EV Range Prediction Using Route Segmentation and Fleet Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional range estimation techniques for electrified vehicles suffer from inaccuracies and variability in predicting remaining range due to unpredictable factors like vehicle acceleration, road conditions, and battery temperature imbalance, leading to 'range anxiety' that hinders market adoption.
Innovation Solution
A range estimation system utilizing a pattern recognition machine learning model that combines real-time and historical data from a pool of OEM vehicles traveling similar routes to estimate range depletion, applying statistical and data management techniques to filter and update data in real-time, segmenting routes, and identifying vehicle combinations for accurate predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional range estimation techniques are used, then the system is simple to implement, but the measurement precision of remaining range is poor due to inaccuracies and variability
Solution Approach 1:
The patent introduces a cloud-based computing server as an intermediary between the vehicle's control system and the range estimation process. The server receives real-time data from multiple vehicles, processes it through pattern recognition models, and returns range predictions. This intermediary approach enables high-precision estimation without overcomplicating the vehicle's onboard systems, as the computational burden is shifted to the cloud.
Solution Approach 2:
The patent uses historical data from other vehicles (copies of operational patterns) to train pattern recognition models that predict range depletion. Instead of relying on complex real-time calculations for each vehicle, the system creates and applies learned patterns from copied operational data, significantly improving accuracy while maintaining computational simplicity on the vehicle side.
2Measurement precision
If real-time data from multiple vehicles is collected and processed, then range estimation accuracy is improved, but the loss of time for data processing increases
Solution Approach 1:
The system performs preliminary actions by pre-processing and storing historical data from multiple vehicles in the cloud before it is needed for prediction. The pattern recognition models are trained in advance on this pre-processed data, allowing rapid inference during actual range estimation without the need for time-consuming real-time analysis of historical datasets.
Solution Approach 2:
The patent implements continuous data collection and processing through the cloud server, where historical data is continuously updated and patterns are continuously refined. This continuous operation allows the system to maintain high accuracy without interrupting data processing, as the cloud server operates asynchronously in the background while vehicles are in motion.
3Adaptability or versatility
If pattern recognition machine learning models are applied, then adaptability to different driving conditions is improved, but the device complexity increases
Solution Approach 1:
The complex pattern recognition machine learning models are hosted on the cloud computing server as an intermediary, shielding the vehicle's control system from computational complexity. The server handles the adaptive pattern matching and model inference, allowing the vehicle to benefit from high adaptability to various driving conditions without carrying complex computational algorithms onboard.
Solution Approach 2:
The system adapts to different driving conditions by dynamically selecting and applying appropriate pattern recognition models based on route characteristics, vehicle type, and environmental factors. The cloud server adjusts model parameters and data processing approaches according to changing conditions, enabling versatile adaptability through parameter variation rather than complex structural changes.
Data Source
AI summary
A range estimation technique for an original equipment manufacturer (OEM) electrified vehicle involves determining, by an OEM computing server, a route for the electrified vehicle and a set of operating parameters of the electrified vehicle and historical data for other OEM vehicles traveling along the determined route or another route that is similar to the determined route, segmenting the determined route into a plurality of route segments and, for each route segment, identifying one or more combinations of OEM vehicles that traveled that route segment, and estimating a range depletion for each route segment based on the historical data for the respective identified combinations of OEM vehicles and a total range depletion for the determined route based on the estimated range depletions for each route segment.