ROS-based Architecture for Autonomous Navigation of an Open-Source Electric Car
Luis M. Bergasa, Rafael Barea, Elena López-Guillén, Pedro Revenga, Eduardo Molinos, Eduardo Romera, Álvaro Saez, Carlos G. Huélamo, Felipe Arango, Miguel Tradacete, Javier Araluce, Javier del Eguido, Esther Murciego, Rafael Sanz, Joaquín López, Enrique Paz, Cándido Otero, Pablo Sánchez

Description of the technology
The software architecture is based on the open-source Robot Operating System (ROS). Our navigation proposal is divided into executive control and lower-level reactive control. The executive layer calculates a path consisting of a sequence of lanelets that can be modified depending on the performed behaviors. The sequence of actions and events that take part of a specific behavior are obtained through a Petri Netimplemented by the RoboGraphtool. The goal of the local navigation is keeping the car within the driving lane in a safety way and avoiding ahead obstacles using the Pure Pursuitsand Beam Curvaturemethods. Environment perception is based on 3D semantic segmentation obtained from our ERFNet.

Experimental set up & Demonstration
The experimental platform is an open-source electric car, based on a TABBY EVO chassis developed by the Open Motors company. The chassis has been equipped with a pack of batteries, a tubular car body, some sensors (RTK-DGPS, LiDAR, stereo-camera and odometry) and it has been modified to be autonomous. The demonstration consists on: 1) navigating autonomously the car along a given path defined by lanelets, while avoiding static and dynamic obstacles detected by the onboard sensors, 2) switching between autonomous/manual mode when the driver touch the steering wheel, the brake pedal or the throttle pedal.







[1] Rafael Barea, Carlos Pérez de Rivas, Luis M. Bergasa, Elena López-Guillén, Eduardo Romera, Eduardo Molinos, Manuel Ocaña, Joaquín López, “Vehicle Detection and Localization using 3D LIDAR Point Cloud and Image Semantic Segmentation”, in IEEE Conference on Intelligent Transportation Systems (ITSC), Maui, Hawaii, USA, November 2018. Accepted.
[2] E. Romera, J.M. Álvarez, L.M. Bergasa, R. Arroyo, “ERFNet: Efficient Residual Factorized ConvNet for Real-time Semantic Segmentation”, in IEEE Transactions on Intelligent Transportation Systems. Vol 19, Issue 1, 263-272, January 2018.[PDF]
[3] L.M. Bergasa, R. Sanz, J. López, E. Paz, C. Otero, P. Sánchez, R. Barea, E. López-Guillén, P. Revenga, E. Molinos, E. Romera, “Desarrollo de un vehículo eléctrico autónomo de código abierto para personas mayores”, in Actas de las Jornadas Nacionales de Robótica 2018, Valladolid, Spain, June 2018. [PDF]
[4] Cándido Otero, Enrique Paz, Jesús López, Rafael Barea, Eduardo Romera, Eduardo Molinos, Roberto Arroyo, Luis Miguel Bergasa, Elena López, “Simulación de vehículos autónomos usando V-Rep bajo ROS”, in Actas de las XXXVIII Jornadas de Automática, Gijón, Spain, September 2017. [PDF]
[5] Fernández, J. L., Sanz, R., Paz, E., & Alonso, C. (2008, May). Using hierarchical binary Petri nets to build robust mobile robot applications: RoboGraph. In Robotics and Automation, 2008. ICRA 2008. IEEE International Conference on (pp. 1372-1377). IEEE.[PDF]