Constraint Optimization: Respects joint limits, velocity, acceleration, and jerk constraints
- 6-DOF Robot Control: Full kinematic modeling for industrial robot arms (KUKA KR60)
- Fluid Mixture Analysis: Supports various water-glycerin mixtures with different viscosities
- Validation Framework: Comprehensive testing against physical constraints and fluid behavior
- KRL Code Generation: Automatic generation of KUKA Robot Language code for deployment
The system was tested with various fluid mixtures to analyze sloshing behavior under different viscosity conditions:
Sloshing angle analysis for pure water (100% H₂O)
X and Y angle analysis during trajectory execution
| Metric | Before Optimization | After Optimization | Improvement |
|---|---|---|---|
| Transfer Time | ~15s | ~8s | 47% faster |
| Spillage Risk | High | Minimal | Safe operation |
| Path Smoothness | Poor | Optimal | Continuous C³ |
┌─────────────────────────────────────────────────────────────┐
│ HiRoCast Master Control │
└─────────────────────────────────────────────────────────────┘
│
┌─────────────┴─────────────┐
│ │
┌───────▼────────┐ ┌───────▼────────┐
│ Path Planning │ │ CFD Simulation │
│ & Kinematics │◄────────┤ (OpenFOAM) │
└───────┬────────┘ └────────────────┘
│
┌───────▼────────┐
│ Optimal │
│ Control │
│ (fmincon) │
└───────┬────────┘
│
┌───────▼────────┐
│ Validation │
│ & Testing │
└───────┬────────┘
│
┌───────▼────────┐
│ KRL Code Gen │
│ (Robot Deploy)│
└────────────────┘
The system solves a constrained optimization problem:
Objective Function:
minimize: Σ(time_intervals)
Subject to constraints:
- Joint angle limits: θᵢ ∈ [θᵢ_min, θᵢ_max]
- Velocity limits: |θ̇ᵢ| ≤ 400°/s
- Acceleration limits: |θ̈ᵢ| ≤ 200°/s²
- Jerk limits: |θ⃛ᵢ| ≤ 1000°/s³
- Sloshing constraints: Φ(acceleration, orientation) ≤ Φ_max
- Boundary conditions: Start/end positions and zero velocity
- Model: KUKA KR60 (6-DOF Industrial Robot)
- Payload: Ladle with liquid (variable mass)
- Workspace: 3D Cartesian space with orientation control
- Control Rate: Real-time trajectory execution
- CFD Solver: OpenFOAM (Volume of Fluid method)
- Fluid Models: Water-Glycerin mixtures (0-100% concentration)
- Sloshing Detection: Angular deflection analysis in X-Y plane
- Validation: Regression models trained on CFD simulation data
HiRoCast/
├── src/ # Source code
│ ├── optimal-control/ # Optimization algorithms
│ ├── kinematics/ # Forward/inverse kinematics
│ ├── validation/ # Constraint validation
│ └── visualization/ # Result plotting
├── data/ # Simulation data
│ ├── fluid-simulations/ # CFD results
│ └── measurement-data/ # Experimental measurements
├── docs/ # Documentation
│ ├── assets/ # Images and diagrams
│ ├── methodology.md # Technical methodology
│ └── api-reference.md # Code documentation
├── results/ # Optimization results
└── legacy/ # Archive of old implementations
% Required MATLAB Toolboxes:
% - Optimization Toolbox
% - Robotics System Toolbox (optional)
% - Curve Fitting Toolbox
% External Dependencies:
% - OpenFOAM (for CFD simulations)
% - RoboDK (for robot visualization)- Clone the repository
git clone https://github.com/yourusername/HiRoCast.git
cd HiRoCast- Run the main optimization
% Open MATLAB and navigate to the project directory
cd('src/optimal-control')
% Set up the environment
KSetUp;
% Run the master control script
HiRoCast_Master;- Visualize results
% View trajectory optimization results
show_spline(optimized_solution, 'Optimized Trajectory');
% Analyze fluid dynamics
CFDBeschleunigungen;- Method: Interior-point algorithm (fmincon)
- Spline Interpolation: Cubic splines for smooth trajectories
- Multi-start: Multiple initial conditions for global optimization
- CFD Integration: Pre-computed sloshing corridors for different orientations
- Regression Model: Machine learning model for real-time prediction
- Angle Calculation: 2D projection analysis in movement plane
- Minimum Jerk Trajectory: Initial path generation
- Collision Avoidance: Workspace constraints
- Reorientation Strategy: Optimal ladle orientation during transport
This project demonstrates expertise in:
- Optimal Control Theory: Constrained nonlinear optimization
- Computational Fluid Dynamics: Multi-phase flow simulation
- Robotics: Kinematics, dynamics, and trajectory planning
- Numerical Methods: Spline interpolation, regression analysis
- Software Engineering: Modular MATLAB architecture
- Industrial Automation: Real robot deployment (KUKA KRL)
For detailed technical documentation, see:
- Project Overview - Executive summary and impact
- Quick Start Guide - Get started in 10 minutes
- Methodology - Mathematical formulation and algorithms
- API Reference - Function documentation
- Results Analysis - Performance evaluation
- Repository Structure - Code organization guide