纯 C 实现 Pure C Implementation
Plain C 编写，无外部库依赖，极致兼容性，可在几乎任意工具链上直接编译。 Written in plain C with no external library dependencies, offering outstanding compatibility and compiling directly with almost any toolchain.
突破算力边界，让在线优化在低算力芯片上触手可及。 Break through compute limits and bring online optimization within reach on low-compute chips.
算法、精度与内存全链路针对嵌入式优化控制场景设计。 The algorithm, accuracy, and memory are all designed for embedded optimization control scenarios.
面向嵌入式优化控制场景专门优化的求解算法，在低算力芯片上也能满足高频控制周期。 The solver is specifically optimized for embedded optimization control scenarios, ensuring high-frequency control cycles even on low-power chips.
同一套算法支持 float 与 double；即使在单精度硬件上也保持稳定的数值表现与求解精度。 The same algorithm supports both float and double; even on single-precision hardware, it maintains stable numerical performance and solution accuracy.
稀疏存储与紧凑数据结构，充分利用问题结构，显著降低 RAM / Flash 占用。 Sparse storage and compact data structures fully utilize the problem structure, significantly reducing RAM / Flash usage.
零动态内存分配，内存需求在编译期确定，运行时行为可预测、无内存碎片风险。 Zero dynamic memory allocation; memory requirements are determined at compile time, making runtime behavior predictable and eliminating the risk of memory fragmentation.
无第三方依赖，拷贝源码即可集成到现有工程。 No third-party dependencies; simply copy the source code to integrate into existing projects.
Plain C 编写，无外部库依赖，极致兼容性，可在几乎任意工具链上直接编译。 Written in plain C with no external library dependencies, offering outstanding compatibility and compiling directly with almost any toolchain.
提供 C / C++ / MATLAB / Simulink 等接口，仿真验证与量产代码共用同一求解内核。 C, C++, MATLAB, and Simulink interfaces are provided, so simulation validation and production code share the same solver core.
已在 MCU、DSP、SoC 等多种平台落地，裸机与 RTOS 环境均可运行。 Already deployed on MCU, DSP, and SoC platforms, running in both bare-metal and RTOS environments.
在 QP 求解器之上提供更上层的 MPC 接口：只需描述预测模型、代价权重、约束与预测时域，
OPTIMicro 自动完成 MPC 到 QP 的转换与求解。
A higher-level MPC interface on top of the QP solver: simply describe the prediction model, cost weights, constraints, and prediction horizon,
and OPTIMicro automatically converts the MPC problem into a QP and solves it.
自动完成 MPC 到 QP 的建模与矩阵构造，无需手工推导。 Automatically builds the MPC-to-QP formulation and matrices, with no manual derivation required.
支持状态 / 输入约束、输入增量约束与软约束等常用配置。 Supports common configurations such as state and input constraints, input-rate constraints, and soft constraints.
权重与参考轨迹可在线更新，同样保持静态内存与确定性执行。 Weights and reference trajectories can be updated online while still keeping static memory and deterministic execution.
我们提供试用版本、演示案例与选型支持：告诉我们您的问题规模、目标芯片与实时性要求，我们将协助您在真实硬件上完成评估与部署。 We provide trial versions, demo cases, and product-selection support: tell us your problem size, target chip, and real-time requirements, and we will help you complete evaluation and deployment on real hardware.