Mitsubishi Heavy Industries Accelerates Drone Development with Simulink Models
Control Algorithms Developed and Verified Safely and Efficiently
“By applying Simulink’s automatic code generation technology to the entire workflow, we were able to improve efficiency and accuracy from algorithm design and implementation to functional verification.”
Key Outcomes
- MHI built a consistent development flow from control design to PX4 implementation using Simulink models.
- MHI improved quality and shortened development time through continuous testing of the PX4 control algorithms with HIL simulation.
- MHI gained insight into utilizing Simulink models for PX4 autopilot integration.
Hardware-in-the-loop architecture for UAV development integrating PX4 control models, flight controller hardware, Speedgoat real-time simulation, and visualization tools.
The widespread use of unmanned aerial vehicles (UAVs) and drones in industry, logistics, and disaster response has driven the need for more sophisticated, high-precision control algorithms. To develop its next-generation drones, the Research & Innovation Center at Mitsubishi Heavy Industries (MHI) needed to efficiently develop and validate control modules that incorporate various control technologies into new algorithms. However, the workflow presented challenges such as bottlenecks in development speed, manual code implementation and verification of control algorithms, and the execution of flight tests in the early stages of development. These challenges introduced risks and constraints related to safety, cost, and repeatability.
MHI built a development environment that could design products and accurately verify functions—without actual equipment. Using Simulink® and UAV Toolbox, engineers can design control algorithms and efficiently interface with sensors and actuators through a PX4-specific blockset. This model-based approach allows them to visualize and verify flight control logic from the earliest stages of development. Using Embedded Coder®, they can automatically generate high-quality C code from the Simulink models. They can then integrate the generated code directly into PX4 autopilot firmware, significantly reducing development effort compared to manual coding. In addition, MHI combined Simulink Real-Time™ with Speedgoat® real-time hardware to build a hardware-in-the-loop (HIL) simulation environment. By simulating sensor signals and flight dynamics in real time, the team can validate control algorithm behavior under realistic operating conditions before proceeding to physical flight tests.
