Software Engineering for Robotics - The RoboStar Technology
1 like369 views
The document describes RoboStar Technology's approach to software engineering for robotics. It involves modeling robot behavior and physical systems, automatically generating simulations from models, and testing and verifying properties through model checking and simulation. The goal is to produce certified control software for safe robotic systems through this model-driven development approach.
Software Engineering for Robotics - The RoboStar Technology
- 1. RoboStar Technology Software Engineering for Robotics Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) November, 2019 Thanks: Alvaro Miyazawa, Pedro Ribeiro, Augusto Sampaio, Jon Timmis, Jim Woodcock Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 1 / 37
- 2. Current approach to development in robotics 1st phase: Abstract model state machine 1st phase: Abstract model state machine controller code hardware simulation discrete environment simulation 2nd phase: Simulation controller code hardware simulation discrete environment simulation 2nd phase: Simulation low-level code robot environment 3rd phase: Implementation Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 2 / 37
- 3. Example state machines Figure 1: Complete state diagram illustrating the operation of the Relay Chain algorithm. The three core states are chain AUVs in the Relay Chain, explore AUVs in the exploratory shoal and navigate AUVs moving between chain and explore. The initialisation flow chart is shown in (a) . end AUVs are special cases. As discussed in the initialis- ing section the start AUV remains at the water’s surface and initialises chain formation. The end AUV is both a member of them in the chain. They aim to stay in a position halfway between the AUVs on either side of them. If the halfway distance exceeds a ‘stress threshold’ then it will stay nearer Becky Naylor, Mark Read, Jon Timmis, and Andy Tyrrell. The Relay Chain: Communication between an Exploratory Underwater Shoal and a Surface Vehicle. ALIFE 14: Proceedings of the 14th International Conference on the Synthesis and Simulation of Living Systems. Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 3 / 37
- 4. Example state machines “A group of e-puck robots transporting an object (blue box) towards a goal (red cylinder).” Jianing Chen, M. Gauci and R. Gross. “A strategy for transporting tall objects with a swarm of miniature mobile robots”. In: Robotics and Automation (ICRA), 2013 IEEE International Conference on. 2013, pp. 863–869. Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 4 / 37
- 5. Our approach RoboChart + RoboWorld Model Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 5 / 37
- 6. Our approach AUTOMATIC GENERATION Proof ModelRoboChart+ RoboWorld Model Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 6 / 37
- 7. Our approach AUTOMATIC GENERATION Proof ModelRoboChart+ RoboWorld Model Model checkers Theorem provers Properties of interest valid? Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 7 / 37
- 8. Our approach AUTOMATIC GENERATION Proof ModelRoboChart + RoboWorld Model Model checkers Theorem provers Properties of interest valid? NO Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 8 / 37
- 9. Our approach AUTOMATIC GENERATION Proof ModelRoboChart + RoboWorld Model Model checkers Theorem provers Properties of interest valid? NO RoboSim Model AUTOMATIC GENERATION YES Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 9 / 37
- 10. Our approach AUTOMATIC GENERATION Proof ModelRoboChart + RoboWorld Model Model checkers Theorem provers Properties of interest valid? NO RoboSim Model YES Robotics Simulator valid? Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 10 / 37
- 11. Our approach AUTOMATIC GENERATION Proof ModelRoboChart + RoboWorld Model Model checkers Theorem provers Properties of interest valid? NO RoboSim Model AUTOMATIC GENERATION YES Robotics Simulator valid? NO Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 11 / 37
- 12. Our approach AUTOMATIC GENERATION Proof ModelRoboChart + RoboWorld Model Model checkers Theorem provers Properties of interest valid? NO RoboSim Model AUTOMATIC GENERATION YES Robotics Simulator valid? NOYESDeployment Code AUTOMATIC GENERATION Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 12 / 37
- 13. Our approach AUTOMATIC GENERATION Proof ModelRoboChart + RoboWorld Model Model checkers Theorem provers Properties of interest valid? NO RoboSim Model AUTOMATIC GENERATION YES Robotics Simulator valid? NOYES Test cases CONVERSION Test cases Test cases Deployment Code AUTOMATIC GENERATION System Testing correct? Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 13 / 37
- 14. Our approach AUTOMATIC GENERATION Proof ModelRoboChart + RoboWorld Model Model checkers Theorem provers Properties of interest valid? NO RoboSim Model AUTOMATIC GENERATION YES Robotics Simulator valid? NOYES Test cases CONVERSION Test cases Test cases Deployment Code AUTOMATIC GENERATION System Testing correct? NO Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 14 / 37
- 15. Our approach AUTOMATIC GENERATION Proof ModelRoboWorld+RoboChart Model Model checkers Theorem provers Properties of interest valid? NO RoboSim Model AUTOMATIC GENERATION YES Robotics Simulator valid? NOYES Test cases CONVERSION Test cases Test cases Deployment Code AUTOMATIC GENERATION System Testing correct? NOYES Code Proofs Test results and assumptions Environment restrictions CERTIFIED FOR SAFE USE Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 15 / 37
- 16. Our approach Control Software Platform Services Environment Assumptions RoboChart + RoboWorld Control Software Platform Physical Model Scenario RoboSim Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 16 / 37
- 17. Our approach Control Software Platform Services Environment Assumptions RoboChart + RoboWorld Control Software Platform Physical Model Scenario RoboSim Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 17 / 37
- 18. Our approach Control Software Platform Services Environment Assumptions RoboChart + RoboWorld Control Software Platform Physical Model Scenario RoboSim Platform Services realisation abstraction Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 18 / 37
- 19. Our approach Control Software Platform Services Environment Assumptions RoboChart + RoboWorld Control Software Platform Physical Model Scenario RoboSim Platform Services realisation abstraction Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 19 / 37
- 20. A very small RoboSim example - d-model Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 20 / 37
- 21. Behaviour for RoboSim: cycle simulation while true read sensors process data write to actuators cycle period { ... Buffered parallelism Visible behaviour: registerRead and registerWrite Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 21 / 37
- 22. Behaviour for RoboSim: cycle simulation while true read sensors process data write to actuators cycle period { ... Buffered parallelism Visible behaviour: registerRead and registerWrite ? Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 21 / 37
- 23. p-model: physical model Inspired by facilities offered by simulators Basic notions: links, joints, sensors, actuators Sensors, actuators, joints: system of differential algebraic equations Diagrammatic notation: block diagram Linked to a module: platform mapping Sensors and actuators → Variables, events, and operations Automatic generation XML-based SDF for simulation CSP-based model for reasoning Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 22 / 37
- 24. p-model: physical model Advantages over XML-based notations Diagrammatic notation Customised editor Well-formedness constraints Simple and clear mechanism to define poses (position and orientation) Indexing for collections of sensors and actuators Structuring mechanisms Extensibility Modelling of behaviour: differential algebraic equations Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 23 / 37
- 25. A very small example - Not the footbot Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 24 / 37
- 26. A very small example - Not the footbot link link body body joint Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 25 / 37
- 27. A very small example - Not the footbot Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 26 / 37
- 28. Joints - a small example Part of a library Behaviour: DAE systems Relate inputs and outputs May involve local variables and constants Constants can be defined at instantiation Sensors and actuators can be defined in a similar way. Equations can be obtained by curve fitting based on experiments. Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 27 / 37
- 29. Actuator - a small example Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 28 / 37
- 30. Parts - a small example Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 29 / 37
- 31. Parts - a small example Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 30 / 37
- 32. Parts - a small example Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 31 / 37
- 33. Platform mappings Association between d-model and p-model Defines variables, events, and operations Variable: input to actuators Input events: predicate and communicated values, if any Output events: assignment to inputs of actuators Operations: actions or DAE systems Very simple definitions: no time Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 32 / 37
- 34. Platform mapping - a small example Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 33 / 37
- 35. Automatic generation of simulations Translation to SDF Automatic Useful tool for validation Various physics engines Equations are (mostly) ignored Proof of properties is possible Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 34 / 37
- 36. RoboTool support for our approach Eclipse plug-ins: www.cs.york.ac.uk/robostar/ RoboChart Models Requirements ARGoS RoboTool C++ PRISM Storm Reactive Modules Formalism CSP and timed-CSP Qualitative Results Simulation Quantitative Results Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 35 / 37
- 37. So, what next? RoboSim Implementation Refinement technique Soundness Case studies: driverless pod, sandwich maker, ... Support to user Properties language Modelling of environment Test generation Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 36 / 37
- 38. RoboSoft - A Royal Academy of Engineering Event www.cs.york.ac.uk/robostar/robosoft 13-14 November, 2019 London, UK Some of the invited speakers Robin Bloomfield, UK Davide Brugali, Italy Kerstin Eder, Bristol Robotics Lab Gusz Eiben, The Netherlands Michael Fisher, Liverpool Arnaud Gotlieb, Norway Felix Ingrand, France Patrizio Pelliccione, Sweden Christian Schlegel, Germany Alan Winfield, Bristol Robotics Lab Ana Cavalcanti University of York, UK - RoboStar (www.cs.york.ac.uk/RoboStar) RoboStar Technology – Software Engineering for Robotics 37 / 37