People monitoring and personal assistance system, in particular for elderly and people with special and cognitive needs

• US 20180368780A1
• Filed: 02/03/2017
• Published: 12/27/2018
• Est. Priority Date: 02/04/2016
• Status: Active Grant

• Current Assignee: Universita' Degli Studi Di Genova (Repubblica Italiana), Teseo S.R.L.
• Original Assignee: Dotvocal Srl, Universita' Degli Studi Di Genova (Repubblica Italiana), Teseo S.R.L.
• Inventors: Bruno, Barbara, Giuni, Antonella, Mastrogiovanni, Fulvio, Reboscio, Enrico, Scalmato, Antonello, Sgorbissa, Antonio
• Granted Patent: US 10,682,097 B2
• Time in Patent Office: Days
• CPC Class Codes: A61B 2503/08 Elderly
• CPC Class Codes: A61B 5/1113 Local tracking of patients, e.g. in a hospital or private home
• CPC Class Codes: A61B 5/1116 Determining posture transitions
• CPC Class Codes: A61B 5/4088 Diagnosing of monitoring cognitive diseases, e.g. Alzheimer, prion diseases or dementia
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 5/746 Alarms related to a physiological condition, e.g. details of setting alarm thresholds or avoiding false alarms alarms per se G08B
• CPC Class Codes: G08B 21/0423 detecting deviation from an expected pattern of behaviour or schedule
• CPC Class Codes: G08B 21/043 detecting an emergency event, e.g. a fall

Abstract

A system for monitoring and private assistance to people includes sensors that detect descriptive parameters of environmental conditions, and sensors and/or devices that monitor a person'"'"'s position and, optionally, her physiological parameters. In particular, the system includes sensors and/or devices that monitor a person'"'"'s position at the topological level and generate tracking signals in time and within a digital topological model of the environment; sensors that monitor posture and/or its variations in time and that provide status signals corresponding to predetermined postures and/or signals, indicating posture variations in time, which are processed by a logic control unit and software that evaluates position and posture and its variation in time and that classifies the signals according to predetermined categories of motor and/or postural activities, and that performs a comparison between detected parameters and predefined patterns of execution.

Claims

1. System for private monitoring and assistance of a person, comprising:

one or both of first sensors detecting parameters describing conditions of an environment where the person is located, or second sensors detecting at least a position or a movement of the person in the environment, and optionally a posture of the person and his/her physiological parameters,

wherein the first and the second sensors comprise;

third sensors or devices monitoring the position of the person at a topological level, the third sensors generating signals tracking the position of the person in time and within a digital topological model of the environment in which the person is located;

fourth sensors monitoring the posture of the person and variations thereof in time, the fourth sensors providing status signals corresponding to predetermined postures or signals of the time variation of the posture of the person;

optionally, fourth sensors monitoring one or more of physiological parameters related to the assisted person or a variation of these parameters over time; and

a processing unit configured to perform processing steps according to a classification algorithm which is coded in instructions executable by said processing unit, to classify one or more of symbols, combinations of the symbols, or temporal sequences of the symbols, and which provides the following operational steps;

representing, with the symbols, combinations of the symbols, or temporal sequences of the symbols, data from said third and fourth sensors;

determining whether the symbols, combinations of symbols, or temporal sequences of the symbols correspond to normative behavior patterns described in a data structure represented by a classification formalism or an equivalent formalism from an expressive and computational point of view;

classifying the symbols, combinations of symbols, temporal sequences of the symbols as compatible or not compatible with respect to the normative behavior patterns describing typical everyday situations or even at a level of specific customs and habits of the person;

deciding whether to activate a procedure for dialogue with the person in order to acquire additional data from the third and fourth sensors to be associated with appropriate symbols, combinations of symbols, or temporal sequences of the symbols; and

deciding whether to activate alarm procedures depending on type and severity of the alarm.

2. The system according to claim 1, wherein the normative behavior patterns, are achieved by providing in the system a learning phase performed on a limited number of experimental users and for a limited period of time, the learning phase collecting information on behavioral characteristics of the person in relation to usual conditions of behavior, habit, or a synthesis of a suitable number of said normative behavioral patterns.

3. The system according to claim 1, wherein the data relating to one or more of the parameters of position of the monitored person her movement in the environment, or her posture and variation thereof are combined with the data related to physiological parameters of the monitored person detected by the fourth sensors.

4. The system according to claim 1, wherein environmental condition sensors are also used.

5. The system according to claim 1, wherein the classification algorithm is encoded as instructions in a memorized program in a memory and is callable from the memory and executable by a processing unit, whereby the classification algorithm configures the processing unit as a control unit that carries out a control procedure which integrates a classification of activities recorded from the data of the position of the monitored person in the environment, the variation of the said position or posture, or a variation of the same with the data of the physiological parameters of the or the data of environmental conditions, thus increasing precision of the classification and the conditions of said monitored person.

6. The system according claim 1, wherein, when the classification of the symbols, the combinations of said symbols, or temporal sequences of said symbols indicates a condition of abnormal behavior of the monitored person or outside behavior of an usual behavioral patterns or an otherwise dangerous situation, making it possible to automatically generate alarms to be sent to a number of people and organizations decided during installation or setup phases said people having competence to intervene addressing situations recognized by the classification procedure.

7. The system according to claim 1, further comprising means for direct passive communication with the monitored person.

8. The system according to claim 1, wherein symbols, combinations of symbols, or temporal sequences of the symbols corresponding to activities of the monitored person are classified periodically, being measured evolutions of cognitive ability or the physiological conditions of said monitored person, thus enabling a prompt intervention not only to protect the monitored person from a condition of immediate danger, but also to put therapeutic foundations necessary to cope with a worsening of her conditions, in a still relatively initial moment of a possible pathology.

9. The system according to claim 1, wherein the system comprises a modular and scalable architecture with four levels, which comprise:

i. a layer consists of a hardware infrastructure;

ii. a layer consists of the classification algorithm, executed on the hardware infrastructure, which implements both a classification procedure based on a taxonomic data structure and a dialogue with the monitored person; and

iii. a layer constituted by a data structure containing symbols representing models, normative behavioural patterns, symbols which describe some characteristics of an environment, or symbols which refer to possible dialogues described in a logical formalism based on description logics;

wherein a data structure includes;

a. models that connect sensory data to the symbols, combinations of the symbols or temporal sequences of the symbols that refer to actions executed by the monitored person and the postures of the monitored person;

b. models that connect the sensory data to the symbols, the combinations of the symbols, or the temporal sequences of the symbols that refer to typical habits of the monitored person;

c. a model that represents the environment in which the monitored person lives and acts, based on characteristics thereof using the symbols, the combinations of the symbols, or the temporal sequences of the symbols, for localization of the monitored person at topological level;

d. models that encode normative aspects related to healthy habits for a the monitored person by the symbols, the combinations of the symbols, or the temporal sequences of the symbols, to define whether the symbols, the combinations of the symbols, or the temporal sequences of the symbols representing the data from the first and the second sensors correspond to potentially dangerous situations or to bad habits of the monitored person;

e. rule-based models that based on a classification of the symbols, the combinations of the symbols, or the temporal sequences of the symbols generate alarms or messages, or receive feedback for dialogues with the monitored person; and

f. models related to rules or based on structures related to formal grammars for encoding dialogues with the monitored person using the symbols the combinations of the symbols, or temporal sequences of the symbols; and

iv. a layer of communication and sharing of the data among system hardware devices and with external communication systems;

further comprising a communication network that uses standard open protocols for interfacing with all the hardware components of the system.

10. A method for monitoring and carrying out private assistance of people of all ages, in situations where the people cannot be considered as independent, the method comprising:

a. generating a taxonomic data structure populated with symbols, combinations of symbols, or temporal sequences of the symbols which refer to various types of models, and which require data on a position of an assisted person related to behaviors in a predetermined environment, movements or displacements of the assisted person in the predetermined environment, the posture of the assisted person, or the changes in the posture in carrying out activities;

b. representing and managing the symbols, the combinations of the symbols, or temporal sequences of the symbols related to the activities of daily life, both simple and composed by different actions or postures together the time information and information about all relevant locations;

c. generating a virtual map of an environment in which the assisted person is present and a representation of the environment in the taxonomic data structure using the symbols or combinations of the symbols;

d. determining position or tracking movements of the assisted person with reference to time, and determining the posture the changes of the posture of the assisted person in time;

e. generating a behavioral model of the assisted person from data provided by sensors and represented by the symbols, the combinations of the symbols, or the temporal sequences of the symbols in the taxonomic data structure;

f. classifying tipology of behavior and adequacy of the behavior of the assisted person compared to normative patterns of the behavior represented in the taxonomic data structure;

g. performing step f, but in relation to sequences or compositions of the actions as defined in point b;

h. generating and issuing audio or visual communications to assistance personnel or to the assisted person according to a type of the behavior detected during classification and to appropriateness of the behavior with normative behavior patterns represented in the taxonomic data structure; and

generating, issuing, and managing, by a classification algorithm, suitable vocal dialogues such that that the classification algorithm enables, by analyzing feedback, an improvement in a monitoring of the assisted person, inferring her state of health, recognizing an existence of dangerous situations, or identifying most important needs case of a rescue operation.

11. The method according to claim 10, further comprising detecting physiological parameters of the assisted person and further, optionally, environmental parameters, and integrating the physiological and environmental parameters in the behavioral model, being the physiological and environmental parameters synchronized with reference to an instant of detection or a duration in time.