Method and system for training machine learning models on sensor nodes

Abstract

Disclosed are apparatus and methods for automatically training a sensor node to detect anomalies in an environment. An indication is sent to the sensor node to initiate training by the sensor node to detect anomalies in object(s) in the environment based on sensor data generated by a sensor operable to detect signals from the one or more objects in the environment. After training is initiated, the sensor node automatically trains a model in communication with the sensor to detect anomalies in the one or more objects in the environment, wherein such training is based on the sensor data. After the model is trained, the model to detect anomalies in the object(s) in the environment is executed by the sensor node.

What is claimed is: 
     
         1 . A method of automatically training a sensor node to detect anomalies in one or more objects in an environment, the method comprising:
 sending to the sensor node an indication to initiate training by the sensor node to detect anomalies in the one or more objects in the environment based on sensor data generated by a sensor operable to detect signals from the one or more objects in the environment;   after training is initiated, the sensor node automatically training a model in communication with the sensor to detect anomalies in the one or more objects in the environment, wherein such training is based on the sensor data; and   after the model is trained, executing by the sensor node the model to detect anomalies in the one or more objects in the environment.   
     
     
         2 . The method of  claim 1 , wherein the sensor resides on at least one of the sensor node and the one or more objects in the environment. 
     
     
         3 . The method of  claim 1 , wherein the sensor node further comprises a controller for initiating training and execution of the model, and wherein the model is trained without being generated by a processor that is external to the sensor node. 
     
     
         4 . The method of  claim 1 , further comprising repeating operations for sending the sensor data and training for a plurality of different sensor nodes, wherein the different sensor nodes reside in different environments. 
     
     
         5 . The method of  claim 1 , wherein the indication to initiate training is received via a user activating a physical input mechanism of the sensor node. 
     
     
         6 . The method of  claim 1 , wherein the indication to initiate training is received as a command by the sensor node. 
     
     
         7 . The method of  claim 1 , wherein the sensor is a motion sensor, acoustic sensor, touch sensor, environmental sensor, vision sensor, temperature sensor, ambient pressure sensor, humidity sensor, vibration sensor, communications sensor, infrared radiation sensor, color sensor, light intensity sensor, magnetic field sensor, air quality sensor, chemical sensor, piezoelectric sensor or electromagnetic interference sensor. 
     
     
         8 . The method of  claim 1 , wherein the sensor node is configured to discriminate between various motion or static operations of the one or more objects in the environment. 
     
     
         9 . The method of  claim 1 , wherein the sensor node is configured to wirelessly transmit at least one of the signals, the sensor data, the training and the model to a cloud storage system. 
     
     
         10 . The method of  claim 1 , wherein the sensor node is configured to transmit continuously, intermittently, with or without interconnect connection during any and all operations. 
     
     
         11 . The method of  claim 1 , wherein the indication to initiate training is received after the sensor node is moved to the environment for a duration of time in which the sensor detects sensor signals that do not indicate anomalies in the one or more objects in the environment. 
     
     
         12 . The method of  claim 1 , further comprising:
 during execution of the model, detecting anomalies in the one or more objects in the environment;   requesting by the sensor node, confirmation of the detected anomalies in the one or more objects in the environment;   repeating the training by the sensor node, if the detected anomalies in the one or more objects in the environment are not confirmed; and   continuing to execute the model to detect anomalies in the one or more objects in the environment if the detected anomalies in the one or more objects in the environment are confirmed.   
     
     
         13 . A sensor node system for automatically training a model to detect anomalies in one or more objects in an environment, the system comprising:
 a sensor operable to detect sensor signals from the one or more objects in the environment and generate sensor data based on such sensor signals; and   a controller for receiving the sensor data from the sensor and performing the following operations: 
 sending an indication to initiate training by the sensor node system to detect anomalies in the one or more objects in the environment based on the sensor data; 
 after training is initiated, automatically training a model in communication with the sensor to detect anomalies in the one or more objects in the environment, wherein such training is based on the sensor data; and 
 after the model is trained, executing the model to detect anomalies in the one or more objects in the environment. 
   
     
     
         14 . The system of  claim 13 , wherein the model is trained without being generated by a processor that is external to the sensor node system. 
     
     
         15 . The system of  claim 13 , wherein the indication to initiate training is received via a user activating a physical input mechanism of the sensor node system. 
     
     
         16 . The system of  claim 13 , wherein the indication to initiate training is received as a command by the sensor node system. 
     
     
         17 . The system of  claim 13 , wherein the sensor is a motion sensor, acoustic sensor, touch sensor, environmental sensor, vision sensor, temperature sensor, ambient pressure sensor, humidity sensor, vibration sensor, communications sensor, infrared radiation sensor, color sensor, light intensity sensor, magnetic field sensor, air quality sensor, chemical sensor, piezoelectric sensor or electromagnetic interference sensor. 
     
     
         18 . The system of  claim 13 , wherein the indication to initiate training is received after the sensor node system is moved to the environment for a duration of time in which the sensor detects sensor signals that do not indicate anomalies in the one or more objects in the environment. 
     
     
         19 . The system of  claim 13 , wherein the controller is further configured for:
 during execution of the model, detecting anomalies in the one or more objects in the environment;   requesting confirmation of the detected anomalies in the one or more objects in the environment;   repeating the training if the detected anomalies in the one or more objects in the environment are not confirmed; and   continuing to execute the model to detect anomalies in the one or more objects in the environment if the detected anomalies in the one or more objects in the environment are confirmed.   
     
     
         20 . A non-transitory machine-readable medium, comprising executable instructions that, when executed by a processor, facilitate performance of operations, comprising:
 sending to the sensor node an indication to initiate training by the sensor node to detect anomalies in one or more objects in an environment based on sensor data generated by a sensor that resides on at least one of the sensor node and the one or more objects in the environment, the sensor operable to detect sensor signals from the one or more objects in the environment;   after training is initiated, the sensor node automatically training a model in communication with the sensor to detect anomalies in the one or more objects in the environment, wherein such training is based on the sensor data; and   after the model is trained, executing by the sensor node the model to detect anomalies in the one or more objects in the environment.   
     
     
         21 . The non-transitory machine-readable medium of  claim 20 , further comprising:
 during execution of the model, detecting anomalies in the one or more objects in the environment;   requesting by the sensor node confirmation of the detected anomalies in the one or more objects in the environment;   repeating the training by the sensor node if the detected anomalies in the one or more objects in the environment are not confirmed; and   continuing to execute the model to detect anomalies in the one or more objects in the environment if the detected anomalies in the one or more objects in the environment are confirmed.