Automated near-fall detector

Abstract

A method of gait data collection, the method comprising collecting movement data, determining from the data a movement parameter that includes a third order derivative of position, comparing the movement parameter with a threshold value, and counting at least a near fall if the movement parameter exceeds the threshold value.

1 - 20 . (canceled) 
     
     
         21 . A method of determining a near fall event, wherein a user recovers from a momentary loss of balance without falling, the method comprising:
 electronically collecting movement data using a detector configured to measure acceleration of the user's body;   using a processor in communication with said detector, electronically determining at least one movement parameter value from said data collected by said detector, wherein said at least one movement parameter value is based on an acceleration pattern; and   using said processor, processing said at least one movement parameter value determined from said data collected using said detector to identify the near fall event, based on at least one threshold value.   
     
     
         22 . A method according to  claim 21 , wherein said at least one movement parameter value includes a measure of maximum acceleration. 
     
     
         23 . A method according to  claim 21 , wherein the method comprises determining a fall. 
     
     
         24 . A method according to  claim 21 , wherein determining comprises matching a pattern with respect to time of the movement data with a reference pattern. 
     
     
         25 . A method according to  claim 24 , wherein the reference pattern represents one of: a proper gait pattern, an improper gait pattern, and a gait pattern exhibiting at least one near fall event. 
     
     
         26 . A method according to  claim 24 , wherein the matching classifies the data as exhibiting a fall, a near fall event, or lack thereof. 
     
     
         27 . A method according to  claim 24 , wherein the matching comprises at least one of correlation, cross-correlation, wavelets matching or neural networks or a combination thereof. 
     
     
         28 . A method according to  claim 21 , wherein said processing said at least one movement parameter value includes electronically comparing said at least one movement parameter value with the at least one threshold value, and wherein said electronically comparing comprises comparing said measure of movement in a substantially vertical direction with said at least one threshold value to identify the near fall event. 
     
     
         29 . A method according to  claim 21 , wherein determining at least one movement parameter value from said data further includes determining a second movement parameter value;
 wherein said processing said at least one movement parameter value includes processing at least a said second movement parameter value;   wherein comparing said at least one movement parameter value further includes comparing said second movement parameter value with a second threshold value; and   wherein said method further includes counting at least the near fall event if a first movement parameter value exceeds a first threshold value and said second movement parameter value exceeds said second threshold value.   
     
     
         30 . A method according to  claim 29 , wherein said second movement parameter value includes one of the group consisting of; a rate of change of acceleration, an angular velocity, an anterior-posterior acceleration, and a medio-lateral acceleration. 
     
     
         31 . A method according to  claim 21 , wherein said at least one threshold value is a predetermined value. 
     
     
         32 . A method according to  claim 21 , wherein said method further includes storing a count of near fall events to provide at least one of: a quantitative measure of effectiveness of therapeutic interventions and quantifiable parameters for assessing a person. 
     
     
         33 . A method according to  claim 21 , wherein said movement data includes cyclic acceleration data; and
 wherein said electronically determining the at least one movement parameter value comprises:
 determining from said acceleration data periods of cycles; and 
 identifying a gait irregularity when a period of a cycle exceeds a threshold. 
   
     
     
         34 . A method according to  claim 33 , wherein said cyclic acceleration data includes peaks;
 wherein each cycle includes a cycle shape; and   wherein said electronically determining comprises at least one of:
 (a) determining from said acceleration data periods between said peaks; and
 identifying a gait irregularity when at least one of:
 a period between said peaks exceeds a threshold; and 
 a cycle shape varies above a threshold; 
 
 
 (b) determining from said acceleration data a cross-correlation between cycles; and
 identifying a gait irregularity when the cross-correlation between cycles exceeds a threshold; and 
 
 (c) determining from said data an acceleration frequency spread; and
 identifying an irregularity of a gait from said acceleration frequency spread. 
 
   
     
     
         35 . A method according to  claim 21 , wherein said at least one movement parameter value relates to movement in at least one of: a substantially an anterior-posterior direction and a substantially vertical direction. 
     
     
         36 . A method according to  claim 21 , wherein said using the processor includes electronically determining the at least one movement parameter value from collected acceleration data alone. 
     
     
         37 . A method according to  claim 21 , wherein said momentary loss of balance is during a gait. 
     
     
         38 . A method according to  claim 21 , wherein said at least one movement parameter value determined from said data collected comprises a plurality of movement parameter values including at least one movement parameter value related to a movement parameter in a substantially vertical direction;
 wherein said processing said at least one movement parameter value includes electronically determining, using said processor, from said data at least one irregularity of a gait, including identifying the near fall event during the gait, said electronically determining including electronically comparing each of said plurality of movement parameter values with an associated threshold value, wherein said electronically comparing comprises comparing said measure of maximum acceleration with said threshold value to identify the near fall event during the gait, including comparing said at least one movement parameter value related to the movement parameter in the substantially vertical direction determined from said data collected using said detector with a threshold value, to indicate the near fall event when each said movement parameter value related to the movement parameter in the substantially vertical direction exceeds an associated threshold value; and   wherein, if a predetermined combination of comparisons indicates the near fall event, said method further includes electronically storing a count of the near fall event in a memory.   
     
     
         39 . A method according to  claim 28 , the method comprising: 
       electronically recording a magnitude of said near fall event. 
     
     
         40 . A method according to  claim 38 , wherein said predetermined combination of movement parameters is a majority of said plurality of movement parameters; and
 wherein said electronically storing the count comprises electronically counting at least a near fall event if a majority of said combination of comparisons indicates a near fall event.   
     
     
         41 . A method according to  claim 21 , wherein said movement data includes a measure of maximum acceleration;
 wherein said method further includes using the processor in communication with said detector to electronically extract an indicator indicating a loss of control from said data collected, wherein said indicator indicating the loss of control includes at least one movement parameter value which exceeds a threshold value, wherein said electronically extracting an indicator comprises comparing said measure of maximum acceleration with said threshold value to identify a near fall event; and   wherein, if said indicator indicates said loss of control, said method further includes:
 electronically storing a count of near fall events in a memory; and 
 electronically recording a date or time for each said near fall event, in said memory.