Vehicle-mounted positioning navigation method based on self-adaptive odometer model

Abstract

The invention discloses a vehicle-mounted positioning navigation method based on a self-adaptive odometer model. The method comprises the steps that multiple motion models and a state vector transfermatrix among all the models are established; the initial motion state of a vehicle is estimated through odometer data of a left driving wheel and a right driving wheel and geometric constraint; in combination with volumetric Kalman filtering, all the motion models are filtered in parallel; and the motion model of the vehicle at the current moment is judged according to the odometer data of the left driving wheel and the right driving wheel, and state information of the vehicle is fused and output to serve as an initial value at the next moment for filtering iteration. According to the method,an optimal gain matrix is determined according to a covariance matrix of estimated vectors and measured vectors by use of volumetric Kalman filtering, the method does not have extra requirements on asystem equation and a measurement equation, and the method has higher superiority under the condition of a nonlinear system; according to the geometric constraint relation among odometers, the positioning precision and instantaneity of vehicle travel are improved; and odometer output information is not influenced by building sheltering, and therefore the continuity of positioning can be improved.

Description

technical field [0001] The invention relates to vehicle positioning and navigation, in particular to a vehicle positioning and navigation method based on an adaptive odometer model. Background technique [0002] Car navigation system has become a necessary configuration for people to travel by car. Most of the current car positioning and navigation systems are nonlinear, and the system equations and observation equations of vehicle maneuvering models are also nonlinear. The existing extended Kalman filter algorithm performs linear approximation to the nonlinear function at a single point, and the approximation to the system state distribution is relatively simple, and when the system state suddenly changes, the convergence is relatively slow, and it cannot be dealt with in time. The safety aspects of emergencies are still not satisfied. [0003] The existing vehicle positioning and navigation systems mostly obtain information such as the speed and position of the positionin...

AI Technical Summary

Problems solved by technology

However, GPS positioning also has certain shortcomings. When the GPS signal is blocked by buildings, the continuity of its positioning is destroyed, and it is difficult to ensure the stability of positioning and navigation functions under special circumstances.

Moreover, the error range of GPS in vehicle positioning is about 14m, and the positioning accuracy still needs to be improved

Images