| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 8.67 MB | Adobe PDF |
Authors
Abstract(s)
Dada a importância de melhorar a segurança rodoviária, é extremamente importante a criação de mecanismos que auxiliem a redução do número de acidentes rodoviários. Posto isto, esta tese visa o desenvolvimento de um algoritmo para alerta de colisões veiculares, capaz de apoiar o condutor na sua tarefa. O algoritmo proposta, inicialmente utiliza ITS-G5 como canal de comunicação entre veículos. No entanto, dada a
necessidade de disponibilizar a aplicação à maioria dos utilizadores, apresenta-se também uma solução para smartphone que permite a utilização da mesma a utilizadores que não tenham acesso a uma OBU ITS-G5.
Neste trabalho, o algoritmo de previsão de colisão foi definido e implementado e, para dar suporte à aplicação móvel, foi desenvolvido um ambiente híbrido entre ITS-G5 e redes celulares. A troca de informação é feita utilizando mensagens CAM. Dado a importância da latência no desempenho da aplicação, foram realizadas medições de latência para validar a solução proposta. Em geral, os melhores resultados foram obtidos aquando da utilização do 4G, obtendo uma latência média de 42 ms. O uso da rede 5G não alcançou as expectativas, pois esperavam-se menores latências que o 4G, mas apenas se verificaram latências similares e com maior instabilidade. Ao testar o algoritmo desenvolvido, foi utilizado um Digital Twin para auxiliar os
testes à solução permitindo a criação de situações específicas de alto risco. Posto isto, a sua utilização foi útil para testar uma situação constante de mensagens recebidas com latência elevada originando uma situação em que a aplicação assume um cenário distinto da realidade que impede a deteção da colisão, quando a mesma existe (falso negativo).
Given the importance of improving road safety, it is crucial to create mechanisms that reduce road accidents. Therefore, this thesis aims the development of a collision warning algorithm capable of supporting the drivers’ tasks by alerting them to possible collisions. The algorithm proposed, initially uses ITS-G5 as a communication channel between vehicles. However, given the need to provide the application to most users, a smartphone application is also presented, making this solution accessible to users that don’t have access to an ITS-G5 OBU. In this work, a collision warning algorithm was defined and implemented and, to support the smartphone application, a hybrid environment between ITS-G5 and cellular network was developed. Information is exchanged among players using CAM messages. Given the role that latency plays in the performance of this kind of application, latency measurements were performed to validate the proposed framework. In general, the best results were obtained when using 4G, obtaining an average of 42 ms in latency. The usage of the 5G network didn’t meet the expectations since it promises lower latencies than 4G, but similar latencies and greater instability have been observed at this stage. When testing the developed application, a digital twin was used to assist the algorithm tests allowing the creation of specific dangerous situations. Therefore, the digital twin was used to test a specific scenario of constant high latency messages creating a false negative situation, where a collision would be detected if latency was lower/admissible.
Given the importance of improving road safety, it is crucial to create mechanisms that reduce road accidents. Therefore, this thesis aims the development of a collision warning algorithm capable of supporting the drivers’ tasks by alerting them to possible collisions. The algorithm proposed, initially uses ITS-G5 as a communication channel between vehicles. However, given the need to provide the application to most users, a smartphone application is also presented, making this solution accessible to users that don’t have access to an ITS-G5 OBU. In this work, a collision warning algorithm was defined and implemented and, to support the smartphone application, a hybrid environment between ITS-G5 and cellular network was developed. Information is exchanged among players using CAM messages. Given the role that latency plays in the performance of this kind of application, latency measurements were performed to validate the proposed framework. In general, the best results were obtained when using 4G, obtaining an average of 42 ms in latency. The usage of the 5G network didn’t meet the expectations since it promises lower latencies than 4G, but similar latencies and greater instability have been observed at this stage. When testing the developed application, a digital twin was used to assist the algorithm tests allowing the creation of specific dangerous situations. Therefore, the digital twin was used to test a specific scenario of constant high latency messages creating a false negative situation, where a collision would be detected if latency was lower/admissible.
Description
Dissertação para obtenção do Grau de Mestre em Engenharia Informática e de Computadores
Keywords
V2X ITS-G5 Collision warning Veículos conectados CAM ITS Ambiente híbrido Redes celulares OBU Collision avoidance algorithm V2X ITS-G5 Collision warning Connected vehicles CAM Hybrid environment Cellular network OBU Collision avoidance algorithm
Citation
REIS, José Diogo Salgado dos - Collision avoidance algorithm for intelligent vehicles using ITS-G5. Lisboa: Instituto Superior de engenharia de Lisboa, 2022. Dissertação de Mestrado