Auto-adaptive robot-aided therapy using machine learning techniques.

This paper presents an application of a classification method to adaptively and dynamically modify the therapy and real-time displays of a virtual reality system in accordance with the specific state of each patient using his/her physiological reactions. First, a theoretical background about several machine learning techniques for classification is presented. Then, nine machine learning techniques are compared in order to select the best candidate in terms of accuracy. Finally, first experimental results are presented to show that the therapy can be modulated in function of the patient state using machine learning classification techniques.

Robotic assessment of the influence of age on upper-limb sensorimotor function.

PURPOSE: This paper examines the influence of age on several attributes of sensorimotor performance while performing a reaching task. Our hypothesis, based on previous studies, is that aged persons will show differences in one or more of the attributes of sensorimotor performance. PATIENTS AND METHODS: Fifty-one subjects (aged 20-80 years) with no known neuromotor disorders of the upper limbs participated in the study. Subjects were asked to grasp the end-effector of a pneumatic robotic device with two degrees of freedom in order to reach peripheral targets (1.0 cm radius), "quickly and accurately", from a centrally located target (1.0 cm radius). Subjects began each trial by holding the hand within the central target for 2000 milliseconds. Afterwards, a peripheral target was illuminated. Then participants were given 3000 milliseconds to complete the movement. When a target was reached, the participant had to return to the central target in order to start a new trial. A total of 64 trials were completed and each peripheral target was illuminated in a random block design. RESULTS: SUBJECTS WERE DIVIDED INTO THREE GROUPS ACCORDING TO AGE: group 1 (age 20-40 years), group 2 (age 41-60 years), and group 3 (age 61-80 years). The Kruskal-Wallis test showed significant differences (P < 0.05) between groups, except for the variables postural speed in the dominant arm, and postural speed and initial deviation in the non-dominant arm (P > 0.05). These results suggest that age introduces significant differences in upper-limb motor function. CONCLUSION: Our findings show that there are objective differences in sensorimotor function due to age, and that these differences are greater for the dominant arm. Therefore for the assessment of upper-limb function, we should take into account the influence of age. Moreover, these results suggest that robotic systems can provide a new and effective approach in the assessment of sensorimotor function.

Pneumatic robotic device for early delivering of rehabilitation therapy.

This paper describes a new pneumatic rehabilitation robot for upper limbs to deliver Proprioceptive Neuro-muscular Facilitation (PNF) therapies to the acute post-stroke patients, even if they are still in supine position. The robotic device assists the therapist in repetitive PNF therapies, learning the defined movement by therapist at the same time that the patient, and then repeating it with different level of assistance. Moreover, the rehabilitation device was designed to be used for relearning daily living skills like: take a glass, drinking, etc. The proposed solution is composed by two robotic arms actuated by pneumatic swivel modules and a virtual environment for the motivation of the patient.

Interfaz cerebral no invasiva para control de un sistema domótico por personas discapacitadas / Non-invasive cerebral interface to control a domotic system by disabled people

Objetivo: Desarrollar una interfaz cerebral no invasiva basada en señales EEG que diferencie estados mentales generados de forma intencionada por una persona para controlar el sistema domótico de una vivienda. Material y método: Participaron 5 voluntarios hombres sanos, con edades comprendidas entre 23 y 28 años. Se procesaron y clasificaron los datos para obtener la configuración de los algoritmos que mejor diferencian entre los diferentes estados mentales. Se realizó una emulación del tiempo real para determinar como se comporta el sistema y medir el tiempo requerido por el usuario para modificar las opciones del sistema domótico. Resultados: En las pruebas offline se obtuvieron el 59.4% de acierto, un 27.7% de no detección y un 12.9% de error. En las pruebas online mejoraron los resultados obtenidos con un 70.7% de acierto, un 23.4% de no detección y un 5.9% de error y un tiempo medio de 15 segundos para activar una opción en el menú domótico. Conclusiones: La interfaz cerebral permite de forma satisfactoria controlar el sistema domótico (AU)

Objetive: To develop an EEG-based non-invasive cerebral interface to differentiate between several mental states intentionally generated by a person with the purpose of controlling the domotic system of a house. Material and method: 5 healthy volunteer subjects, all men between 23 and 28 years old, have participated in the study. Offline data have been collected, processed and classified in order to obtain the best configuration of the algorithms that allow differentiate between the mental states. Then, an emulation of the real time has been done to analyze the behaviour of the system and to measure the time required by the user to modify the options of the domotic system. Results: in the offline tests, means % of 59.4% of success, a 27.7% of non-detection and a 12.9% of error have been obtained. In the online tests, the results have been improved. Means % of 70.7% of success, a 23.4% of non-detection and a 5.9% of error, and an average time required of 15 seconds to activate an option of the domotic menu have been obtained. Conclusions: based on the results with the system we can conclude that the brain interface allows successfully control the domotic system (AU)