Haier Smart Home Uses Model-Based Design Approach to Improve Test Coverage
New Workflow Streamlines Development While Reducing Errors
“One by one, we tested the whole system. Model-Based Design and testing have improved efficiency.”
Key Outcomes
- A Model-Based Design workflow increased development efficiency by 30% and allowed functional decoupling, component visualization, and automatic one-click code generation
- The testing time of the reservation heating function was reduced from 45 days to 40 minutes
- Efficient automated testing improved 100% of test coverage and saved 30% of testing resources and development costs
Haier Smart Home, a division of Haier Group Corporation, specializes in developing smart home appliance ecosystem products. Water network products are a major part of Haier Smart Home’s business. Engineers at Haier Smart Home have adhered to traditional embedded software development processes to build their legacy products. This traditional approach involves complex coding, sometimes reaching up to 20,000 lines, which can increase the probability of errors. Additionally, it can lead to long debugging cycles, low test coverage, and inefficiencies that can impact project deadlines.
Haier engineers implemented Model-Based Design to integrate advanced features into their products. Architectures and modules were split to develop separate models based on the set of requirements. The team used Simulink® to construct fundamental blocks and standardize shared libraries to offer cross-functional and collaborative development for different product lines.
Model-Based Design enabled algorithm design by creating a digital model of the final software, with the team performing logic modeling via Stateflow® to test various scenarios. Embedded Coder® was utilized to automate the generation of control algorithm code, streamlining the development process. Simulink also allowed for the development of a simulation platform for motor control systems to verify core algorithms. Software testing was done under abnormal conditions—such as short circuits, open circuits, and system failures—without reliance on physical hardware. In conclusion, the team performed V&V automation using Simulink Test™ to improve test coverage.
