This paper presents the preliminary results of MUSE, a MUlti-sensor State Estimator for quadruped robots, combining data from IMU, encoders, and cameras to accurately estimate pose and velocity, even in challenging environments such as uneven terrain. Experiments on a Unitree Aliengo robot, tested on stairs, rocks, and slippery surfaces, show MUSE's superior performance compared to using only a T265 tracking camera, providing reliable and high-frequency state estimation.