Accurate and precise distance estimation for noisy IR sensor readings contaminated by outliers

Abstract

Infrared(IR) sensors are widely used to estimate distance due to their practicality and cost-effectiveness. The distance is estimated by measuring the reflectance amplitude of IR light from a targeted object. However, the non-linearity characteristic of the reflected IR light makes distance estimation difficult. Additionally, IR sensors have low accuracy and precision. Some researches superficially studied distance estimation using IR sensors. However, a deeper study is needed on this subject as competitive market conditions push companies to produce high-quality devices without increasing the production cost. Therefore, this paper proposed a procedure on improving the accuracy and precision of distance estimation of a low-cost, but noisy IR sensor. The proposed procedure contains six complementary steps. Experimental and simulation validation is performed to prove the effectiveness of the proposed procedure. In the validation, the distance estimation error never exceeds +/- 2 mm between 10 cm and 400 cm. (C) 2020 Elsevier Ltd. All rights reserved.