CPSE Labs

Contact us about robotic software architectures, safety assessments, artificial intelligence, dynamic reconfiguration and single-chip cloud computing.

Our French design centre aims to make dependability affordable for companies that are developing autonomous vehicles, ranging from car with advanced driver systems to ground mobile robot or unmanned aerial vehicle. This is a pre-requisite to prove that such vehicles can be operated safely and can be more widely deployed.

Autonomous vehicles range from cars that integrate advanced driver systems (ADAS) to remotely piloted aircrafts fullling various missions (e.g. agriculture, rescue, surveillance). They often interact with a control station according to predefined procedures; they also have to deal with the environment uncertainties (e.g. because of human presence). One obstacle to their development for a wider application is the lack of guidance about the technologies and the design methods which ensure that they can be operated with an acceptable level of risks. There is also a lack of safety culture (especially for robotics SME), and the experiments will provide an excellent means to train new dependability comers and to help the formalisation of new needs.

However, the deployment of the dependability and safety culture is slowed by several factors such as the versatility of the concepts of operation and the variety of applicable regulations and standards, the robustness of complex technologies or the cost of system validation. CPSE Labs France will:

• Assist companies to clarify their dependability needs according to their concept of vehicles and operation;
• Provide a set of methods, tools and technologies to address efficiently dependability issues specifically raised by autonomous vehicles; and
• Teach companies the most appropriate safety practices.

The French CPSE-Labs is located in Toulouse, a city famous for its aeronautical and space achievements, its university founded in 1229, its laboratories and research centers. It is led jointly by by ONERA Midi-Pyrenees (Toulouse Centre of the French Aerospace Lab) and LAAS-CNRS.

Technology platforms: Engineering platform for autonomous dependable vehicles

Autonomous vehicles raise specific dependability issues and this platform gathers a set of methods and tools to address some of these issues at complementary design stages. Follow links to read more about each of these platforms.

In early design phases, the platform assists the specification of the dependability requirements for innovative concepts of vehicle or operation. Starting from UML model of the vehicule operation, the method HAZOP UML method enables to identify and classify operational risks due to human-robot or robot-robot interactions.

Then the AltaRica based safety assessment method is proposed to assess the robustness of complex concept of operation and preliminary system architectures. This method has been deployed in the aeronautics domain and the Design Centre provides libraries of models and the DALculator tools to check or allocate safety levels and budgets.

During more detailed design phase, SMOF research framework is proposed to assist the specification of safety rules executed by an independent monitor. This safety rules will be the last safety barrier applicable by the autonomous vehicle to ensure the safety of the whole system.

The design of real-time software architecture which orchestrate a lot of sensors acquisition as well as decision making or more classical control algorithms is supported by languages and tools defined primarily for the robotics community. MAUVE is a domain specific language and associated tools used to design robotics components. It enables generation of OROCOS components. It also enables static analysis of worst-case execution time of the module and model-checking of functional requirements. MAUVE has been already applied to OROCOS components embedded in ONERA robots. These components can be adapted to re-implement the case study in OROCOS if needed.

The Generator of Modules GenoM is a tool to design real-time software architectures. It encapsulates software functions inside independent components. GenoM is more specifically dedicated to complex on-board systems, such as autonomous mobile robots or satellites.

Finally, our design centre proposes also the MORSE simulation facilities for the functional validation of the function embedded in the autonomous vehicle. MORSE is an generic simulator for academic robotics. It focuses on realistic 3D simulation of small to large environments, indoor or outdoor, with one to tenths of autonomous robots.

Funded experiments

You can see the funded experiments which are partnered with our France design centre under CPSE Labs here.

• FormalRob (Rigorous framework for developing and validating robotic applications). This experiment targets innovative approaches of formally validating and verifying Cyber-Physical Systems (CPS) programmed using the GenoM framework. It is relevant to suppliers of formal Verification & Validation (V&V) technology willing to transfer their technology to the field of robotics.
• SafeAM (Safety enhancement for autonomous robotic airfield management): application to lighting control and management. This experiment targets innovative approaches of realizing safe autonomous CPS, with a focus on mobile ground robot applications. The experiment provides an assessment of the applicability of a number of methods and tools proposed by the French CPSE Labs Design Centre for the safety analysis, the architectural design and validation of such robotic systems. A subset of these techniques has been successfully applied in the aeronautics domain and the experiment explores their applicability to the robotic field.
• 4MOB-AIRMAN (Value chain completion for robots in airfield management). This experiment at securing the design of the 4MOB robotic mobile platform and at assessing the safety of this design for one safety critical operation: the maintenance and control of the lighting system on an airport runway. It will be focussed on the use of the SMOF and Altarica methods to respectively add safety monitors in the 4MOB platform and to assess the safety level of the resulting platform.
• RAST (Risk Analysis and Simulation Testing for Agricultural Robots). This experiment applies safety analysis and simulation-based testing to agricultural robots from Naïo Technologies. HAZOP-UML provides structured and systematic analysis of the operational risks induced by the robots, and MORSE-based testing is transferred into a new test platform based on an alternative simulator.
• SafeNav (Formal Verification Applied to the Navigation of an Autonomous Shuttle). This experiment uses formal verification methods on an autonomous vehicle, the EZ10. This EZ10 is an autonomous shuttle vehicle produced by EasyMile that can transport 10 people, typically for last-mile use cases or for short travels of some kilometers in city centers, large industrial sites, etc. Safety is at the heart of such a system and there is thus a need to apply verification techniques on the software. The approach consists of modelling the software architecture with components using Genom3 and defining properties about some attributes of these components that shall be satisfied. The BIP framework is used to perform both static analysis and real-time verification about these properties.