Search

DE-102025145869-A1 - EDGE PROCESSING OF TIMESERIES DATA IN A VEHICLE

DE102025145869A1DE 102025145869 A1DE102025145869 A1DE 102025145869A1DE-102025145869-A1

Abstract

In one embodiment, a method includes receiving, by a vehicle control system for a vehicle, one or more first time-series inputs associated with a first time window, wherein the one or more first time-series inputs relate to one or more signals generated in the vehicle. The method also includes applying, by the vehicle control system, a first function to the one or more first time-series inputs to generate a first value of a first metric for the first time window. Furthermore, the method includes storing, by the vehicle control system, the first value of the first metric in the vehicle in a first bin associated with the first time window.

Inventors

  • Aditya Purohit
  • Sunil Bhagwath

Assignees

  • RIVIAN IP HOLDINGS, LLC

Dates

Publication Date
20260513
Application Date
20251107
Priority Date
20241108

Claims (20)

  1. A method for in-vehicle edge processing of time-series data, comprising: Receiving, by a vehicle control system, one or more first time-series inputs associated with a first time window, wherein the one or more first time-series inputs relate to one or more signals generated in the vehicle; Applying, by the vehicle control system, a first function to the one or more first time-series inputs to generate a first value of a first metric for the first time window; and Store, by the vehicle control system, the first value of the first metric in the vehicle in a first bin associated with the first time window.
  2. Procedure according to Claim 1 , wherein the first bin comprises a plurality of values of the first metric for a plurality of time windows of a period, wherein the plurality of values includes the first value of the first metric for the first time window.
  3. Procedure according to Claim 2 , wherein the first bin implements a ring buffer configured to store a predetermined number of values of the first metric; and storing the first value of the first metric involves overwriting at least one other value of the plurality of values of the first metric.
  4. Procedure according to Claim 2 , where the multitude of time windows corresponds to a predetermined time resolution for the first metric.
  5. Procedure according to Claim 2 , further comprising: Receiving, by the vehicle control system, one or more second time-series inputs that are assigned to a second time window; Applying, by the vehicle control system, the first function to the one or more second time-series inputs to generate a second value of the first metric for the second time window; and storing, by the vehicle control system, the second value of the first metric in the first bin in assignment to the second time window.
  6. Procedure according to Claim 2 , further comprising: Receiving, by the vehicle control system, a request for a value of the first metric over an aggregation window; aggregating, by the vehicle control system, the multitude of values of the first bin based on the aggregation window to generate the requested value; and outputting the requested value that was generated.
  7. Procedure according to Claim 6 , wherein the requested value that was generated is output to a machine learning model in the vehicle and the method further includes using the machine learning model based on the requested value that was generated.
  8. Procedure according to Claim 1 , further comprising: applying, by the vehicle control system, a second function to the one or more first time series inputs to generate a value of a second metric for the first time window; and storing, by the vehicle control system, the value of the second metric in a second bin in association with the first time window.
  9. Procedure according to Claim 8 , wherein the second bin comprises a plurality of values of the second metric for a plurality of time windows of a period, wherein the plurality of values includes the value of the second metric for the first time window.
  10. Procedure according to Claim 1 , furthermore, including the rejection of one or more initial time series inputs in response to the fact that the first value of the first metric is stored in the first bin.
  11. Procedure according to Claim 1 , where the first value of the first metric comprises an aggregated value that is associated with one or more of the first time series inputs.
  12. Procedure according to Claim 1 , where the first value of the first metric includes at least one of the following: maximum, minimum, sum or number of the one or more first time series inputs.
  13. System for in-vehicle edge processing of time-series data, comprising: one or more memories containing executable instructions; and one or more processors that exchange data with one or more memories and are configured to execute the executable instructions for receiving one or more initial time-series inputs associated with an initial time window, wherein the one or more initial time-series inputs relate to one or more in a Retrieve vehicle-generated signals; apply a first function to the one or more first time series inputs to generate a first value of a first metric for the first time window; and store the first value of the first metric in the vehicle in a first bin in association with the first time window.
  14. System according Claim 13 , wherein the first bin comprises a plurality of values of the first metric for a plurality of time windows of a period, wherein the plurality of values includes the first value of the first metric for the first time window.
  15. System according Claim 14 , where: the first bin implements a ring buffer configured to store a predetermined number of values of the first metric; and storing the first value of the first metric involves overwriting at least one other value of the plurality of values of the first metric.
  16. System according Claim 14 , wherein the one or more processors are further configured to execute the instructions for receiving one or more second time-series inputs associated with a second time window; applying the first function to the one or more second time-series inputs to generate a second value of the first metric for the second time window; and storing the second value of the first metric in the first bin in association with the second time window.
  17. System according Claim 14 , wherein the one or more processors are further configured to execute the instructions for receiving a request for a value of the first metric over an aggregation window; aggregating the plurality of values of the first bin based on the aggregation window to produce the requested value; and outputting the requested value that was produced.
  18. System according to Claim 17 , wherein the requested value that was generated is output to a machine learning model in the vehicle, wherein the one or more processors are further configured to execute the executable instructions for using the machine learning model based on the requested value that was generated.
  19. The system according to Claim 13 , wherein the one or more processors are further configured to execute the executable instructions to discard the one or more first time-series inputs in response to the first value of the first metric being stored in the first bin.
  20. A computer program product comprising a non-transitory, computer-usable medium embodying computer-readable program code, wherein the computer-readable program code is adapted to be executed to implement a method comprising: receiving one or more first time-series inputs associated with a first time window, wherein the one or more first time-series inputs relate to one or more signals generated in a vehicle; applying, by the vehicle control system, a first function to the one or more first time-series inputs to generate a first value of a first metric for the first time window; and storing, by the vehicle control system, the first value of the first metric in the vehicle in a first bin associated with the first time window.

Description

CROSS-REFERENCE TO RELATED REGISTRATIONS This application incorporates the rights arising from the priority of the preliminary US patent application with serial number 63/718,443 , filed on November 8, 2024, claim which is hereby incorporated by reference in its entirety as if fully and completely set forth herein. INTRODUCTION This disclosure relates to edge processing. In particular, this disclosure relates to techniques for processing time-series data on an edge device, such as a vehicle. SUMMARY In one embodiment, a general aspect includes a method for in-vehicle edge processing of time-series data. The method includes receiving, by a vehicle control system, one or more first time-series inputs associated with a first time window, wherein the one or more first time-series inputs relate to one or more signals generated in the vehicle. The method also includes applying, by the vehicle control system, a first function to the one or more first time-series inputs to generate a first value of a first metric for the first time window. Furthermore, the method includes storing, by the vehicle control system, the first value of the first metric in the vehicle in a first bin associated with the first time window. In one embodiment, another general aspect includes a system for in-vehicle edge processing of time-series data. The system includes one or more memories containing executable instructions. The system also includes one or more processors that communicate with one or more memories and are configured to execute the executable instructions to receive one or more initial time-series inputs associated with a first time window, the one or more initial time-series inputs relating to one or more signals generated in a vehicle. The one or more processors are further configured to execute the executable instructions to apply a first function to the one or more initial time-series inputs to generate an initial value of a first metric for the first time window. The one or more processors are further configured to execute the executable instructions to store the initial value of the first metric in the vehicle in a first bin associated with the first time window. In one embodiment, a further general aspect includes a computer program product comprising a non-transitory, computer-usable medium containing computer-readable program code. The computer-readable program code is adapted to be executed to implement a method. The method includes receiving one or more first time-series inputs associated with a first time window, wherein the one or more first time-series inputs relate to one or more signals generated in a vehicle. The method also includes applying a first function to the one or more first time-series inputs to generate a first value of a first metric for the first time window. Furthermore, the method includes storing the first value of the first metric in the vehicle in a first bin associated with the first time window. BRIEF DESCRIPTION OF THE DRAWINGS 1A illustrates an exemplary vehicle according to specific embodiments.1B illustrates a vehicle chassis according to certain embodiments.2A is a schematic block diagram of components of a vehicle according to specific embodiments.2B is a schematic block diagram of alternative components of a vehicle according to specific embodiments.3 illustrates an exemplary environment for a telematics control module (TCM) according to certain embodiments.4 illustrates an exemplary operation of TCM by 3 according to certain embodiments.5 illustrates an exemplary environment for estimating the charging time of the vehicle battery according to certain embodiments.6 illustrates an example of a process for managing metric values based on time series inputs in a vehicle, according to certain embodiments.7 This illustrates an example of a process for responding to queries of metric values stored in a vehicle, according to certain embodiments. DETAILED DESCRIPTION Machine learning (ML) models are increasingly being implemented on edge devices to provide these devices with enhanced functionality. For example, an ML model can provide more accurate predictions of a vehicle's battery charging time than traditional physical models. Since performing inferences using an ML model generally requires significant computational and bandwidth resources, implementing an ML model on an edge device presents a number of challenges. One such challenge is the limited computing capacity of edge devices. To circumvent this limitation, conventional approaches typically perform inference procedures using a cloud-based service. While such approaches significantly reduce the computing and power requirements of edge devices, they are less effective when inference procedures are performed with large input datasets of time-series data. For example, in vehicles, electronic control units (ECUs) can generate large amounts of raw time-series data at high sampling rates (e.g., every 10 ms or less). Consequently, transf