Publicação
Optical camera communications for platooning applications
| Resumo: | Platooning is a technology that corresponds to all the coordinated movements of a collection of vehicles, or, in the case of mobile robotics, to all the coordinated movements of a collection of mobile robots. It brings several advantages to driving, such as, improved safety, accurate speed control, lower CO2 emission rates, and higher energy efficiency. This dissertation describes the development of a laboratory scale demonstrator of platooning based on optical camera communications, using two generic wheel steered robots. For this purpose, one of the robots is equipped with a Light Emitting Diode (LED) matrix and the other with a camera. The LED matrix acts as an Optical Camera Communication (OCC) transmitter, providing status information of the robot attitude. The camera acts as both image acquisition and as an OCC receiver. The gathered information is processed using the algorithm You Only Look Once (YOLO) to infer the robot motion. The YOLO object detector continuously checks the movement of the robot in front. Performance evaluation of 5 different YOLO models (YOLOv3, YOLOv3-tiny, YOLOv4, YOLOv4-tiny, YOLOv4-tiny-3l) was conducted to assess which model works best for this project. The outcomes demonstrate that YOLOv4-tiny surpasses the other models in terms of timing, making it the ideal choice for real-time performance. Object detection using YOLOv4-tiny was performed on the computer. This was chosen since it has a processing speed of 3.09 fps as opposed to the Raspberry Pi’s 0.2 fps. |
|---|---|
| Autores principais: | Silva, Beatriz Dias |
| Assunto: | Platooning OCC YOLO Camera Object detection |
| Ano: | 2022 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Aveiro |
| Idioma: | inglês |
| Origem: | RIA - Repositório Institucional da Universidade de Aveiro |
| Resumo: | Platooning is a technology that corresponds to all the coordinated movements of a collection of vehicles, or, in the case of mobile robotics, to all the coordinated movements of a collection of mobile robots. It brings several advantages to driving, such as, improved safety, accurate speed control, lower CO2 emission rates, and higher energy efficiency. This dissertation describes the development of a laboratory scale demonstrator of platooning based on optical camera communications, using two generic wheel steered robots. For this purpose, one of the robots is equipped with a Light Emitting Diode (LED) matrix and the other with a camera. The LED matrix acts as an Optical Camera Communication (OCC) transmitter, providing status information of the robot attitude. The camera acts as both image acquisition and as an OCC receiver. The gathered information is processed using the algorithm You Only Look Once (YOLO) to infer the robot motion. The YOLO object detector continuously checks the movement of the robot in front. Performance evaluation of 5 different YOLO models (YOLOv3, YOLOv3-tiny, YOLOv4, YOLOv4-tiny, YOLOv4-tiny-3l) was conducted to assess which model works best for this project. The outcomes demonstrate that YOLOv4-tiny surpasses the other models in terms of timing, making it the ideal choice for real-time performance. Object detection using YOLOv4-tiny was performed on the computer. This was chosen since it has a processing speed of 3.09 fps as opposed to the Raspberry Pi’s 0.2 fps. |
|---|