CoFLY

The use of Micro Aerial Vehicles (MAV) has significantly increased during the past few years where operators could be from simple users to advanced administrators. Commercially available MAVs were exploited in a wide variety of applications including entertainment and surveillance. Nonetheless, much effort has been made by both the market and the research community to introduce specific advances in such systems regarding their hardware components, mostly. For entertainment usage, the vehicles are equipped with low cost features so that they could be affordable to a common user. On the contrary, specialized sensors and complex manipulations are required to benefit from their exploitation in business applications increasing the overall operational cost.

In both cases innovative advances are required in order to transform simple MAVs from “hobbyists’ toys” to business essentials and mitigate restrictions inserted by complex scenarios.

Description of the solution

The main objective of the CoFly is to insert pioneering functionalities in a MAV system aiming at reducing the operational cost as well as provide intelligent modules assisting non-expert users in proper manipulation. An “average” business operator should be able to exploit the robotic advances inserted with the CoFly in order to accomplish complex tasks accurately and without the prerequisite of hiring a highly skilled MAV operator. The proposed architecture aims at enhancing the software-dependent system abilities in various robot-related research areas while the hardware requirements are kept at the most significant so that the cost could be kept at low levels.

More specific, innovations will be developed for every operational module so that a fully functional autonomous MAV system (while keeping the user in the loop) with enhanced capabilities could be demonstrated. The operational MAV will be equipped with low cost sensors based on which the robotic technologies can be deployed. The navigational module will be enhanced by utilizing semantic representations of the MAV’s environment through the corresponding algorithm so that adaptability between the defined tasks could be accomplished. The system will manage in real-time the required objectives (defined either by the decision support module or the operator) in a way that smoother transitions and manipulations can be achieved. This enhanced capability prerequisites the extraction of a semantic representation of the robot’s operational area meaning that perception and cognition abilities must be deployed. Localization and mapping, semantic representation and object recognition techniques will enhance the corresponding cognition capabilities in order to improve the accuracy of both the human-robot and robot-environment interactions. All services and robotic software abilities of the proposed framework will operate on mobile devices to support maximum mobility. Thus, an intuitive graphical user interface will be developed including higher level user commands and proper data representations depicting the results of the MAV’s operation.

The expected impact of the developed modules includes a verifiable increase in the level of the robotic system abilities in business related applications as well as significant improvements in the corresponding technologies. Therefore, the system will be thoroughly tested in agricultural applications in order to ensure accurate operations in complex environments with high diversity. The operator/farmer will be provided with a system which will be able to accomplish tasks previously completed manually (e.g. crop growth monitoring) and play a supportive role in decision making objectives (e.g. possible insect infestations).