SERVANT-ES
SERVANT-ES: Experimental Unmanned Aerial Vehicle Research System
Unmanned Aerial Vehicles (UAVs) and Remotely Piloted Aircrafts (RPAs) are becoming more important every day. Unfortunately, the design of these systems is very complex and prone to errors. Some projections suggest that the volume of the UAV global market will exceed USD$ 21 billion in 2022. These figures help to justify the interest of research institutions and of start-ups for the development and applications of UAVs. There are some problems to overcome though:
-
Although the variety of UAV models available on the market is constantly growing, the costs of professional-level UAV may be prohibitive for low-capital research institutions and/or start-ups;
-
So-called avionics, i.e., flight control software for electronic aircraft, has a high development cost due to a large number of potential errors, and the difficulty in finding these errors and correcting them.
Avionics software currently accounts for most of the costs of developing new aircraft. Traditional methods of analysis and validation include tests and simulations, but the failure coverage of these methods is very low because the number of possible behaviors in a system increases exponentially with the size of the system.
An alternative to these traditional validation methods is the use of formal methods of software engineering, which have the characteristic of exhaustively exploring all possible executions of a system being evaluated. In this way, the results obtained are guaranteed, unlike tests and simulations, which explore only a portion of the possible behaviors. The use of formal methods is encouraged by international avionics certification standards such as DO-178C and DO-333 published by RTCA.
The SERVANT-ES project aims to develop an open-source software platform for research with UAVs and RPAs using modern engineering techniques based on the application of formal methods.
This project is being founded by FAPES – Fundação de Amparo à Pesquisa do Espírito Santo.