Autonomous Driving Control via Multi-Sensor Wheel Dynamics Fusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing autonomous driving systems lack the capability to effectively control vehicle operations based on a combination of sensor data to achieve precise and real-time adjustments in wheel speed, inclination, and suspension settings for optimal performance in various driving conditions.

Innovation Solution

An information processing device that calculates control variables for wheel speed, inclination, and suspension using a combination of sensor information through deep learning, enabling precise control of autonomous driving operations every billionth of a second.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple sensor information combinations are used to calculate index values for controlling wheel speed, inclination, and suspension, then the precision of autonomous driving control is improved, but the calculation complexity and processing time increase

Engineering Contradiction:
Improvecontrol precisionVSAvoidcalculation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the autonomous driving control into multiple independent control targets (wheel speed, wheel inclination, suspension), with each target having dedicated sensor combinations and calculation processes. This allows parallel processing of different control aspects, improving overall precision without proportionally increasing sequential complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a combinatorial dimension by calculating index values from multiple different combinations of sensor information for each control target. Instead of using a single sensor set, it processes multiple sensor combinations (e.g., combining data from various sensors in different configurations) to generate multiple index values that are then aggregated, thereby enhancing control precision through multi-dimensional data analysis

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If control variables are calculated and aggregated from multiple sensor information combinations, then the reliability of autonomous driving control is improved, but the processing time increases

Engineering Contradiction:
Improvecontrol reliabilityVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary calculations by pre-defining multiple sensor information combinations and their corresponding index value calculations for different control targets. By preparing these calculation frameworks in advance, the system can rapidly aggregate results from multiple sensor combinations during real-time operation, improving reliability without proportionally increasing processing time

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements continuous calculation and aggregation of index values from multiple sensor combinations for all control targets simultaneously. Rather than processing control targets sequentially, the system maintains continuous parallel computation across wheel speed, inclination, and suspension controls, ensuring uninterrupted reliable control while minimizing total processing time

Inventive Principle:
Principle #20Continuity of useful action

3Measurement precision

If deep learning multivariate analysis is used to calculate control variables from index values, then the accuracy of autonomous driving control is improved, but the computational requirements and system complexity increase

Engineering Contradiction:
Improvecontrol accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces index values as intermediary variables between raw sensor information and final control variables. Multiple sensor combinations first generate multiple index values for each control target, which are then aggregated and processed through deep learning multivariate analysis to produce control variables. This intermediary layer simplifies the input requirements for deep learning models while maintaining high control accuracy

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4620764A1Information processing device and program
Publication Date: 2025.09.24 SOFTBANK GROUP CORP
  • EP4620764A1 patent drawingFigure 1
  • EP4620764A1 patent drawingFigure 2
  • EP4620764A1 patent drawingFigure 3

AI summary

An information processing device disclosed herein includes: a calculation unit that calculates, for each of plural combinations of a predetermined number of pieces of sensor information among plural pieces of sensor information included in a vehicle, index values for controlling a wheel speed and an inclination of each of four wheels of the vehicle, and suspensions that support the wheels for each of the wheel speed, the inclination, and the suspension, and calculates a control variable for each of the wheel speed, the inclination, and the suspension by aggregating the index values; and a control unit that controls autonomous driving on the basis of the control variable.