Purpose / Objective

• Develop an AI-driven tool using Large Language Models (LLMs) to analyze and migrate embedded projects from non-STM32 MCUs to the most suitable STM32 device.
• Automate peripheral, feature and constraint mapping to speed up MCU selection and reduce manual porting effort.

Main Tasks / Work to Be Done

• Structured Data Generation from source MCU documentation and drivers (extract datasheets, reference manuals, driver APIs and peripheral descriptions).
• Static and Dynamic Analysis of the embedded project to identify used peripherals, middleware, interrupts, clock configurations and system features.
• STM32 MCU selection based on feature matching and constraints (memory, peripherals, pinout, power, real-time requirements).
• Software emulation of missing hardware features where direct hardware mapping is not available.
• Integration of LLM-assisted automation tools to propose code refactorings, peripheral adaptation and migration patches.
• Validation and testing on STM32 hardware: deploy migrated firmware, run functional tests and compare behavior against original project.

Keywords / Technologies / Deliverables

• Keywords: STM32, LLM, MCU Migration, Peripheral mapping, Code refactoring, Feature extraction.
• Expected technologies/tools: STM32 HAL/LL/CubeMX, static analysis tools, dynamic tracing/profiling tools, emulation frameworks, LLM frameworks or APIs (for assistant automation).
• Deliverables: a migration assistant/toolchain (prototype), migration reports for sample projects, test procedures and validation logs, documentation describing mapping rules and limitations.

Validation & Evaluation

• Functional validation on target STM32 hardware with representative test cases from source projects.
• Quality metrics: correctness of peripheral mapping, compilation success rate, runtime behavior parity and documented manual interventions.

Context / Constraints

• Work involves a combination of reverse-engineering (feature extraction), automated code transformation and hardware validation on STM32 platforms.
• Project requires careful handling of MCU-specific constraints (timing, DMA, interrupts, peripheral variations) and producing traceable migration decisions.

Application

• Position referenced as: STTunis 2026 Project_ID08 : LLM-Based Migration of Embedded Projects to STM32.
• To apply use the online link provided below.