System Based on Compact mmWave Radar and Natural Body Movement for Assisting Visually Impaired People

The aim of this paper is to show how compact mmWave radar systems, attached to a person, can take advantage of their movements, in order to significantly improve the image resolution by means of synthetic aperture radar (SAR) techniques. Thus, high resolution imaging becomes possible even with tiny on-chip systems, opening up a horizon of possibilities to assist visually impaired people. The approach is illustrated for movements in terms of arm swings. For this purpose, it is firstly shown that the information from an inertial measurement unit (IMU) can be used to provide a trajectory estimation, which is enough to apply synthetic aperture radar processing. As a consequence, the resolution of the radar is greatly improved when compared with that achieved considering a single position. In order to avoid the continuously growing position error, commonly observed in dead-reckoning systems relying just on IMU, zero velocity updates (ZUPT) will be systematically applied. For doing so, two different zero velocity detectors (ZVD) are considered, one based on acceleration data and other on velocity estimation. Although the positioning information is extracted from the IMU, demonstrating the enhanced obstacle detection capabilities obtained with this approach, other sensors that can be carried by a person and that are capable of providing trajectory estimations to combine measurements acquired with compact mmWave radars could have been used. The system is tested using different targets and a good performance comparable to the one achieved by using a high accuracy reference positioning system is attained. The compactness of the system enables the possibility of using it as a wearable device, e.g., attached to the wrist. This shows that an IMU is enough for tracking relatively simple movements, though more complex positioning systems can be used for more general actions.