Vehicle Weight Estimation Using Sensor Fusion for Grip Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous Vehicles (AV) and Advanced Driver Assistance Systems (ADAS) face challenges in accurate localization, vehicle weight estimation, and grip determination due to limitations in visual sensors' performance, especially in adverse weather conditions and varying road conditions, which affect vehicle behavior and safety.
Innovation Solution
A method and system that utilize vehicle sensor measurements, such as height, fuel consumption, and velocity data, to calculate an evaluated weight of the vehicle using energy coefficients, applying machine learning to determine weight estimates for each path segment and associating quality attributes with weight estimates, enabling real-time weight calculation and improved grip determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual sensors (camera, LiDAR, RADAR) are used for localization, then localization accuracy can reach centimeter level, but the sensors fail to function 100% of the time due to bad weather, headlight blindness, jamming, and scene changes
Solution Approach 1:
The patent introduces map data as an intermediary reference system. Instead of relying solely on visual sensors comparing current scenes with stored images, the system uses pre-stored map data (including elevation, road geometry, and feature information) as a mediator to validate and supplement sensor-based localization, maintaining accuracy when visual sensors fail due to weather or lighting conditions
Solution Approach 2:
The patent creates a multi-functional localization system that combines visual sensor data with map-based localization. The map database serves multiple purposes: providing geometric constraints, elevation information, road feature references, and validation data. This universal approach allows the system to function across diverse conditions where single-sensor systems would fail
2Measurement precision
If vast computational and storage resources are allocated to estimate vehicle behavior parameters, then the estimation accuracy improves, but the system complexity and resource requirements increase significantly
Solution Approach 1:
The patent transforms the vehicle behavior estimation problem by changing the parameters used for estimation. Instead of directly estimating complex behavior parameters from raw sensor data, the system uses map-based geometric constraints, elevation changes, and road gradient information as input parameters. This parameter transformation reduces computational complexity while maintaining estimation accuracy through physics-based relationships between vehicle state and terrain characteristics
3Reliability
If the vehicle uses ABS system to prevent wheel locking, then braking safety improves, but energy consumption increases due to continuous brake pumping
Solution Approach 1:
The patent applies preliminary action by estimating grip conditions and road characteristics before braking events occur. Using map data about road surface type, gradient, and historical grip measurements, the system predicts optimal braking parameters in advance. This allows the vehicle to prepare appropriate braking force levels before wheel locking becomes a risk, reducing the need for reactive brake pumping and associated energy consumption
Solution Approach 2:
The system implements feedback by continuously monitoring actual vehicle deceleration, wheel speed, and comparing it with predicted values based on map data and estimated grip conditions. This feedback loop allows the system to adjust braking force in real-time, maintaining safety while minimizing energy-wasting brake pumping by correcting deviations before they require ABS intervention
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A vehicle monitor and a method for monitoring a vehicle.