A New Adaptive Estimation Algorithm Based on CT Model and Ellipsoid Constraint

Kalman filtering algorithm is frequently used in the kinematic measurement of Global Navigation Satellite System (GNSS) for positioning calculation. The vehicle often appears maneuvering turns in GNSS navigation and positioning with randomness and unpredictability, which makes it impossible to be described with an accurate models. If the Kalman filtering algorithm based on Constant Velocity (CV) or Constant Acceleration (CA) model is still applied for navigation solution, the accuracy and reliability of the filtering results will be significantly reduced. Aiming at the problem that the current dynamic model has a large error under the maneuvering turns, this paper combines the Coordinated Turn (CT) model and the fixed elevation algorithm to estimate the horizontal turning rate of the vehicle in real time, and limits the deviation of Z direction through the ellipsoid constraint, which ameliorates the impact of the dynamic model deviation on navigation solutions. Experiments show that no matter in left turn, right turn or compound road section, the adaptive filtering algorithm proposed in this paper can better control the influence of dynamic model deviation, and effectively improve the positioning accuracy and reliability of the vehicle during maneuvering turns.