Method for recognizing and identifying emergency situations in an anesthesia system by means of a self-organizing map

• US 5,584,291 A
• Filed: 03/25/1994
• Issued: 12/17/1996
• Est. Priority Date: 03/26/1993
• Status: Expired due to Fees

• Current Assignee: Instrumentarium Corporation
• Original Assignee: Instrumentarium Corporation
• Inventors: Vapola, Mauri, Merilainen, Pekka
• Primary Examiner(s): Sykes, Angela D.
• Assistant Examiner(s): Huang, Stephen D.
• Application Number:
  US08/217,889
• Time in Patent Office: 998 Days
• Field of Search: 128/630, 128/633, 128/670, 128/671
• US Class Current: 600/301
• CPC Class Codes: A61M 16/0833 T- or Y-type connectors, e.g. Y-piece
• CPC Class Codes: A61M 16/085 Gas sampling
• CPC Class Codes: A61M 16/104 specially adapted for anaesthetics A61M16/18 takes precedence
• CPC Class Codes: A61M 2205/15 Detection of leaks
• CPC Class Codes: A61M 2230/432 partial CO2 pressure (P-CO2)
• CPC Class Codes: A61M 2230/435 partial O2 pressure (P-O2)
• CPC Class Codes: G16H 20/10 relating to drugs or medications, e.g. for ensuring correct administration to patients
• CPC Class Codes: G16H 20/40 relating to mechanical, radiation or invasive therapies, e.g. surgery, laser therapy, dialysis or acupuncture
• CPC Class Codes: G16H 40/63 for local operation
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: Y10S 128/925 Neural network

Abstract

A method for recognizing and identifying emergency situations in an anesthesia system measures a plurality of variables associated with an anesthesia delivery, the measurement values of said measured variables are formed into pattern vectors characterizing the instantaneous states of the system, the system is trained prior to the actual measurement situation with a reference material comprising pattern vectors corresponding to actual situations, and on the basis of measured pattern vectors corresponding to actual situations, and on the basis of measured pattern vectors and trained material, the emergency situations are recognized and possibly identified. The reference material is trained to the system using self-organizing maps, and the pattern vectors formed from the measurement results are compared with weight vectors (m₁) of the self-organizing map to recognize and identify the emergency situations.

Claims

1. A method for recognizing and identifying emergency situations in an anesthesia system comprising:

a) measuring a plurality of variables associated with an anesthesia delivery to produce measurement values associated therewith,

b) forming the measurement values of said measured variables into pattern vectors characterizing instantaneous states of the system,

c) providing a emergency identifying self organizing map;

d) training the emergency identifying self organizing map, prior to an actual measurement situation, by providing reference material comprising pattern vectors corresponding to actual situations, the pattern vectors being arranged in said self organizing map as weight vectors, each weight vector representing a particular combination of system parameters representing an actual condition, and

e) comparing the pattern vectors formed from the measurement results with weight vectors (m₁) of the emergency identifying self-organizing map to recognize and identify emergency situations.

2. A method as defined in claim 1, wherein the emergency identifying self organizing map of claim 1 is a single level self organizing map, said method being capable of concurrently identifying both the existence and cause of an emergency situation by means of level 1 and level 2 self-organizing maps, said step e) identifying the cause of the emergency situation defined by the closest weight vector in said level 1 self organizing map to the pattern vector corresponding to the current measurement values produced in said step b), each weight vector of the level 1 self organizing map being associated with one of a plurality of level 2 self organizing maps;

said method further including,

f) utilizing the level 2 self organizing map corresponding to the weight vector in said level 1 self organizing map to further identify the cause of the emergency situation.

3. The method of claim 2 wherein said level 2 self organizing maps are also formed of a plurality of map elements each having a weight vector representing a particular combination of system parameters representing an actual condition.

4. The method of claim 3 wherein the level 2 self organizing map contains only weight vectors corresponding to weighing vectors arranged closer to one another on said level 1 self organizing map than other weighing vectors on said level 1 self organizing map.

5. The method of claim 4 where said level 2 self organizing map is trained after said level 1 self organizing map.