I am developing a mobile robot that operates outdoors within an area of approximately 1 km², with a maximum speed of around 2 meters per second. I plan to set up a private RTK base station using the ZED-F9P module.
For the rover unit mounted on the robot, I would like to ask for your recommendation:
Should I use ZED-F9P or ZED-F9R as the rover, considering the following requirements?
Accurate RTK positioning for waypoint-based navigation and autonomous missions
Reliable heading information, even when the robot is moving slowly or stationary
Robust performance under varying outdoor conditions
Ease of integration with microcontrollers (e.g., ESP32, Raspberry Pi)
Additionally, I would like to ask:
Will overcast skies, heavy cloud cover, or low visibility weather conditions significantly affect the RTK positioning accuracy or fix stability?
Personally I think using F9P is a better choice, you can integrate with your own IMU/AHRS, and have a lot more control of the experience and performative behaviour of the robot.
Now the F9R has a RLM (Robot Lawn Mower) dynamic which is workable at these lower speeds (1-2 m/s), but does require you provide wheel tick or motor drive data for both rear wheels (left/right). Most of the other dynamics expect velocity getting to 30 km/h to attain sensor fusion / calibration.
Heading, you kind of want TWO F9P in a moving base configuration so you can get orientation whilst static, or spinning on the spot. GNSS isn’t inherently directional, just determines via motions from A to B
Reception is mostly impacted by physical obstructions in the signal path, so walls, buildings, trees, etc.
Weather is less impactful, but are differences day/night and with scintillation, that latter is more of an issue in Brazil / near equator.
Micro-controllers, shouldn’t be an issue I2C and UART are commonly used.
Robustness of RTK starts at your BASE, visibility of sky from antenna, as this establishes the common view the ROVER can work from. Data via various sources, the type of radio link you have with your BASE(s), redundancy and throw there. Can use SPARTN / PointPerfect data via L-Band (satellite) using NEO-D9S
Today might consider ZED-X20P for new designs, will have support for Galileo HAS in 26Q1, and also internal L-Band reception for SPARTN
My personal opinion is the ZED-F9R (Dead Reckoning) is useful when a GNSS Solution isn’t possible (in a temporary, short time-frame).
So depending how how critical this is to you, :
Reliable heading information, even when the robot is moving slowly or stationary
A Moving Base “might” be a better choice verses Dead Reckoning…
Naturally, there can be an impact on the RF signals.
But I’m generally more concerned with Space Weather verses my “Local” Weather.
Locally, a storm nearby impacts a smaller portion of the observable sky (sats).
I would never argue that bad weather wont impact RTK performance. Newer receivers have ways to mitigate that somewhat.
[Edit] Ah… @clive1 was a little faster pulling the trigger
