Indoor localization and position tracking of people and robots are well-known measurement problems, which have recently attracted the attention of researchers both in industry and academia [1]. While many different localization techniques exist, the most effective solutions usually rely on the fusion of heterogeneous sensor data, and particularly dead reckoning systems (eg based on inertial measurement units) and wireless, ultrasonic or optical systems able to measure the absolute position and/or the orientation of the target within a given reference frame [2]. The key advantage of combining dead reckoning with absolute positioning techniques is the possibility to achieve excellent localization coverage and good scalability, while keeping positioning accuracy and costs within desired boundaries. Moreover, when wheeled robots are considered, dead reckoning can be effectively implemented through odometry, ie by installing incremental encoders on the wheels. Unfortunately, the path estimated through odometry is typically affected by growing uncertainty due to the accumulation of noise and systematic contributions perturbing encoder-based measurements. This problem can be mitigated if suitable landmarks (eg Quick Response (QR) codes associated with a given position and orientation in the chosen reference frame) are detected by a camera [3]. In this way, the drift of estimated position and orientation inherently due to dead-reckoning can be partially compensated thus keeping localization uncertainty bounded. In this paper, the results of the characterization of the sensors used to estimate the position of a smart robotic walker are briefly reported.