TY - JOUR
T1 - Lightweight obstacle detector based on scattered IR and Lock-In filtering
AU - Rosas-Flores, B.
AU - Hernández-Zavala, A.
AU - Huerta-Ruelas, J.
N1 - Publisher Copyright:
© 2019 Elsevier B.V.
PY - 2020/3
Y1 - 2020/3
N2 - Obstacle detection in autonomous navigation involves the processing of the environment information obtained from sensors, such as LIDAR, radar, ultrasonic, magnetic and capacitive. Imaging techniques like stereoscopy, optical flux, and deep image, have been used for precise object detection but with a high computational cost. These sensors also have disadvantages like weight, its size or power consumption. The case of systems with lasers added like in LIDAR can be harmful to people. Ultrasonic sensors are safer for people but the signal to noise ratio decreases in outdoors. IR sensors are cheap and can be clustered to strength signal, however, additional sources of radiation can affect the measurement especially in outdoors. In this work, we propose an IR system with adequate spatial configuration and electronic filters to detect objects indoors and outdoors by means of scattered IR radiation centered at 940 nm. The developed sensor has the following specs: Front area dimensions is 5x5 cm that includes 10 IR LEDs at the center and a Si detector in each corner. The approximated detection volume is 17,000 cm3, considering a detection cone length of 300 cm and a solid angle of 30 degrees. Total system weight is 40 g. System was tested with four experiments to evaluate signal filtering, effect of object surface roughness, repeatability and working sensitivity region. This evaluation allowed to make a comparison with traditional Obstacle Detection System techniques, giving advantages in almost all parameters considered, allowing us to conclude that developed system can detect objects in indoors and outdoors, in a region suitable for aerial and surface vehicles or blind persons.
AB - Obstacle detection in autonomous navigation involves the processing of the environment information obtained from sensors, such as LIDAR, radar, ultrasonic, magnetic and capacitive. Imaging techniques like stereoscopy, optical flux, and deep image, have been used for precise object detection but with a high computational cost. These sensors also have disadvantages like weight, its size or power consumption. The case of systems with lasers added like in LIDAR can be harmful to people. Ultrasonic sensors are safer for people but the signal to noise ratio decreases in outdoors. IR sensors are cheap and can be clustered to strength signal, however, additional sources of radiation can affect the measurement especially in outdoors. In this work, we propose an IR system with adequate spatial configuration and electronic filters to detect objects indoors and outdoors by means of scattered IR radiation centered at 940 nm. The developed sensor has the following specs: Front area dimensions is 5x5 cm that includes 10 IR LEDs at the center and a Si detector in each corner. The approximated detection volume is 17,000 cm3, considering a detection cone length of 300 cm and a solid angle of 30 degrees. Total system weight is 40 g. System was tested with four experiments to evaluate signal filtering, effect of object surface roughness, repeatability and working sensitivity region. This evaluation allowed to make a comparison with traditional Obstacle Detection System techniques, giving advantages in almost all parameters considered, allowing us to conclude that developed system can detect objects in indoors and outdoors, in a region suitable for aerial and surface vehicles or blind persons.
KW - Avoidance sensor
KW - Lock-In filter
KW - NIR scattering
KW - Obstacle detection
KW - Unmanned vehicle
UR - http://www.scopus.com/inward/record.url?scp=85081139597&partnerID=8YFLogxK
U2 - 10.1016/j.infrared.2019.103157
DO - 10.1016/j.infrared.2019.103157
M3 - Artículo
AN - SCOPUS:85081139597
SN - 1350-4495
VL - 105
JO - Infrared Physics and Technology
JF - Infrared Physics and Technology
M1 - 103157
ER -