[alexchunlin]

Description

Adds espp::Bno055, a driver for the Bosch BNO055 9-axis absolute-orientation IMU over I2C, plus an example and docs. espp has no BNO055 driver today. It follows the same BasePeripheral pattern as the ICM20948 and BMI270 drivers: the caller passes in probe, write and write_then_read, so the driver runs on any espp I2C bus.

• Init: resets the chip, waits for chip ID 0xA0, selects normal power mode, sets SI units (m/s², rad/s, rad, °C), optionally enables the external 32 kHz crystal, then switches to the configured fusion mode. The default is IMU mode, accel + gyro fusion with heading relative to power-on. NDOF and the non-fusion modes are available.
• update(): reads accel, mag, gyro, Euler angles, quaternion, linear acceleration, gravity, temperature and calibration status in a single 46-byte burst, and caches the result. data() returns a copy under a mutex.
• Example: components/bno055/example prints a CSV row at 100 Hz. Pins and address are set in menuconfig. The defaults are SDA 7 / SCL 8 on esp32p4 (the ESP32-P4-ETH internal bus) and 21 / 22 elsewhere, at address 0x28.
• Docs: doc/en/imu/bno055.rst, the example page, and the Doxyfile and imu/index.rst entries.

Axis remapping is left out on purpose. On an earlier robot the BNO055's register-level axis remap scrambled the quaternion axes, and doing the mounting correction in software was more reliable.

Motivation and Context

We use a BNO055 on the RAMMP MIB, an ESP32-P4-ETH wheelchair base controller (rammp-org/RAMMP-mib-fw#38). The driver started life there. This PR moves it upstream so the MIB can depend on espp/bno055 instead of carrying a local copy.

How has this been tested?

• Example build: ESP-IDF v6.0.1, esp32p4. No warnings. clang-format clean.

• Hardware: the example on an ESP32-P4-ETH (rev v1.3) with a BNO055 breakout on GPIO7/8, board still on a desk.

  • Initialized in IMU mode.
  • Gyro calibration reached 3 within about 1 s.
  • Roll −0.0533 rad and pitch −0.0444 rad, which matches the tilt from the accelerometer's gravity vector (−0.49, −0.41, 9.48 m/s²).
  • About 91 Hz at the default 115200 baud, where the CSV line length is the limit, not the I2C read. A burst read takes about 1.5 ms.

• In-application: the same driver, under mib:: instead of espp::, has run in the MIB firmware at 100 Hz alongside Ethernet and RTPS.

• Tilt by hand, in the MIB firmware (same driver):

  • I tilted the sensor to +34° pitch and −32° roll at 100 Hz.
  • Over the roll test, fused roll agreed with roll computed from the driver's own accelerometer output (atan2(a_x, |a_yz|)) to within a median 0.19°. That supports the driver's unit scaling and axis mapping.
  • The left panel also shows the sensor unplugged and replugged. The recovery there is application logic (probe, then init() again), not part of this component.

  

Types of changes

• New feature (non-breaking change which adds functionality)
• Documentation Update
• Software change

Checklist:

• My change requires a change to the documentation.
• I have added / updated the documentation related to this change via either README or WIKI

Software

• My code follows the code style of this project.

🤖 Generated with Claude Code

[github-actions]

✅Static analysis result - no issues found! ✅