In the positioning process

the positioning accuracy of mobile robot drops sharply because of the sensor measurement error and the pose error caused by the robot model. In view of this phenomenon

a new algorithm based on multi-innovation Kalman filtering（ MR-AKF） is proposed. Based on the standard Kalman filter

when the measured value of the sensor is error

the robust weighting factor is introduced

and the estimation accuracy of the filter is improved by changing the weight of the error measurement value; when the robot pose is error

an adaptive factor is introduced to resist the filtering divergence caused by the pose error by adjusting the size of the state covariance matrix. At the same time

introduce multi-innovation which is the innovation vector of multiple moments to further improve the accuracy of the nonlinear system. The experimental results show that the filtering algorithm can effectively improve the positioning accuracy in the presence of measurement errors and pose errors.