University of Thessaly (UTH)
Programming and testing support for drone based applications
Abstract
dc:descriptionThe evolution and the widespread adoption of low cost and versatile manufacturing technologies such as 3D printing along with the recent software and hardware advances in embedded computing, wireless communications and control systems have sparked a development boom in the field of unmanned vehicles (UVs) for civilian use. Nowadays, UVs which are commonly known as drones are widely available, come at low cost, and are capable of operating largely autonomously in aerial, ground and aquatic environments with limited human intervention. Besides their considerable and continuously improving mobility and navigation properties, drones can carry a diverse set of application-specific payloads, including various types of sensors/actuators as well as powerful computing and communication equipment.Thanks to these capabilities, drones will become key components of cyber-physical systems. They will revolutionize existing applications, and may very well give birth to entirely new ones. Recently, drone platforms have been adopted in several civilian application domains, spanning from smart agriculture and structural health monitoring to the entertainment industry. However, their full potential has not been unleashed yet. Despite the significant work that has been done by the scientific community over the last decades in various aspects of wireless sensor networks, robotics and distributed systems, application development for drones remains a hard task, especially when multiple drones are involved and need to perform a joint mission in a coordinated manner. In this thesis we focus on supporting mission program development for unmanned vehicles by reducing the amount and complexity of the programming that is required in order to deal with the problems of communication, coordination, heterogeneity and dynamics in multi-drone missions. We introduce a programming model that provides suitable coordination abstractions and mechanisms regarding resource heterogeneity and dynamics, allowing mission programs to utilize multiple drones in a flexible way. We adopt a centralized service-oriented approach whereby a distinguished entity which is called mission controller runs the mission program logic, which retrieves information from the drones and issues high-level control commands to them. In addition, our programming model provides unified abstractions for controlling individual drones as well as teams of drones with minimal development effort. The model along with its supportive runtime environment comes with an integrated fault tolerant mechanism, which combines checkpointing and logging in order to allow seamless recovery from mission controller failures.We support our programming model with an efficient coordinated transport protocol that performs reliably 1-N request/reply interactions within a group of processes. The protocol utilizes the broadcast nature of the wireless medium to avoid collisions, while allowing a high rate of interactions between the coordinator process (mission controller) and the ordinary processes (UVs) with low and predictable latency. It also comes with integrated fault handling and group management. The protocol is used as a transport for the commands sent by the coordinator process to the ordinary processes. The same protocol can be used to let the coordinator retrieve state/sensor updates from the ordinary processes via polling. We have also done work to support a dynamic adjustment of the polling rate, based on a method that automatically infers the actual data rate of the ordinary processes. Last but not least, we have developed a modular simulation infrastructure to perform tests with multiple virtual UVs (vUVs) in a flexible, controlled and safe way. Each vUV is a separate virtual machine, which runs the entire software stack including autopilot coupled to a physics model that simulates the mobility-related behavior of the UV. The system currently supports multi-copters with a virtual camera sensor, which can communicate with each other over a simulated WiFi network. We have successfully used the simulation environment to test various functional aspects of our system software stack, which would be practically impossible or prohibitively expensive to do using real drones.
Degree
thesis:*- Grantor dc:publisher
- University of Thessaly (UTH)
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Koutsoubelias, Manos
- Κουτσουμπέλιας, Εμμανουήλ
Subjects
dc:subject × 12- Προγραμματιστικά μοντέλα
- Συντονισμός σμήνους drones
- Ανοχή σφαλμάτων
- Προσομοίωση σμήνους drone
- Programming model
- Drone coordination
- Fault tolerant support
- Multi - drone simulation
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/44411