Analyzing the Effect of Deceiving Agents in a System of Self-Driving Cars at an intersection - a computational model
The creation of protocols for autonomous intersection management is an active research topic with the potential of increasing the capacity of intersections addressing the increasing demand on roads. Most of the proposed protocols assume that all the vehicles involved will behave pro-socially, that i...
- Autores:
-
Morales Chavarro, Javier Mauricio
- Tipo de recurso:
- Fecha de publicación:
- 2021
- Institución:
- Universidad Nacional de Colombia
- Repositorio:
- Universidad Nacional de Colombia
- Idioma:
- eng
- OAI Identifier:
- oai:repositorio.unal.edu.co:unal/79380
- Palabra clave:
- 000 - Ciencias de la computación, información y obras generales::003 - Sistemas
Autonomous vehicles
Internet of vehicles (IoV)
deceiving agents
Traffic model
Autonomous intersection
Vehículos autónomos
Internet de los vehículos
Agentes engañosos
Modelo de tráfico
Intersección autónoma
Inteligencia artificial
Artificial intelligence
Programación informática
Computer programming
Automatización
Automation
- Rights
- openAccess
- License
- Atribución-NoComercial-CompartirIgual 4.0 Internacional
| Summary: | The creation of protocols for autonomous intersection management is an active research topic with the potential of increasing the capacity of intersections addressing the increasing demand on roads. Most of the proposed protocols assume that all the vehicles involved will behave pro-socially, that is, in a way that improves the outcome of the system over their individual gain. We simulated three different autonomous intersection protocols, two centralized and one decentralized, introducing some egoistic agents that we call deceiving vehicles. Deceiving vehicles may decide to transmit false information while using a protocol if they detect that doing so can result in a lower delay in the intersection. Our simulations show that in two of the protocols, it is possible for a deceiving vehicle to experience lower delay times compared to its non-deceiving counterparts. Additionally, as more deceiving vehicles enter the system the overall capacity of an intersection can be reduced, increasing delays for non-deceiving vehicles which creates an incentive for more vehicles to deceive. We pose that, given that vehicles have an incentive to deceive, autonomous intersection protocol's authors need to consider deceiving vehicles in their design and include measures to prevent them, thus avoiding the performance degradation they produce. |
|---|