Global Navigation Satellite Systems (GNSS)

GPS (Global Positioning System) is a specific constellation of satellites owned by the USA. GNSS (Global Navigation Satellite Systems) is the overarching term that includes GPS, but also Europe's Galileo, Russia's GLONASS, and China's BeiDou. Modern robots use GNSS receivers to access all these satellites simultaneously for better accuracy and availability.

Generally, no. Satellite signals are too weak to penetrate roofs and thick walls effectively, leading to complete signal loss. For indoor navigation, robots rely on LiDAR SLAM, visual SLAM, or infrastructure-based systems like magnetic tape or QR codes.

RTK (Real-Time Kinematic) is a technique that improves GNSS accuracy from meters to centimeters. If your robot operates in large open spaces without tight constraints (like a drone), standard GNSS might suffice. However, for ground robots that need to stay on a sidewalk or dock precisely, RTK is absolutely necessary.

When tall buildings block satellite signals, robots switch to "Dead Reckoning." This involves using onboard IMUs (accelerometers/gyroscopes) and wheel encoders to estimate position based on the last known GNSS coordinate. This drift increases over time, so the signal must be reacquired quickly.

Standard phone GPS updates at 1Hz (once per second). Robotics applications typically require 5Hz, 10Hz, or even 20Hz. Faster update rates ensure the control loop has fresh position data, which is critical for smooth path following at higher speeds.

Not necessarily. While you can install a local base station, many regions offer NTRIP services (Networked Transport of RTCM via Internet Protocol). This allows the robot to receive correction data over a cellular connection from a network of existing reference stations.

GNSS is largely weather-independent, working through rain, fog, and snow, which gives it an advantage over cameras or LiDAR in some scenarios. However, heavy ionospheric activity (solar storms) can degrade signal accuracy, though RTK helps mitigate these atmospheric errors.

Costs have dropped significantly. While survey-grade equipment costs thousands, modern dual-band RTK modules for robotics (like u-blox F9P based chips) are available for a few hundred dollars. The main cost driver is often the integration engineering and antenna quality.

Dual-band receivers listen to two different frequencies from satellites simultaneously. This allows the receiver to better calculate and correct for ionospheric delays directly. For robotics, this results in faster "time to first fix" and much more robust signal holding in difficult environments.

In ROS/ROS2, GNSS data is typically handled via the `sensor_msgs/NavSatFix` message type. Packages like `robot_localization` are used to fuse this GPS data with IMU and odometry to publish a continuous `map` to `base_link` transform, allowing the navigation stack to plan global paths.

Multipath occurs when signals bounce off buildings or ground surfaces before hitting the antenna, confusing the receiver about the distance. Robots mitigate this using high-quality antennas (choke ring or helix), placing antennas high up, and using advanced filtering algorithms in the receiver firmware.

GNSS modules are generally power-efficient, consuming between 100mW to 500mW during active tracking. This is negligible for large AGVs but can be a consideration for small, battery-constrained IoT sensors or micro-robots. The active antenna also draws a small amount of current via the bias tee.