Vehicle Wheel Load Prediction for Road Disturbance Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicle computers are unable to accurately determine road disturbances due to limitations in sensor configurations and packaging, leading to inadequate accounting for road surface deviations, which affects vehicle operation.
Innovation Solution
A system utilizing a machine learning program to predict wheel load and vertical displacement, enabling identification of road disturbances and updating map data to account for these deviations, with communication between vehicles and remote servers for aggregated data updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vehicle sensors are configured to collect environmental data, then the vehicle can operate based on acquired data, but the sensors cannot accurately determine road disturbances due to packaging constraints
Solution Approach 1:
The patent introduces machine learning models as an intermediary between the limited sensor data and road disturbance detection. The model processes sensor inputs (accelerometer, gyroscope, steering angle, etc.) to infer road conditions that cannot be directly measured, effectively mediating the gap between sensor capabilities and detection requirements
Solution Approach 2:
The patent replaces direct mechanical measurement of road disturbances with a computational approach using machine learning. Instead of using complex mechanical sensors to directly measure road surface deviations, the system uses software-based prediction models that process available sensor data to estimate road conditions
2Reliability
If the vehicle computer uses existing sensor data to operate the vehicle, then basic vehicle control is achieved, but the computer cannot account for road disturbances affecting vehicle performance
Solution Approach 1:
The patent performs preliminary prediction of road disturbances before the vehicle encounters them. The machine learning model continuously predicts upcoming road conditions based on current sensor data, allowing the vehicle control system to prepare appropriate responses in advance, thereby improving operational reliability
Solution Approach 2:
The patent implements a feedback mechanism where predicted road disturbances are fed back into the vehicle control system. The model continuously compares predicted disturbances with actual vehicle responses, using this feedback to refine predictions and improve the accuracy of road condition assessment
3Productivity
If map data is updated with road disturbance information, then future vehicle operations can be improved, but real-time accurate detection is still limited by sensor constraints
Solution Approach 1:
The system performs preliminary detection and characterization of road disturbances, storing this information in map data for future reference. By detecting and recording disturbance patterns in advance, the system enables more efficient vehicle operations when encountering known road conditions, even though real-time measurement precision remains limited
Data Source
AI summary
Based on inputting collected data of a host vehicle to a machine learning program, a predicted load on a wheel of the host vehicle and a predicted vertical displacement of the wheel are determined via output from the machine learning program. A road disturbance traversed by the host vehicle is identified based on at least one of the predicted load and the predicted vertical displacement. Map data is updated to include the road disturbance.