DroneMap: A Cloud Robotics System for Unmanned Aerial Vehicles in Surveillance Applications
- Principal Investigator: Anis Koubaa
- Duration: 1 year
- Funding Agency: King Abdul-Aziz City for Science and Technology
- WP 1. Conceptual Design of the MyBot Robot Model
- WP2. Software Engineering of MyBot
- WP 3. Implementation and Development
- WP 4. Experimentation, testing and validation of MyBot
This research proposal is a research and development project that aims at the design and development of a cloud based architecture for Micro Unmanned Aerial Vehicles (UAVs). The core contribution of this project is the coupling between cloud computing and UAV technologies, where the cloud provides computing and storage resources for micro UAVs that collect sensory data from the environment and send it to the cloud for further processing and sharing. This coupling will allow the UAV to go beyond their limited processing capabilities and take profit from cloud computing resources.
In this project, we plan to design and implement a cloud robotics system for UAVs for surveillance applications. The cloud should be able to track a team of UAVs, receive sensory data, perform processing and control it by sending commands to the UAVs. Several research challenges arise from this coupling, but we will focus on (1) the real-time communication between the cloud and the UAV, (2) middleware design and implementation for interaction between the cloud the UAV. We will consider remote monitoring and surveillance as a target application. Furthermore, we aim at implementing a prototype that demonstrates the applicability of the prospective cloud system to promote such real-world applications.
The project encompasses both research and development (i.e. engineering) contributions. Research contributions consist of the design of a cloud robotics system for UAVs that allows storage, sharing and processing of information from different clients. The Cloud services provided by the system would allow users to easily access this information in a user friendly browser window, and allow the UAVs to (1) offload their intensive computation tasks to the cloud (2) share sensor data coordinate with other UAVs through the cloud to effectively accomplish their surveillance mission.
The engineering part consists in the implementation of a real-world prototype that demonstrates how the UAVs can benefit from the cloud to promote the effectiveness of the surveillance and monitoring application. We intend to use a middle ware framework such as Robot Operating System (ROS) middleware framework for robot programming as it provides software components that allow to easily building complex robotic applications. We will choose UAVs with 3G and/or WiFi connections for ensure long-range communication. We also intend to use Hadoop Distributed File System (HDFS) for data storage and Hadoop Map/Reduce framework for implementing the batch processing of sensor data and visual information. At the end of the 12-month period, a proof-of-concept will be demonstrated showing how the proposed platform allows micro UAVs to send and share sensor information from a disaster recovery site.