System and method for providing localization using inertial sensors

Abstract

A system and method for providing localization, including, during a training phase: obtaining a training dataset of accelerations, angular velocities, and known locations over time of vehicles moving in a defined area; and training a machine learning model to provide location estimation in the defined area based on the accelerations and angular velocities using the training dataset; and during runtime phase: obtaining runtime accelerations and angular velocities over time of a vehicle moving in the defined area; and using the trained model to obtain current location of the vehicle based on the runtime acceleration and angular velocities.

1 . A method for providing localization, the method comprising, using a processor:
 during a training phase:
 obtaining a training dataset of accelerations, angular velocities, and known locations over time of vehicles moving in a defined area; and 
 training a machine learning model to provide location estimation in the defined area based on the accelerations and angular velocities using the training dataset; and 
   during runtime phase:
 obtaining runtime accelerations and angular velocities over time of a vehicle moving in the defined area; and 
 using the trained model to obtain current location of the vehicle based on the runtime acceleration and angular velocities. 
   
     
     
         2 . The method of  claim 1 , wherein the accelerations, angular velocities of the training set and the runtime acceleration and angular velocities are measured using at least one inertial measurement unit (IMU). 
     
     
         3 . The method of  claim 2 , wherein the IMU comprises at least one three-dimensional accelerometer and at least one three-dimensional gyroscope. 
     
     
         4 . The method of  claim 1 , wherein the machine learning model is a neural network. 
     
     
         5 . The method of  claim 1 , comprising:
 during the training phase:
 extracting features from the accelerations and angular velocities of the training dataset and adding the features to the training dataset; and 
   during the runtime phase:
 extracting runtime features from the runtime accelerations and angular velocities; and 
 using the trained model to obtain the current location of the vehicle based on the runtime acceleration, the runtime angular velocities and the runtime features. 
   
     
     
         6 . The method of  claim 5 , wherein the features are selected from the list consisting of velocity and horizontal slope. 
     
     
         7 . The method of  claim 1 , wherein during the training phase, the known locations are obtained from at least one of the list consisting of: a global navigation satellite system (GNSS) receiver and a real-time kinematic (RTK) positioning system. 
     
     
         8 . The method of  claim 1 , wherein the defined area comprises a route. 
     
     
         9 . The method of  claim 1 , comprising dividing mapping of the defined area into segments, wherein the location is provided as a segment in which the vehicle is located. 
     
     
         10 . The method of  claim 1 , comprising performing anomaly detection to find changes in defined area. 
     
     
         11 . The method of  claim 1 , comprising:
 obtaining readings from at least one sensor selected from the list consisting of: a camera, a GNSS receiver, a Lidar sensor and radio frequency (RF) sensor; and   using the readings to enhance an accuracy of the current location provided by the trained ML model.   
     
     
         12 . A system for providing localization, the system comprising:
 a memory; and   a processor configured to:
 during a training phase:
 obtain a training dataset of accelerations, angular velocities, and known locations over time of vehicles moving in a defined area; and 
 train a machine learning model to provide location estimation in the defined area based on the accelerations and angular velocities using the training dataset: 
 
 during a runtime phase:
 obtain runtime accelerations and angular velocities over time of a vehicle moving in the defined area; and 
 use the trained model to obtain current location of the vehicle based on the runtime acceleration and angular velocities. 
 
   
     
     
         13 . The system of  claim 12 , comprising at least one inertial measurement unit (IMU) attached to each of the vehicles and configured to measure the accelerations, angular velocities of the training set and the runtime acceleration and angular velocities. 
     
     
         14 . The system of  claim 13 , wherein the IMU comprises at least one three-dimensional accelerometer and at least one three-dimensional gyroscope. 
     
     
         15 . The system of  claim 12 , wherein the machine learning model is a neural network. 
     
     
         16 . The system of  claim 12 , wherein the processor is configured to:
 during the training phase:
 extract features from the accelerations and angular velocities of the training dataset and adding the features to the training dataset; and 
   during the runtime phase:
 extract runtime features from the runtime accelerations and angular velocities; and 
 use the trained model to obtain the current location of the vehicle based on the runtime acceleration, the runtime angular velocities and the runtime features. 
   
     
     
         17 . The system of  claim 16 , wherein the features are selected from the list consisting of velocity and horizontal slope. 
     
     
         18 . The system of  claim 12 , wherein the defined area comprises a defined route. 
     
     
         19 . The system of  claim 12 , wherein the processor is configured to divide mapping of the defined area into segments, wherein the processor is configured to provide the location as a segment in which the vehicle is located. 
     
     
         20 . The system of  claim 12 , wherein the processor is configured to perform anomaly detection to find changes in defined area.