TWI486914B - An information processing system based on multi-layer inference architecture - Google Patents

Description

Information processing system with multi-level inference structure

The present invention relates to an information processing system for a multi-hierarchical inference architecture, and more particularly to an information processing system that can analyze various contextual features in an environment.

For environments that require intensive care or high risks, such as an elderly care environment, a medical institution, or an infant care center, special monitoring of the immediate state of the observed person is required. The usual practice is to set up in this environment. A plurality of electronic sensing devices sense sensing signals associated with various contextual features of the environment in which the viewer is located. At present, with the advancement of modern technology, a variety of electronic sensing devices have been developed, which can sense various situational features in the environment. For example, the sound sensor can sense the sound intensity and optical inductance of the surrounding environment. The detector can sense the change of light intensity in the environment, the temperature sensor can sense the temperature value in the environment, the pressure sensor can sense the pressure in the environment, and the displacement sensor can sense the environment. The distance between an object and the open/close door sensor can sense the opening/closing of the environment.

However, in order to achieve comprehensive monitoring, it is necessary to install a large number of electronic sensing devices in the environment. For example, each door needs to be equipped with an on/off door sensor to sense the opening of each door. / Closed state, each bed needs to be equipped with a pressure sensor to sense the pressure on the bed, etc. These various and numerous electronic sensing devices convert the sensed situational features into electronic signals , And often accompanied by a lot of noise,. Therefore, it is conceivable that the amount of data processed by the analysis unit that receives these electronic signals for data analysis at the back end is so large that the instant state of the observed person or the frequent false positives cannot be obtained immediately, so that the conventional Information processing systems are unable to perform effective and immediate analysis of complex contextual features, and there is also a high false positive rate.

In order to simplify the analysis of the calculation process and the amount of sensing data, the conventional information processing system introduces radio frequency identification (RFID) technology, so that the observer can wear an RFID tag to detect the immediate state of the observed person. However, although this method can effectively obtain the position of the observed person, the general acceptance of the extra RFID tag is not high, and the effect of the traditional information processing system is still limited. Therefore, there is a need for an information processing system that is non-intrusive and provides instant access to the immediate state of the observed person.

It is an object of the present invention to provide an information processing system for a multi-level inference structure, in which an inference unit outputted by a lower-level inference unit is input to a higher-level inference unit through an inference unit in a multi-level inference module. Inductive signals for various contextual features are inferred and/or further inferred by inferred information to show inferred results.

Another object of the present invention is to provide an information processing system with a multi-level inference structure, which can effectively infer a result of semantic inference by using an intrusive RFID technology through an inference unit in a multi-level inference module. In order to get instant status of the observed person.

A further object of the present invention is to provide an information processing system for a multi-hierarchical inference architecture, in which a derivation message outputted by a lower-level inference unit in a multi-level inference module is a hierarchical structure of input of a higher-level inference unit. The complex information processing process is divided by each level to save processing time and reduce computational complexity.

Another object of the present invention is to provide an information processing system with a multi-level inference architecture, such that an embedded microcontroller is provided in the sensing unit to place the first level inference module in the sensing unit, and thus transmitted. It is no longer a large amount of noise-containing signals, but processed information, which can effectively reduce the amount of communication data.

Another object of the present invention is to provide an information processing system for a multi-hierarchical inference structure, which performs an inference level analysis together with an inference unit in a multi-level inference module to infer the sensing signals of various context features.

According to the present invention, an information processing system for a multi-level inference architecture is provided, comprising: a plurality of sensing units, a multi-level inference module, and a display unit. The sensing unit senses a context feature to generate a sensing signal. The multi-level inference module receives and analyzes the sensing signals generated by the sensing unit to generate at least one inference result, and the multi-level inference module includes a plurality of inference units, which are divided into a plurality of levels, infering an inference information and outputting, respectively. The inferential message of the output of the lower-level inference unit is the input of the higher-level inference unit. The display unit receives the inference results from the output of the multi-level inference module and displays the inference results.

According to the present invention, there is provided an information processing comprising the steps of: respectively sensing a context feature to generate a sensing signal; receiving and analyzing the sensing signals to generate a plurality of inference information; and receiving and integrating the inference information to Produce at least one inference result; and display the inference result.

According to an embodiment of the present invention, the detailed structure of the multi-level inference module may be a two-level, three-level, and N-level inference module. For example, the two-level inference module may include a plurality of The first hierarchical inference unit and the at least one second hierarchical inference unit. The first level inference unit respectively receives the sensing signals generated by the sensing unit, and infers a first level inference information through the sensing signals. The second hierarchical inference unit receives the first hierarchical inference information from the first hierarchical inference unit, synchronizes the first hierarchical inference information, and integrates at least one second hierarchical inference information, and the multi-level inference module outputs the second hierarchical level. Inferred information as the result of the above inference, if the inference information of the first level has semantic meaning, the multi-level inference module output can also be used as the inference result, but the two-level inference module is not limited to this. The three-level inference module can include a plurality of first hierarchical inference units, a plurality of second hierarchical inference units, and at least a third inference unit. The first level inference unit respectively receives the signals generated by the sensing unit, and infers a first level of inference information via the sensing signals. The second hierarchical inference unit receives the first hierarchical inference information from the first hierarchical inference unit, synchronizes the first hierarchical inference information, and separately infers a second hierarchical inference information. The third level inference unit receives at least two of the second level inference information and the sensing signal from the second level inference unit, so that the second level infers information or sensed information Synchronization of the number and integration of a third-level inference information, and the multi-level inference module outputs the second-level inference information or the third-level inference information as a result of inference. If the first-level inference information has semantic meaning, then The multi-level inference module output can be used as the inference result, but the three-level inference module is not limited to this. The N-level inference module can be based on the above two-level or three-level inference modules, and then add other units of inference units, but not limited to this.

According to an embodiment of the present invention, the information processing system of the multi-hierarchical inference architecture of the present invention can be applied to a medical institution, an nursery center, a home environment, or any other environment in which a sensing signal requiring a large amount of environmental characteristics is analyzed. The type of sensing unit is not limited and may be selected from the group consisting of a pressure sensor, a displacement sensor, a temperature sensor, a laser range finder, and a switch door sensor. For hospitals, nursing homes, homes or other care settings, the following framework can be devised for hospital single-person wards: the sensing unit senses a bed pressure to determine the patient's sleeping position and a ground laser to determine how many A foot, a sofa bed pressure to determine whether the caregiver is lying on the sofa bed, a bath toilet dynamic detection to determine whether the toilet is a person and a door switch switch context feature to generate the relevant sensing signal. The first level inference unit includes a bed inference unit, a ground inference unit and an auxiliary inference unit. The bed inference unit infers a bed-bed inference information from the sensing signal associated with the bed pressure, and the ground inference unit is measured by the ground laser. A ground-based inference information is deduced from the associated sensing signal. Several auxiliary inference units are inferred from the sensing signals associated with the sofa bed pressure, the toilet dynamic detection, and the door switch. The second hierarchical inference unit includes a bed deduction unit and a caregiver inference unit. The bed inference unit receives the bed inference information and the ground inference information, and deducts a bed-off inference information based on the bed inference information and the ground inference information to determine whether the patient is leaving the bed. The caregiver inference unit receives bed inference information, ground inference information, and auxiliary inference information, and infers the caregiver's inference information based on bed inference information, ground inference information, and auxiliary inference information to determine the condition of the caregiver. The third-level inference unit is a warning level inference unit, which receives the in-bed inference information, the ground inference information, the auxiliary inference information, and the inferential information from the bed inference unit, the ground inference unit, the auxiliary inference unit, the out-of-bed inference unit, and the caregiver inference unit output. The caregiver deduces the information and integrates a warning level inference information to indicate the current alert level of the observed state. If the patient is leaving the bed and accompanied by an unaccompanied person, it is the highest alert level. Therefore, the information processing system through the multi-level inference structure can be inferred to determine the state of the observed person in the hospital, nursing home, home or other care environment, and display the caretaker's inference information for the caregiver. In order to provide timely care services.

According to an embodiment of the present invention, the step of performing inference processing by the first hierarchical inference unit may include, but is not limited to, k-mans grouping algorithm operation, Hidden Markov Model operation, and fault tolerance analysis of the backup device. Any one of the sensing signals can be selectively processed to remove noise. The steps of the second hierarchical inference unit for integrating the inference processing of the first hierarchical inference information may include, but are not limited to, inference of information synchronization, context analysis, Hidden Markov Model operation, and alert level analysis.

According to an embodiment of the present invention, the multi-level inference structure of the present invention is The processing system may additionally include an emergency notification unit connected to the multi-level inference module to transmit an emergency message when the inference result represents an emergency. The transmission of emergency messages can take many forms, such as: telephone, internet, or other types of electrical connections. The information processing system of the multi-level inference architecture of the present invention may additionally include a plurality of transmission units respectively corresponding to the sensing units, and the sensing signals generated by the sensing unit are respectively transmitted to the multi-level inference module.

Therefore, the information processing system of the multi-hierarchical inference architecture of the present invention is complicated by the inference mechanism of the multi-level inference module having the inference information output by the lower-level inference unit as the input of the higher-level inference unit. The information processing process is carried out by division of labor, and the calculations required for semantic inference results are distributed to save processing time and reduce computational complexity.

To further illustrate the various embodiments, the invention is provided with the drawings. The drawings are a part of the disclosure of the present invention, and are mainly used to explain the embodiments, and the operation of the embodiments may be explained in conjunction with the related description of the specification. With reference to such content, those of ordinary skill in the art should be able to understand other possible embodiments and advantages of the present invention. Elements in the figures are not drawn to scale, and similar elements are generally used to represent similar elements.

The information processing system of the multi-level inference architecture of the present invention can be applied to an environment, such as any environment that needs to monitor or analyze its context characteristics, for example, a medical institution, a care center, an care center, and a home environment. Or other suitable It is in places where humans live, but the present invention is not limited thereto, and may be any other environment in which a sensing signal requiring a large amount of environmental characteristics is analyzed. It is said here that human beings living in the environment are "observed." Secondly, the contextual features mentioned herein generally refer to changes in various physical values or physical values in the environment, such as temperature values, pressure values, distances, displacements, accelerations, light intensities and their variations, etc., but are not limited thereto. this. The information processing system of the multi-level inference architecture of the present invention comprises: a plurality of sensing units, a multi-level inference module and a display unit. The sensing unit senses a context feature to generate a sensing signal. The multi-level inference module receives and analyzes the sensing signals generated by the sensing unit to generate at least one inference result, and the multi-level inference module includes a plurality of inference units, which are divided into a plurality of levels, infering an inference information and outputting, respectively. The inferential information of the output of the low-level inference unit is the input of the higher-level inference unit. The display unit receives the inference results from the output of the multi-level inference module and displays the inference results. The type of the sensing unit is not limited, and may be a sensing unit that detects changes in any physical value or physical value, for example: pressure sensor, displacement sensor, temperature sensor, laser range finder And the switch door sensor is not limited to this. The inference result obtained by analyzing the situational features by the multi-level inference module is a semantic information, which immediately reflects the immediate state of the observed person. For example, the inference result can be "the observer is lying in bed" or " More accurate and accurate inference results are obtained when the observed person leaves the bed with the caregiver.

The information processing system of the multi-hierarchical inference architecture of the present invention may additionally include an emergency notification unit in addition to displaying the inference result using the display unit. The multi-level inference module is connected to transmit an emergency message when the inference result represents an emergency. The transmission of emergency messages can take many forms, such as telephone, internet, or other types of electrical connections, and is not limited to this. It should be noted that the multi-level inference module in the information processing system of the multi-hierarchical inference architecture of the present invention is not limited to the number of classes, and may be an inference of two hierarchical types, three hierarchical levels, and four hierarchical levels. Module.

For example, please refer to FIG. 1 , which shows a functional block diagram of an information processing system of a multi-hierarchical inference architecture according to a first embodiment of the present invention. Here, an N-level multi-level inference module is shown. As shown in the figure, the information processing system 1 of the multi-level inference architecture includes a plurality of sensing units 10, 11, 12, a multi-level inference module 20, a display unit 30, and an emergency notification unit 40. The sensing unit 10 represents a first sensing unit, the sensing unit 11 shows a second sensing unit, the sensing unit 12 of the N ₁ represents the sensing unit, a plurality of N ₁ is any integer. The sensing units 10, 11, 12 respectively sense a context feature to generate a sensing signal. The number of sensing units 10, 11, 12 is not limited, and the types of sensing characteristics are not limited, and are required to monitor the environment, such as: a temperature sensor, three pressure sensors, and a sound. A sensor, a displacement sensor, and the like. Each of the sensing units 10, 11, 12 can transmit the sensing signals generated by the sensing units 10, 11, 12 to the multi-level inference module 20 via a corresponding transmitting unit 13, 14, 15.

The multi-level inference module 20 receives and analyzes the sensing signals generated by the sensing units 10, 11, 12 to generate a plurality of inference results, and the multi-level inference module preferably includes a plurality of layers 21, 22, and 23 Inference units 211, 212, 213, 221, 231. The hierarchy 21 indicates the first hierarchy, the hierarchy 22 indicates the second hierarchy, the hierarchy 23 indicates the N ₂ hierarchy, and N ₂ is an arbitrary complex integer. The inference information outputted by the lower-level inference units 211, 212/213, 221 is the input of the higher-level inference unit 221/231. For example, the inference information output by the inference units 211, 212 is the inference unit 221 of the previous hierarchy. The inference information output by the input, inference units 213, 221 is the input of the higher level inference unit 231. The higher level inference unit 221/231 integrates the inference information from the different inference units 211, 212/213, 221 of the lower level into an inference process. The first hierarchical inference unit, such as the inference unit 211, 212, 213, may perform the inference processing steps including, but not limited to, k-mans grouping algorithm operation, Hidden Markov Model operation, and fault tolerance of the backup device. Any of the analysis, and the noise removal process can be selectively performed on the sensing signal. The higher-level inference unit, such as: the inference unit 221/231, the step of integrating the inference from the different inference units from the lower level, such as the inference units 211, 212/213, 221, may include, but is not limited to: the situation Analyze, hide any of the Hidden Markov Model operations, and alert level analysis.

The inference result obtained after the integration of the multi-level inference unit 20 is inferred is semantic information, which can reflect the immediate state of the observed person, and the result of the inference is output to the display unit 30 and the emergency notification unit 40 for display on the display unit 30. And for the observer to know, or when the inference result represents an emergency, the emergency notification unit 40 transmits an emergency message to notify the observer or the emergency management order. The location is for emergency elimination. If the information inferred by the first hierarchical inference unit 21 has semantic meaning, it may also be output to the display unit 30 and the emergency notification unit 40. The observer can provide care and monitors, but is not limited to this.

Please also refer to FIG. 2, which shows a functional block diagram of an information processing system of a multi-hierarchical inference architecture according to a second embodiment of the present invention. Here, a two-level multi-level inference module is shown. Herein, in order to briefly describe the features of the embodiment, only the differences between the embodiment and the first embodiment will be described. As shown in the figure, the multi-hierarchy inference module 20A of the information processing system 2 of the multi-hierarchical inference architecture of the present embodiment is different from the first embodiment in that the information processing system 2 of the multi-hierarchical inference architecture of the present embodiment is The hierarchical inference module 20A is specifically a two-level multi-level inference module 20A. The number of inference units 211, 212, 221 and the number of sensing units 10, 11 in each level 21, 22 are not limited to the numbers shown in the figures, but may be other numbers. The inference units 211 and 212 of the first level 21 respectively receive the sensing signals generated by the sensing units 10 and 11 and remove the noise of the sensing signals to infer a first level of inference information. The inference unit 221 of the second level 22 receives the first level inference information from the inference units 211, 212 of the first level 21, synchronizes the first level inference information, and integrates at least one second level inference information. The multi-level inference module 20A outputs this second-level inference information as an inference result for use by 30, 40.

Please also refer to FIG. 3, which shows a functional block diagram of an information processing system of a multi-hierarchical inference architecture according to a third embodiment of the present invention. Here, a three-level multi-level inference module is shown. Herein, in order to briefly describe the features of the embodiment, Only the differences between the embodiment and the first embodiment will be described. The information processing system 3 of the multi-hierarchical inference architecture of the present embodiment can understand that the internal architecture of the information processing system 3 of the multi-level inference architecture and the data processing path performed can vary with the requirements. As shown in the figure, the sensing unit 10 of the information processing system 3 of the multi-level inference architecture of the present embodiment can sense a first context feature, the sensing unit 11 can sense a second context feature, and the sensing unit 12 A third context feature can be sensed to generate a sensing signal. At this time, the multi-hierarchy inference module 20B of the information processing system 3 of the multi-level inference structure of the present embodiment includes the inference units 211, 212, 213, 221, 222, and 231 of the three levels 21, 22, and 23. The inference units 211, 212, 213 of the first level 21 may illustratively include a first first hierarchical inference unit 211, a second first hierarchical inference unit 212, and a third first hierarchical inference unit 213, first A hierarchical inference unit 211 infers a first first hierarchical inference information from the sensing signal associated with the first context feature, and the second first hierarchical inference unit 212 infers a sensing signal associated with the second context feature. The second first level inference information 213 infers a third first level inference information from the sensing signal associated with the third context feature. The inference units 221, 222 of the second level 22 may illustratively include a first second hierarchical inference unit 221 and a second second hierarchical inference unit 222, a first second hierarchical inference unit 221 and a second second hierarchical inference The unit 222 receives at least two of the first first hierarchical inference information, the second first hierarchical inference information, and the third first hierarchical inference information, for example, the first second hierarchical inference unit 221 receives the first first hierarchical inference information. And the second first level inference information, the second second level inference unit 222 receiving first first level inference information, second first level inference information, and third first level inference information. The first second hierarchical inference unit 221 and the second second hierarchical inference unit 222 respectively synchronize the received first first hierarchical inference information, the second first hierarchical inference information or the third first hierarchical inference information, and The inferences of a second class are inferred separately. The inference unit 231 of the third level 23 receives at least two of the second level inference information and the sensing signal from the first second level inference unit 221 and the second second level inference unit 222 of the second level 22, such as: Receiving the first first level inference information, the second first level inference information, the third first level inference information, and the second level inference information, and causing the second level to infer information such as the first first level inference information, The first level of inference information, the third first level of inference information, and the second level of inference information or sensing signals are synchronized, and a third level of inference information is inferred. The first and second hierarchical inference information and the third hierarchical inference information are displayed on the display unit 30, and the third hierarchical inference information is provided to the emergency notification unit 40 to ascertain whether an emergency situation is currently generated.

When the information processing system 3 of the multi-level inference architecture is applied to the aforementioned environment, such as a hospital, a nursing home, a home, or other care environment, the following architecture may be specified: the sensing units 10, 11, 12A, 12B of varying amounts 12C is a pressure bar 10 for sensing a bed pressure, a ground laser ranging sensor 11 for sensing a pair of feet on the ground, and an opening/closing door sensing unit 12A for sensing a door switch, sensing A sofa bed pressure to determine whether someone is lying on the sofa bed pressure bar 12B and determine whether the bathroom is a person's bath dynamic detector 12C The sensing signal is generated and its sensing signal is transmitted to the inference units 211, 212, 213A, 213B, 213C of the first level 21 of unequal numbers via the corresponding transmitting units 13, 14, 15A, 15B, 15C. The inference units 211, 212, 213A, 213B, and 213C of the first level 21 may be a bed inference unit 211, a ground inference unit 212, and a plurality of auxiliary inference units 213A, 213B, and 213C, and the bed inference unit 211 is provided by the bed surface. The pressure-related sensing signal infers a bed-bed inference information. The ground inference unit 212 infers a ground inference information from the sensing signals associated with the ground pressure, and the auxiliary inference units 213A, 213B, and 213C infer a number of auxiliary inference information. The inference unit 221, 222 of the second level 22 includes a bed deduction unit 221 and a caregiver inference unit 222. The bed deduction unit 221 receives bed inference information from the bed inference unit 211 and ground inference information from the ground inference unit 212. The bed deduction information and the ground inference information infer a bed deduction information, and the caregiver inference unit 222 receives the bed inference information from the bed deduction unit 211, ground inference information from the ground inference unit 212, and from several auxiliary inference units 213A, 213B. And the auxiliary inference information of 213C, based on the bed inference information, the ground inference information and the auxiliary inference information, infers that one of the semantic caregiver inference information is transmitted to the display unit 30. The inference unit 231 of the third level 23 is a warning level inference unit 231, which receives the inference information from the bed and the inference information of the caregiver from the output of the bed deduction unit 221 and the caregiver inference unit 222, and integrates a warning level inference information, indicating that the current The alert level of the observed state is transmitted to the emergency notification unit 40 and the display unit 30. Therefore, information through the aforementioned multi-level inference structure The treatment system can be inferred to determine the status of the observed person in the hospital, nursing home, home or other care setting, and to display a careless caregiver's inference information for the caregiver to provide timely care.

It should be noted that the N-level inference module of the present invention can be based on the two-level or three-level inference module, and the inference unit of other classes is added, but not limited thereto.

Please also refer to FIG. 4, which shows a schematic diagram of an information processing method according to a fourth embodiment of the present invention. As shown in the figure, the information processing method of the present embodiment can be applied to an information processing system of a multi-hierarchical inference architecture, such as the information processing system of the multi-hierarchical inference architecture shown in FIGS. 1 to 3, but Limited thereto, it can also be applied to other information processing systems in accordance with the present invention. First, a context feature is respectively sensed to generate a sensing signal (step S410). The context feature generally refers to changes in various physical values or physical values in the environment, such as: temperature value, pressure value, distance, displacement, acceleration, Light intensity and its changes, etc., are not limited to this. The context feature is preferably generated by a plurality of sensing units, such as a pressure sensor, a displacement sensor, a temperature sensor, a laser range finder, and a switch door sensor. Thereafter, the sensing signals are received and analyzed by a multi-level inference module to generate a plurality of inference information (step S420), preferably by the inference unit of the first level in the multi-level inference module. In this step, any processing for removing noise from the sensing signal and any of the following inference operations may be combined: k-mans grouping algorithm operation, Hidden Markov Model operation, and fault tolerance analysis of the backup device. However, it is not limited to this. Multi-level inference module The inference unit of the second and higher classes receives and integrates the inference information to generate at least one inference result (step S430), in which any of the following integrated inference operations are performed on the inference information: the situation Analysis, hiding Hidden Markov Model calculations, and warning level analysis are not limited to this. Next, the inference result is displayed by a display unit such as a screen or other device capable of displaying information (step S440). If the information generated by S420 is already semantic, it can be directly output to S440 as the inference result.

Therefore, as can be seen from the above, the information processing system of the multi-hierarchical inference architecture of the present invention uses the inference information output by the lower-level inference unit of the multi-level inference module as the input of the higher-level inference unit. The inference mechanism divides the complex information processing process from different levels of division, and decentralizes many semantic inference results to save processing time and reduce computational complexity.

The above description is based on a number of different embodiments of the invention, wherein the features may be implemented in a single or different combination. Therefore, the disclosure of the embodiments of the present invention is intended to be illustrative of the embodiments of the invention. Further, the foregoing description and the accompanying drawings are merely illustrative of the invention and are not limited. Variations or combinations of other elements are possible and are not intended to limit the spirit and scope of the invention.

1, 2, 3 ‧ ‧ multi-level inference architecture information processing system

10, 11, 12A, 12B, 12C‧‧‧ Sensing unit

13, 14, 15A, 15B, 15C‧‧‧ transmission unit

20, 20A, 20B‧‧‧ multi-level inference module

21, 22, 23‧‧ ‧

30‧‧‧Display unit

40‧‧‧Emergency Notification Unit

211, 212, 213, 221, 222, 231A, 231B, 231C‧‧‧Inference units

S410, S420, S430, S440‧‧ steps

1 is a diagram showing an information processing system of a multi-level inference architecture according to a first embodiment of the present invention.

Figure 2 is a diagram showing an information processing system of a multi-level inference structure according to a second embodiment of the present invention.

Figure 3 is a diagram showing an information processing system of a multi-level inference architecture according to a third embodiment of the present invention.

Figure 4 is a diagram showing an information processing method according to a fourth embodiment of the present invention.

1‧‧‧Multi-level inference architecture information processing system

10,11,12‧‧‧Sensing unit

13, 14, 15‧ ‧ transmission unit

20‧‧‧Multi-level inference module

21, 22, 23‧‧ ‧

30‧‧‧Display unit

40‧‧‧Emergency Notification Unit

211, 212, 213, 221, 231‧‧‧ Inference units

Claims (11)

An information processing system of a multi-level inference structure includes: a plurality of sensing units respectively sensing a context feature to generate a sensing signal; a multi-level inference module receiving and analyzing the generated by the sensing units Sensing the signal to generate at least one inference result, the multi-level inference module includes a plurality of inference units, is divided into a plurality of levels, infers an inference information and outputs, respectively, and the inference information output by the inference unit of the lower level is The input of the inference unit of the high level includes: a plurality of first level inference units respectively receiving the sensing signals generated by the sensing units, and inferring a first level inference through the sensing signals respectively And at least one second-level inference unit that receives the first-level inference information from the first-level inference units, synchronizes the first-level inference information, and integrates at least one second-level inference a display unit that receives at least one inference result from the output of the multi-level inference module and displays the at least one inference result; The multi-stratum inference inference module outputs the first hierarchical information and / or the second level information as the at least one inference inference results. For example, the information processing system of the multi-level inference structure of claim 1 of the patent scope, wherein the multi-level inference module includes: a plurality of first level inference units respectively receiving the sensing signals generated by the sensing units, and inferring a first level inference information through the sensing signals respectively; and a plurality of second hierarchical inference units receiving the information from The first-level inference information of the first-level inference units synchronizes the first-level inference information and separately infers a second-level inference information; and a third-level inference unit receives the comparison information. The inference information of the lower-level inference unit synchronizes the inference information of the lower-level inferences and integrates a third-level inference information; wherein the multi-level inference module outputs the first-level inference information, The second level of inference information and/or the third level of inference information is the result of the at least one inference. For example, the information processing system of the multi-level inference structure of claim 2, wherein: the sensing unit senses a bed pressure, a ground laser ranging sensor, a door switch, a sofa bed pressure, and a toilet Dynamically sensing the contextual features to generate the sensing signals; the first hierarchical inference unit includes a bed inference unit, a ground inference unit, and a plurality of auxiliary inference units, the bed inference unit being associated with the bed pressure The sensing signal infers a bed-bed inference information that is inferred from the sensing signal associated with the ground laser ranging sensing a ground inference information, the auxiliary inference unit deduces a plurality of auxiliary inference information by the sensing signals associated with the door switch, the pressure on the sofa bed, and the dynamic sense of the toilet; the second level inference unit includes one a bed deduction unit and a caregiver inference unit, the bed deduction unit receives the bed inference information and the ground inference information, and infers a bed departure inference information based on the bed inference information and the ground inference information, the caregiver inference unit receives the bed rest Inference information, the ground inference information, and the auxiliary inference information, inferring a caregiver inference information based on the bed inference information, the ground inference information, and the auxiliary inference information; and the third level inference unit including a warning level The inference unit receives the off-the-bed inference information output by the off-bed inference unit and the caregiver inference unit and the inference information of the caregiver, and integrates a warning level inference information. For example, the information processing system of the multi-level inference structure of any one of the first to third aspects of the patent application, wherein the first-level inference units perform the process of removing noise from the sensing signals. For example, the information processing system of any multi-level inference structure of claim 1 to 3, wherein the first hierarchical inference unit performs any of the following inferences on the sensing signals: k-mans grouping calculus Method operations, Hidden Markov Model operations, and fault tolerance analysis of backup devices. For example, the information processing system of any multi-level inference structure of any one of claims 1 to 3, wherein the at least one or the second hierarchical inference unit performs any of the following integrated inferences on the first hierarchical inference information : Multi-information synchronization processing, situation analysis, Hidden Markov Model calculation, and warning level analysis. For example, the information processing system of the multi-hierarchical inference structure of claim 1 or 2, wherein the multi-level inference module outputs the first-level inference information as the result of the at least one inference result as semantic information. The information processing system of the multi-layer inference structure of the first to third aspects of the patent application, wherein the sensing units respectively comprise an embedded microcontroller, and the first hierarchical inference units are disposed in the sensing units. Among these embedded microcontrollers. For example, the information processing system of the multi-level inference structure of claim 1 further includes an emergency notification unit connected to the multi-level inference module to transmit an emergency message when the at least one inference result represents an emergency situation. . For example, the information processing system of the multi-level inference structure of claim 1 of the patent scope further includes a plurality of transmission units respectively corresponding to the sensing units, The sensing signals generated by the sensing units are respectively transmitted to the multi-level inference module. The information processing system of the multi-hierarchical inference structure of claim 1, wherein the sensing units are selected from the group consisting of: a sound sensor, a temperature sensor, a pressure sensor, a displacement sensor, Switch door sensor, and laser ranging sensor.