Formal methods for autonomous systems
Docent lecture by Jana Tumova, RPL
Time: Wed 2020-10-14 10.30
Location: Fantum and Zoom
Participating: Docent lecture by Jana Tumova
Autonomous robots permeated a variety of application areas from industrial automation to transport to household services. Robot vacuum cleaners clean our homes and mobile robotic solutions have been deployed in warehouses to carry shelves of goods. Gradually, autonomous mobile robots are moving from enclosed environment to the wild. Visions for future include driverless vehicles seamlessly cruising streets, drones delivering goods to our doors, or robot butlers delivering room service in hotels. Such robots will need to interact with each other to coordinate on safe, effective, and efficient execution of their tasks. Even more importantly, they will inevitably interact with people, if not directly, then by co-existing in the same environment. This brings a particular importance to the question of making sure that autonomous systems execute complex functionalities as intended.
Recently popular formal methods-based approaches to task planning, motion planning, and control combine two great benefits by allowing i) to specify complex tasks and constraints in a rigorous way and ii) to deploy automated formal synthesis techniques to obtain correct-by-construction plans for a system model. In this lecture, we will take a peek into formal methods for autonomous systems, their main principles, and several applications.