Abnormal motion detector and monitor

Abstract

In an embodiment, a seizure monitor provides intelligent epilepsy seizure detection, monitoring, and alerting for epilepsy patients or people with seizures. In an embodiment, the seizure monitor may be a wearable, non-intrusive, passive monitoring device that does not require any insertion or ingestion into the human body. In an embodiment, the seizure monitor may include several output options for outputting the accelerometer/gyro or other motion sensor data and video data, so that the data may be immediately validated and/or remotely viewed. The device alerts are communicated to the outside care givers via wireless or wired medium. The device may also support recording of accelerometer or other motion sensor data and video data, which can be reviewed later for further analysis and/or diagnosis. The device and invention is also used and applicable for other body motion disorders or detection and diagnostics.

What is claimed is: 
     
         1 . A method to detect a seizure, comprising:
 electronically collecting motion data produced by a sensor physically associated with a person;   determining from the collected motion data via a processor communicatively coupled to the sensor at least one value characterizing the motion data;   comparing the at least one value characterizing the motion data to at least one corresponding value characterizing abnormal motion;   determining whether the seizure has occurred based on the comparing; and   activating a seizure alert signal when the determined seizure has occurred.   
     
     
         2 . The method of  claim 1 , wherein the at least one corresponding value characterizing abnormal motion includes a motion pattern model characterizing a seizure, and wherein the determining includes determining whether the motion data matches the motion pattern model characterizing the seizure within a predetermined tolerance. 
     
     
         3 . The method of  claim 1 , further comprising:
 retrieving historic motion data previously stored in a memory; and   applying the historic motion data to generate the at least one corresponding value characterizing abnormal motion.   
     
     
         4 . The method of  claim 1 , wherein the motion data is video data, the method further comprising:
 identifying at least one distinguishable feature in the video data;   determining locations of the distinguishable feature across multiple picture frames of the video data; and   determining a motion path for the distinguishable feature, the motion path for the distinguishable feature being the at least one value characterizing the motion, wherein the comparing includes comparing the motion path for the distinguishable feature to a motion path characterizing the seizure, and wherein the determining whether a seizure has occurred includes determining whether the motion path for the distinguishable feature matches the motion path characterizing the seizure within a predetermined tolerance.   
     
     
         5 . The method of  claim 1 , wherein the motion data represents oscillatory motion. 
     
     
         6 . The method of  claim 5 , wherein the at least one value characterizing the motion data represents a frequency of oscillation, and wherein the at least one corresponding value characterizing abnormal motion represents a predetermined threshold, and wherein determining whether the seizure has occurred is based on the frequency of oscillation being higher than the predetermined threshold. 
     
     
         7 . The method of  claim 6 , the motion data is derived from an optical flow or feature point analysis. 
     
     
         8 . The method of  claim 6 , wherein the motion data is derived from a plurality of motion vectors. 
     
     
         9 . A method to detect abnormal motion in a person, comprising:
 producing first sensor data at a first time, the first sensor data representing a first physical state of a portion of a body of a person, the first sensor data produced by at least one of an accelerometer, a gyro sensor, and a camera;   producing second sensor data at a second time, the second time after the first time, the second sensor data representing a second physical state of the portion of the body of the person, the second sensor data produced by the at least one of the accelerometer, the gyro sensor, and the camera;   mathematically associating the first sensor data with the second sensor data to obtain a motion value;   comparing the motion value to an abnormal motion threshold; and   activating an abnormal motion signal based on the comparing.   
     
     
         10 . The method of  claim 9 , wherein the first sensor data and the second sensor data are digital values representing at least one of amplitude, frequency, and acceleration. 
     
     
         11 . The method of  claim 9 , further comprising:
 repeating, a plurality of times, the acts of producing first and second sensor data, the act of mathematically associating the first and second sensor data, and the act of comparing;   tracking at least one point of the portion of the body of the person; and   deriving oscillation information as the motion value.   
     
     
         12 . The method of  claim 9 , further comprising:
 repeating, a plurality of times, the acts of producing first and second sensor data, the act of mathematically associating the first and second sensor data, and the act of comparing;   computing at least one motion vector;   generating a signature representing the at least one motion vector, the signature being the motion value.   
     
     
         13 . The method of  claim 9 , further comprising:
 physically attaching the at least one of the accelerometer, the gyro sensor, and the camera to either the person or furniture the person is in contact with.   
     
     
         14 . The method of  claim 9 , wherein the abnormal motion is indicative of a motion disorder, and wherein the motion disorder is one of epilepsy, ataxia, dystonia, dyskinesia, Parkinson's disease, chorea, tremor, tics, myoclonus, and restless leg syndrome. 
     
     
         15 . A system for detecting a seizure or abnormal motion, said system comprising:
 an input means for inputting motion parameters, said input means configured for inputting a template;   a seizure detection engine; and   a processor configured for activating an alert,   wherein said processor is configured to be part of said seizure detection engine or separate from said seizure detection engine or both.   
     
     
         16 . The system of  claim 15 , further comprising at least one of a location determination hardware and a global positioning software for determining a location of a user, further comprising at least one of a camera for video capture, a microphone for audio capture, and a sensor for motion capture. 
     
     
         17 . The system of  claim 15 , wherein the template comprises threshold values comprising one or more of a value relating to at least one of a frequency of oscillations of a body part, oscillatory motion, frequency of oscillations, frequency of motion within a time frame, amplitude of the motion, duration of the motion, seizure intensity or abnormal motion intensity based on amplitude, seizure intensity or abnormal motion intensity based on frequency, acceleration magnitude peaks in a time frame, a first derivative of a magnitude of acceleration, and a second derivative of a magnitude of acceleration. 
     
     
         18 . The system of  claim 15 , wherein the seizure detection engine is configured for matching a motion pattern to the template, wherein the motion pattern is captured by one or more of an accelerometer, a gyroscopic sensor, a video-capture camera, and an audio-capture microphone, wherein said motion pattern is represented as a sum of basis functions multiplied by coefficients and determining if a magnitude of one or more coefficients crosses a predetermined threshold as an indication that at least one of a seizure or abnormal motion has occurred, wherein the matching the motion pattern is performed at least in part in a device that is remote from other functionality of the system, and wherein the template comprises a seizure data or abnormal motion threshold for at least one motion parameter, and past seizure data or abnormal motion data for a user of the system. 
     
     
         19 . The system of  claim 15 , further comprising at least one of a statistical model, a neural network, or a training routine for learning behaviors associated with a specific type of motion for detecting at least one of a seizure or abnormal motion. 
     
     
         20 . The system of  claim 15 , further comprising at least one or more of an option for a user to manually trigger an alert to a concerned party, an option for a requirement of a user confirmation prior to an alert being sent to a concerned party, an option for cancelling an alert to a concerned party, and an option to snooze an alert, and further comprising a transmitter capable of sending the alert to a remote location.