Vehicle automation software development using software-only simulation
Ruusiala, Arttu (2022)
2022
Tietotekniikan DI-ohjelma - Master's Programme in Information Technology
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2022-11-04
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-202210107532
https://urn.fi/URN:NBN:fi:tuni-202210107532
Tiivistelmä
Automatic driving and driver assistance systems are gaining attraction in the automotive industry. Their development is not an easy task and requires enormous amounts of testing and validation. However, conducting all testing with a real car is expensive and ineffcient. A possible solution to streamline testing is simulation, especially software-only simulation. In software-only simulation, everything is simulated using just software. It does not require any specialized hard ware making it cheaper and easier to establish and scale up the number of testing environments.
The goal of this thesis was to study how a software-only simulation environment could be built using readily available open-source components. A simulator environment based on an open-source driving simulator, CARLA, was built, and an example application was developed and integrated into it using Robot Operating System 2 (ROS2). The example application, Carlabot, supports manual driving with a gamepad and utilizes a LiDAR sensor to implement a simple collision avoider, which slows down or stops the car if something is detected in front of the car.
The process of setting up a CARLA simulator environment using predefned assets, such as vehicle and world model, proved to be straightforward, and integrating a simple example application was fairly uncomplicated. However, using the environment for real product development would require customizing at least the assets.
Software-only simulation brings benefts to the software development of automatic vehicles. It allows testing on a scale that is not viable using just real hardware, and it enables using test automation already in integration testing. Software-only simulation supports agile software development, where testing begins early, already during the development.
The goal of this thesis was to study how a software-only simulation environment could be built using readily available open-source components. A simulator environment based on an open-source driving simulator, CARLA, was built, and an example application was developed and integrated into it using Robot Operating System 2 (ROS2). The example application, Carlabot, supports manual driving with a gamepad and utilizes a LiDAR sensor to implement a simple collision avoider, which slows down or stops the car if something is detected in front of the car.
The process of setting up a CARLA simulator environment using predefned assets, such as vehicle and world model, proved to be straightforward, and integrating a simple example application was fairly uncomplicated. However, using the environment for real product development would require customizing at least the assets.
Software-only simulation brings benefts to the software development of automatic vehicles. It allows testing on a scale that is not viable using just real hardware, and it enables using test automation already in integration testing. Software-only simulation supports agile software development, where testing begins early, already during the development.