Mobile robot control or an assistive system for the visual disabled needs real-time and stable estimation results for the position and heading information. If the information is estimated in time difference of arrival (TDOA) based sensor network, a linear estimator can be applied to this problem which is not suffering from the initial guess problem and can be derived in a recursive formula. However, it requires a computational burden to compensate the correlation error and biased error caused by TDOA measurement noise. Additionally, many literatures have treated TDOA measurements rather focusing on the stationary target. In this work, we propose a position and heading estimator based on the instrument variable (IV) method which is able to flee from the correlation problem and reflects a constant velocity model. Through simulations, the performance of the proposed method is verified by showing the position and heading estimation results for a moving target.