Development of firmware (Embedded Firmware) for a BLE wearable device (Nordic nRF52840 / nRF5340 + PPG)

Content Management Systems, Web Programming — incorrectly specified categories?

607 USD

Project translated automatically. Log in or register, to view the original

We are looking for an experienced Embedded developer to create energy-efficient firmware for a wearable device (form factor: chest strap) for monitoring psychophysiological state. The circuit design is complete, and development will be based on a development kit with subsequent transfer to a custom board.

Stack and hardware base:

MCU: Nordic Semiconductor nRF52840 or nRF5340.
Development environment: nRF Connect SDK / Zephyr RTOS (strictly C/C++).
Peripherals:
1. Optical pulse sensor (PPG) — operates via I2C.
2. 6-axis inertial sensor (IMU accelerometer + gyroscope) — operates via SPI.
3. Haptic motor driver (LRA 1030) — controlled via PWM.
4. External SPI Flash memory for cyclic logging.

Work scope by stages

Stage 1: Bring-up, peripheral drivers, and BLE telemetry

Goal: Bring the hardware to life, set up stable data collection, and transmit it to a smartphone/PC.

Configure I2C and SPI buses using DMA (Direct Memory Access) to offload the processor.
Write/integrate drivers for PPG and IMU (configure polling frequency, sensor FIFO buffers).
Deploy a custom BLE profile (GATT service) for real-time transmission of raw data from sensors.
Basic initialization of PWM for controlling the haptic LRA motor (launch test vibration patterns).

Stage 2: Algorithmic layer (DSP and multi-level feedback logic)

Goal: Implement mathematical signal processing directly on the chip (Edge Computing) and device reaction logic.

Implement algorithms for filtering motion artifacts: clean the PPG signal using accelerometer data (IMU).
Extract R-R intervals (HRV) from the cleaned PPG signal.
Implement 4-level system stability logic:
1. Drift Level: Track gradual changes in HRV and breathing.
2. Prediction Level: Calculate a mathematical trend for deterioration.
3. Micro-correction Level: Generate barely noticeable haptic patterns on the LRA motor when a target data pattern is detected.
4. Hardware Reset Level: Implement safety logic (high-intensity haptic pulse at critical metrics).

Stage 3: Energy saving, logging, and fault tolerance

Goal: Prepare the device for autonomous operation from a miniature battery and ensure "24/7" stability.

Deep optimization of energy consumption. Put the MCU into sleep mode (System ON/OFF) between measurement cycles. The processor should wake up only on interrupts from sensor FIFO buffers or a timer.
Write a fault-tolerant file system in external SPI Flash (e.g., LittleFS) for autonomous cyclic log saving.
Configure a hardware watchdog timer to prevent firmware hangs.

Requirements for the performer:

Experience in commercial development for Nordic chips (nRF52/nRF53) of at least 3 years.
Proficient in Zephyr RTOS and debugging tools (Segger Embedded Studio / VS Code, J-Link).
Understanding of digital signal filtering principles (DSP) and low-power design.
Having your own debugging kit (nRF52840-DK or similar) for starting work is welcome.

Work format:

Milestone payments with results documented on GitHub. Please indicate your hourly rate or approximate cost and timeline estimates for each of the three stages.

Proposals 4 Discussions

Please select

Oleg Grigoryev

32 0

Budget: 27000 UAH Deadline: 84 days

Look, there's a nuance here - the full scope of development does not fit into 27,000 UAH. Within this budget, we can reasonably allocate an engineering mini-stage for 5-7 days - to check the board, component set, PPG risks, BLE, and energy saving, after which we can provide a solid estimate.

For the complete development of all three stages, I would estimate starting from 360,000 UAH and 10-12 weeks. The breakdown for the stages is as follows - stage 1 from 90,000 to 140,000 UAH and 3-4 weeks, stage 2 from 160,000 to 240,000 UAH and 4-6 weeks, stage 3 from 110,000 to 170,000 UAH and 3-4 weeks. Hourly rates for a senior firmware developer start from 45 USD per hour, but for this task, it's more comfortable to go by stages rather than counting every micro-pause of the oscilloscope.

Technically, I would conduct the work as follows - first nRF Connect SDK and Zephyr, I2C and SPI buses with DMA, checking FIFO sensors, a custom BLE GATT for raw data, then the processing layer for PPG+IMU, HRV, and tactile patterns, after that energy saving, LittleFS, watchdog timer, and failure scenarios.

From you, I will need the schematic, BOM, datasheets, the chosen model nRF52840 or nRF5340, access to the GitHub repository according to marketplace rules, and acceptance criteria for each stage. Well, without raw PPG+IMU signals, the algorithmic layer will be a bit like guessing on coffee grounds - and that's not engineering, but my internal circus with a calculator.

We need to clarify two things -
> what exact model of PPG, IMU, SPI Flash, and LRA driver has been chosen
> is there a test set of raw PPG+IMU data or do we need to collect it during the initial launch

Similar projects in terms of subject matter and complexity by Ingello -
> https://business.ingello.com/lita - a medical application and logic for sensitive data
> https://business.ingello.com/rapport - a platform in the field of psychology and psychological health
> https://business.ingello.com/vorfahr - a complex product system with AI and integrations

Our FLH landing -
> https://systems-fl.ingello.com

Similar project: Рефаткоринг приложения

Vasil Bodnaruk

198 2

Budget: 27000 UAH Deadline: 18 days

Denis, I see there is a complex but very interesting embedded project here: BLE telemetry, PPG/IMU drivers, PWM for LRA, and further DSP logic with power-saving mode. I have worked with C/C++ and low-level integrations, quickly bringing up and stabilizing peripherals during the debugging phase. I can break the work into 3 milestones, evaluate each stage separately, and keep the code on GitHub. I am ready to discuss the details regarding the sensors, BLE profile, and test board.

Lilia N.

1 0

Projects -
Rating -
Rating 583

Budget: 27000 UAH Deadline: 1 day

Hello! I have experience with Nordic nRF52840/nRF5340, Zephyr RTOS, BLE GATT, and PPG/IMU signal processing. I would be happy to help you!

Daria Kratofil

0 0

Projects -
Rating -
Rating 196

Budget: 27000 UAH Deadline: 53 days

We already have a practically ready groundwork in healthtech, BLE telemetry, and signal processing that can be quickly adapted for Nordic and Zephyr, so we can discuss the details here and proceed step by step.
Regarding the budget - 27,000 UAH for such a set of works will likely only cover a short audit of the schematic and a plan for reviving the board.
The rate guideline is 35-45 USD per hour if we consider the entire cycle from DK to custom board.
Stage 1 - 65,000-95,000 UAH and 10-15 working days.
Stage 2 - 140,000-240,000 UAH and 25-40 working days, because DSP and HRV usually require iterations on real logs.
Stage 3 - 95,000-160,000 UAH and 18-30 working days.
The preliminary cost for the entire cycle is from 300,000 UAH, more precisely after the PPG model, IMU, and sampling frequencies.
We can start with the first stage - drivers, BLE GATT, PWM patterns, basic logs, and consumption measurements.
From you, we need the schematic, pinout table, selected PPG and IMU, access to GitHub, requirements for the BLE packet format, and criteria for critical states.
Note, there is a nuance - regarding algorithms, it is important to clarify whether there are already formulas and thresholds or if a research part on logs needs to be done.
The second question is what the target autonomy mode is, hours or days, and what battery capacity is required.
Healthtech and medical processes - https://business.ingello.com/lita - seems similar in medical logic, data, and reliability.
Medical platform with roles and data quality - https://business.ingello.com/rapport - close to the requirements for stability and verifiability.
It feels like AI is better added after the logging stage - to find personal patterns, rather than embedding magic in the firmware at the first step.
Main page for FLH - https://systems-fl.ingello.com.