Gyroscope Drift Estimation of a GPS/MEMSINS Smartphone Sensor Integration Navigation System for Kayaking

In 2016, Google released an application programming interface on Android smartphones which allowed users to access the raw Global Positioning System (GPS) measurements. This recent accessibility of raw GPS measurements coupled with the rising demand for handheld navigation and tracking for consumer applications has created an opportunity to test the capabilities of a smartphone tracking system for river kayaking. However, GPS-only navigation can be intermittent or unreliable due to satellite signal outages in environments such as river canyons, therefore an inertial navigation system (INS) can be integrated to provide a continuous solution during these outages. In this paper, a system to continuously track a kayak through a river using a low-cost, smartphone GPS/INS based positioning system was developed. Low-cost inertial sensors, such as those found in a smartphone, are incapable of providing an accurate solution for a long period of time during GPS outages due to the accumulation of sensor errors. To overcome this, kayak-specific motion constraints are proposed that estimate the gyroscope bias, pitch and roll angles, and accelerometer bias with the assumption that the kayak’s pitch and roll angles are centered around zero, cyclic in nature, and return to the same orientation at regular time intervals. To evaluate this proposed solution, multiple field datasets were collected using a Google Pixel 4 that was fixed to the deck of a kayak. The experimental results indicate an improvement in the navigation performance verifying that the proposed kayak-specific motion constraints increased the accuracy of the solution during GPS outages.