Pushing the limit of robotics cognition (AI, Data and Robotics Partnership) (RIA) -

Proposal results are expected to contribute to the following expected outcome:

• New generation of AI-Powered Robotics: Enabling robots to have more profound impacts than they currently have, in powering them with a deeper kind of AI, endowing them with a better perception and understanding of the world (up to semantic and explainable representations), This would allow the next generation of autonomous robots, with increased capabilities to work without/with limited supervision, as well as the next generation of interactive robots, with greatly improved intuitive, safe and efficient cognitive, social and physical capabilities, to assist humans.

In addition, depending on the focus of the proposal, the results are expected to contribute to at least one of the following expected outcomes

• Smarter robots with improved capabilities, functionalities (including complex functionalities such as manipulation of delicate, irregular, dynamic or deformable objects, navigation in un-controlled and variable or challenging and harsh environments, and continuous human-physical interactions) and an increased level of autonomy over the current state of the art, necessary to address real-world problems, while ensuring safety and reliability.
• Smooth and trustworthy (including safety and reliability) human-robot collaboration through advanced reactivity and mutual understanding, and human-centric automated adaptation of robots in human-robot interactions.

Proposals are expected to develop technologies and systems that significantly enhance the cognitive ability of robots from the current state of the art to achieve greater levels of interaction and autonomy. Proposals will address as appropriate the following:

• Develop enabling technologies, both new and existing, that extend the current state of the art in robotics perception, cognition, interaction and action as well as develop novel or advanced tools for the design and configuration of robots and robot systems that speed up the process of integration thereby reducing the time taken to deploy robot solutions. To do so by addressing the modularity and composability of solutions both in the operational context of a wide range of action and interaction use cases.
• There is also a need to address concepts such as trustworthiness, privacy, security and ethics already at the technology design phase.
• Develop lifelong autonomous robotics able to tackle unknown situations and adapt in the long term in pushing the state of the art of AI-based robots that combine monitoring, learning, planning and acting in order to evolve in difficult environments over long period of time. Support from simulation tools could be considered, as appropriate.
• Increase robot acceptance by handling adequately both human and robotic actions, with human-centric, advanced behavioural and elaborated planning models, and adopting multidisciplinary approaches including SSH[1], as well as end-user involvement in the design of solutions addressing human factors and interaction. Gender and intersectionality dimension[2] analysis should be a part of the proposals, where relevant.
• Push the limits of robotics interaction, adopting an interdisciplinary approach to integrate methods and techniques that allow the machines to engage in physical interactions with people or the environment, safely and intelligently, through specific enabling technologies: intuitiveness and responsive human-robot interfaces; integration of robot perception with natural and artificial intelligence; ability to physically, stably, dependably and safely interact with the environment, including users and surrounding people; development of advanced control tools fully integrating the human in the loop when performing a task; development of advanced control tools for dexterous and safe manipulation, assistance, and locomotion in diverse environments (ground, air, water, space, in-vivo and including safety critical and hazardous environments that are corrosive, explosive, nuclear or at extremes of pressure or temperature) and in general for improved performance of robots; energy autonomy and resilience to highly limited and imperfect communication networks in on-field applications.

Two types of proposals are expected, either focusing on higher level of autonomy, expecting less reliance on human supervision, or focusing on human-machine collaboration.

In each case, improvement in the level of robotics cognition should be demonstrated through at least three real-world scenarios (including measurements of functional performance), showing also the potential added value of such improvement in such use-cases scenarios. Scientific and technological progress should be demonstrated by qualitative and quantitative KPIs, demonstrators, benchmarking and progress monitoring. Activities are expected to achieve TRL 4-5 by the end of the project

• The first type of proposals will further develop the level of autonomy in building on the latest developments in areas such as advanced perception, smart sensors, intelligent action and interaction, reasoning and learning, increased interpretation and understanding of the complex real-world environments (possibly involving human actions), anticipation of the effect of actions, adaptation and re-planning, graceful degradation, safety and security, etc. They will, as appropriate, further develop such components, and integrate them in an advanced robotics system, consider the balance of on-board vs off-board processes and the access and utilisation of external data and cloud resources to guide tasks and missions by adding external knowledge to internal reasoning and decision-making processes.
• The second type of proposals will further develop and integrate physical human-robot interaction, verbal/non-verbal communication as well as robot-environment/object interaction, embedding, as appropiate, safety, mutual understanding perception and interpretation of human actions, interaction situated in complex real-world environments and related motivations and social structures, joint goals, shared and sliding autonomy, ethical human-centric behaviour by understanding of physiological responses and emotions, etc. to reach truly smooth human-robot collaboration. This should as well integrate advanced control developments, and further develop them as necessary to guarantee the necessary speed for the required reactivity, ensuring natural, safe and smooth interactions with humans. Appropriate use should be made of data and knowledge accumulation from internal and external sources in order to guide reasoning and decision-making and the inclusion of explainability/transparency mechanisms[3] appropriate to the use case. Such proposals should adopt a multidisciplinary approach and involve the necessary expertise in SSH[1], in particular in ethics and human-centric design to enhance trust and acceptability.

When possible proposals should build on and reuse public results from relevant previous funded actions. Proposals should make use of connections to the Digital Innovation Hub networks, particularly those in Robotics, Data and AI. Full use should be made of the common resources available in the AI-on-Demand platform[5], Digital Industrial Platform for Robotics[6], data platforms[7] and, if necessary other relevant digital resource platforms. Communicable results from projects should be delivered to the most relevant of these platforms so as to enhance the European AI, Data and Robotics ecosystem through the sharing of results and best practice.

All proposals should also take into consideration trustworthy AI principles[8] including respect of human dignity and agency, as appropriate, given the technology focus.

This topic implements the co-programmed European Partnership on AI, Data and Robotics.

All proposals are expected to allocate tasks to cohesion activities with the PPP on AI, Data and Robotics and funded actions related to this partnership, including the CSA HORIZON-CL4-2021-HUMAN-01-02. Where relevant, synergies with other PPPs are encouraged.

[1] Social Sciences and Humanities

[2] https://en.wikipedia.org/wiki/Intersectionality - intersectional aspects could cover gender, age, social level, education, ethnic origin, etc.

[3] Explainability mechanisms can use multiple communication modalities, verbal, non-verbal, gestural etc.

[4] Social Sciences and Humanities

[5] Initiated under the AI4EU project https://cordis.europa.eu/project/id/825619 and further developed in projects resulting from H2020-ICT-49-2020 call

[6] https://robmosys.eu/newsrobmosys-rosin-towards-an-eu-digital-industrial-platform-for-robotics/

[7] E.g.: https://www.big-data-europe.eu/

[8] https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai

News flashes