|HEAL9000 HEALTHCARE AGENTS AND LEARNING ROBOTS|
The HeAL9000 project proposes the creation of a “Service Robot” in the Healthcare area. It sets out to develop a platform that offers cognitive abilities in a Robot capable of: (i) learning information through the perception of the environment in which it works, (ii) interacting with people and things so that it can look after patients in a natural way, (iii) monitoring events and actions to be activated in the event of specific cases.
The support software platform, when the robotic agent is released, will have the task of managing the life cycle (from training to testing) of the Machine Learning and Process Mining processes that enable the cognitive functions underlined above, to improve them based on the information obtained in the field and to adapt them to their area of use and to the user.
The production of the pre-industrial prototype will develop a complex “cognitive agent”, able to anticipate user needs based on physiological/environmental monitoring and to provide assistance both in terms of response to adverse situations as well as in terms of physical/cognitive stimulation to improve the quality of the therapy. The architecture proposed should guarantee the installation and updating of new validated models via OTA (Over The Air) so as to maintain the Robot's intelligence and align its features to sector's research and studies.
The Robot, with its sensor system and computing ability, will have to adapt its functioning model to the ecosystem in which it operates. It will be able to learn the habitual behaviours of the subject it interacts with, the actions the subject must carry out, monitor the main parameters perceived directly or by interacting with environmental sensors, wearable sensors and specialised devices. The Robot will therefore be able to analyse the volume of information learned in the field (sensing) (i) directly on site or (ii) using an external service (e.g. cloud); at the same time, it will be able to carry out direct actions, deciding necessary actions (act) or the appropriate responses to the questions of the monitored subject: the initiatives may involve independent actions or responses/actions towards other connected devices, thus acting as an intelligent gateway.
The goals of the project are therefore:
- Study therapist-patient interaction mechanisms in order to make for a more effective robot-mediated rehabilitative treatment than the current state of the art. Analyse the mechanisms that lie behind multi-modal interaction between the therapist and the patient during traditional treatment in order to develop technologies capable of repeating them in the rehabilitative treatment with robot.
- Develop a verbal communication and physical and cognitive interaction system. In order to repeat the communication and multi-modal interaction methods typical of the therapist-patient combination, a patient interface system will be developed which integrates a vocal synthesis system for verbal communication with the patient, a viewing system to recognise the patient's gestures and facial expressions, and a system to monitor the patient's biomechanical and psycho-physiological state. The system will be able to orchestrate the motor and dialogue control components so as to exploit the reciprocal dependencies and make an integrated plan.
- Develop an intelligent robotic therapist The robotic system will have a manipulator and gripper to manipulate the patient's limbs. An autonomous control system will be developed for the robot, capable of establishing symbiotic interaction with the patient. The system will have language abilities capable of holding a natural dialogue with the patient and, at the same time, controlling the dialogue based on the objectives of the therapy and on the state of the patient, using explicit models of his physical and emotional condition. The system will also be able to learn and progressively improve its behaviour, adapting it to the patient and to the changes in his motor and psycho-physiological condition, during each individual rehabilitative session and over time.
- Testing the system in the field of operation The HeAL9000 system will be tested at the Campus Bio-Medico University, initially on healthy subjects and later on patients suffering from neurological and musculoskeletal diseases.
The partners involved in the project are:
- Reveal S.r.l.
- Exprivia S.pA.
- Campus Bio-Medico University of Rome
PROJECT CO-FINANCED BY THE EUROPEAN UNION
POR FESR LAZIO 2014/2020 – REGION OF LAZIO
Public notice “Progetti Strategici - 2019”
Protocol number A0320-2019-28108
HeAL9000 CUP Code B84I20001880002