Modelling, Analysis, Simulation and Implementation of Robotic Applications

Identification

Title: Modelling, Analysis, Simulation and Implementation of Robotic Applications

Coordinator: Augusto Sampaio

Contact: acas@cin.ufpe.br

 

Summary

This project aims at creating a systematic and rigorous methodology to specify, verify, design and implement robotic applications. The idea is to build on the RoboCalc project (https://www.cs.york.ac.uk/circus/RoboCalc/), already under development at York and led by Ana Cavalcanti, Jin Timmis, and Jim Woodcock. Currently, the RoboCalc project offers the following facilities:

  • A state machine based notation (RoboChart) to model individual robots;
  • A translation from RoboChart to the process algebra CSP to use the FDR tool in background for the purpose of functional analyses (both of classical properties as deadlock and livelock, as well as doman specific properties); and
  • RoboTool, implemented as an Eclipse plugin, to support the model edition and analysis as described in the previous items.

The aim is to extend this scope with facilities for simulation and implementation of robot applications. We also plan to evolve RoboChart with features to specify swarm robots. Finally, we will also consider modelling of the environment and develop three case studies.

 

Goals and Activities

The central objective of this project is to contribute with a systematic and rigorous methodology to specify, verify, design and implement robotic applications. The proposal is to build on the RoboCalc project (https://www.cs.york.ac.uk/circus/RoboCalc/), already under development at York. Currently, the RoboCalc project offers the following facilities:

  • A state machine based notation (RoboChart) to model individual robots;
  • A translation from RoboChart to the process algebra CSP, for the purpose of analysis (both of classical properties as deadlock and livelock, as well as domain specific properties) using the FDR tool in background; and
  • RoboTool, implemented as an Eclipse plugin, to support the model edition and analysis as described in the previous items.

The objective is to extend this framework in several directions. These are captured by the following activities.

A1. Extension of the RoboChart notation to allow the specification of swarm robots, involving collections of (possibly heterogeneous) robots;

A2.  Extension of the RoboChart notation with probabilistic transitions;

A3. Definition of a simulation language;

A4. Translation from RoboChart (with the above extensions) to a simulation notation;

A5. Translation from the simulation notation to the language of a target platform (possibly C++);

A6. Definition of a controlled natural language (CNL) to model the environment;

A7. Translation from CNL to RoboChart;

A8. Extension of the RoboTool to incorporate theabovefacilities;

A9. Development of case studies (from an initial model, through simulation to a concrete implementation) that illustrate the new features, as listed below;

A10. Integration with the platform FIWARE.

 

Some practical applications

The group led by Jones Albuquerque has contributed with their expertise on several applications involving drones. Prototypes of the “HealthDrones” software platform are described here: https://bioandhealthdrones.wordpress.com/.

An example of the simulation environment for real applications can be viewed here: https://www.youtube.com/watch?v=N9z9pCL_Syk

HealthDrones provide an overview and integration of the software platforms available today for development of algorithms for unmanned aerial vehicles (UAVs) in autonomous flight to be used in IOT applications. We are primarily interested in their use for epidemiological monitoring of areas under risk, but these platforms can be used to monitor any kind of crowd event. We develop prototypes for applications in environment monitoring, biosecurity, illicit drugs, and malacological studies. In particular, in this project, images were captured and processed using NOAA and Meteosat 8 Satellites. They are operated by GOESERE-UFRPE (Laboratory of GIS and Remote Sensing of the Federal Rural University of Pernambuco and were used to compose the monitoring scenarios.  Images and other data are also captured and processed by the drones platforms and its own sensors. This information will compound automatic surveillance scenarios of the observed phenomena. These data can be used for monitoring environments at risk or under specific control (epidemiological or any crowd phenomena).

 

Research Team

Name Highest degree Institution
Adolfo Duran PhD UFBA
Alexandre Mota PhD UFPE
Álvaro Miyazawa PhD York
Ana Cavalcanti PhD York
Augusto Sampaio PhD UFPE
Daniel Lopez-Codina PhD UPC
Gustavo Carvalho PhD UFPE
Jim Woodcock PhD York
Jon Timmis PhD York
Jones Albuquerque PhD UFRPE
Juliano Iyoda PhD UFPE
Madiel Souza Conserva Filho PhD UFPE
Marcel Vinicius Medeiros Oliveira PhD UFRN
Marcio Cornélio PhD UFPE
Pedro Ribeiro PhD York
Sidney Nogueira PhD UFRPE
Silvana Bocanegra PhD UFRPE