I am looking for a freelancer to design the data-processing and trajectory estimation solution for a compact embedded IMU/GPS logger project. What I Need: Develop (and deliver) a robust, documented Python algorithm to reconstruct the trajectory or path from logged IMU data (acceleration and orientation, optionally with GPS). The emphasis is post-processing/analysis with Python. The freelancer should build and demo the workflow using a readily available, minimal embedded system (such as STM32, ESP32 or similar + IMU, with or without GPS). The hardware should be simple and reproducible, just enough to prove the workflow and verify results. If some pre-processing/filtering (e.g., bias removal, simple fusion) is best done on the embedded device, you can implement it there as well, but all position/trajectory estimation must be reproducible in Python using the raw or pre-processed logs. Resulting script(s) must: Take CSV logs as input (format: timestamp, yaw/pitch/roll or quaternion/Euler, free acceleration, and GPS if available). Output: position/trajectory reconstruction. Generate 2D and 3D trajectory plots. Requirements: All deliverables in Python; no MATLAB or proprietary toolchains. Algorithm, plotting, and data processing must be clear, well commented, and documented so I (or others) can reuse or extend them. Use open-source tools and libraries only. Hardware must be built from readily available microcontrollers (ESP32/STM32/Arduino/etc), IMUs, and (optionally) GPS modules; specify all part numbers and wiring. If you propose useful on-device algorithms (e.g., motion/noise filtering) that simplify post-processing, implement them as Arduino/ESP32 code and supply full source. When you reply, include: A detailed project proposal on your approach to IMU-only and IMU+GPS trajectory reconstruction—what Python libraries and algorithms you’d use, and why. A summary of the hardware you propose to use or can assemble for testing. An example input CSV format and example output (trajectory plot screenshot or sample data). Timeline for both hardware setup and code delivery. Notes: I am not asking for a full finished product—just the proof-of-concept hardware and the complete algorithmic and Python side. Strong preference for Python (pandas/numpy/matplotlib) only. If your method can be partly accelerated or cleaned up by processing/filtering on the microcontroller (e.g., bias compensation) before logging, you can implement and document that too. If you have modifications or suggestions that would further improve the practical utility or accuracy (especially for indoor/IMU-only use), feel free to mention them in your proposal.