In order to improve ability of bionic quadruped robot to percept location information of obstacles in complex and dynamic environment

information fusion between binocular sensor and ultrasonic ranging sensor is researched under condition that the robot moves in walking gait on structured road surface. Firstly

distance information of obstacles from these two sensors is processed by Kalman filtering algorithm

which is helpful to alleviate interference caused by noise. Then

use two groups of state vector estimation value and covariance matrix obtained by filtering

fusion processing can be carried out by STF fusion algorithm. Simulation results show that curve of estimation value of distance information after filtering track curve of real value well

which shows that Kalman filtering algorithm plays an effective role in inhibiting noise; after fusion

the variance of the estimation value is distinctly smaller

which illuminates that the information of the obstacles becomes more accurate. Finally

the need for precision is met when the robot works in complex and dynamic environment.