我喜欢把真实问题做成可运行、可验证、可维护的小项目。
I build practical software across WeChat Mini Programs, embedded systems, ROS, MQTT, and scientific modeling. My current focus is local-first tools, reproducible analysis, and clear evidence boundaries.
| Project | What it demonstrates | Quality signal |
|---|---|---|
| Pugo Toolbox | 31 个本地优先的微信小程序实用工具；无需登录、后端或 API 密钥 | 74 automated checks, release preflight, GitHub Actions |
| Heliostat Field Optimization | 根据 2023 数学建模参赛论文与留存结果重构的定日镜光学效率、数据审计和小规模优化 | Python, 16 tests, generated figures, documented model limits |
| Smart Home Transport Robot | 2023 国家级三等奖团队项目的导航算法赛后重构，展示 Dijkstra 全局规划与 DWA 动态避障 | Interactive replay, deterministic demo, explicit evidence boundary |
| ROS–MQTT Vehicle Bridge | ROS 1 与 MQTT 之间可配置的 JSON 桥接，面向机器人与车辆数据链路 | C++, security-focused configuration, documented architecture |
| Arduino RGB Plant Lamp | Arduino RGB 植物灯、串口 PWM 控制、功耗估算与接线说明 | Offline tests, wiring docs, versioned releases |
- Useful first — 先完成能解决具体问题的端到端版本。
- Evidence over claims — 用测试、CI、复现步骤和发布记录说明质量。
- Privacy by design — 能在本地完成的处理，不依赖账号、云端或长期凭据。
- Open collaboration — 欢迎通过项目 Issue 提交问题、建议或使用场景。
- Improve Pugo Toolbox with real user feedback and contributor-friendly issues.
- Extend reproducible mathematical-modeling work without overstating reconstructed results.
- Develop explainable robot navigation and reliable bridges between embedded devices, middleware, and useful interfaces.
- Share implementation notes that help other learners reproduce the result.
如果某个项目对你有帮助，欢迎 Star、试用并留下反馈。真实的使用意见比单纯的数字更有价值。