Autonomous Fleet Coordination via Regional Profile Transfer

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous vehicle fleets face challenges in efficient coordination due to high costs and complexity of sensors, leading to underutilization and difficulty in predicting passenger demand outside regular hours, which complicates route and charging planning.

Innovation Solution

A vehicle fleet coordination system that generates a regional environmental profile based on similar regions, using historical and real-time data to optimize route planning, battery charging, and idle time management, while adapting to spatiotemporal and event-specific factors, including machine learning for real-time adjustments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If autonomous vehicles are deployed in small numbers at the beginning, then costs are reduced, but vehicle utilization is insufficient and downtime increases

Engineering Contradiction:
Improvenumber of autonomous vehiclesVSAvoidvehicle utilization rate
Core Design Contradiction:
Quantity of substanceVSProductivity

Solution Approach 1:

The system performs preliminary actions by predicting passenger demand in advance using historical data and machine learning models. It proactively assigns vehicles to predicted high-demand locations and times, and schedules charging during low-demand periods, thereby maximizing vehicle utilization without requiring a large fleet from the outset

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system continuously collects real-time data on vehicle locations, passenger demand, and charging status, then uses this feedback to dynamically adjust vehicle assignments and charging schedules. This closed-loop control enables optimal utilization of limited vehicles by adapting to changing conditions

Inventive Principle:
Principle #23Feedback

2Productivity

If vehicles travel with multiple passengers throughout the entire journey, then productivity increases, but route planning complexity and charging strategy complexity increase

Engineering Contradiction:
Improvevehicle utilization rateVSAvoidcoordination system complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system segments the journey into distinct phases (passenger pickup, transportation, passenger dropoff, charging) and optimizes each phase separately. It divides the fleet into individually manageable vehicles with specific assignments, allowing complex multi-passenger routes to be coordinated through simple, standardized vehicle-level control instructions

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The coordination system acts as an intermediary between passenger demand and vehicle operations. It processes complex demand patterns and translates them into simplified vehicle assignment instructions, managing route planning and charging strategies centrally without requiring complex onboard decision-making at each vehicle

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If predictive models use historical and real-time data, then demand prediction accuracy improves, but data processing requirements and computational complexity increase

Engineering Contradiction:
Improvedemand prediction accuracyVSAvoiddata processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system creates simplified representations (copies) of complex demand patterns by training machine learning models on historical data. These trained models then quickly predict future demand without requiring real-time processing of all raw historical data, reducing computational complexity while maintaining prediction accuracy

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system performs preliminary data processing by pre-training machine learning models offline using historical data. This preliminary action creates ready-to-use prediction models that can operate with minimal real-time computational resources, balancing accuracy requirements with processing complexity

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3997537B1Method for coordinating an autonomous vehicle fleet, and vehicle fleet coordination system
Publication Date: 2024.11.06 MERCEDES BENZ GROUP AG
  • EP3997537B1 patent drawingFigure 1

AI summary

The invention relates to a method for coordinating an autonomous vehicle fleet (12) having a multiplicity of autonomous motor vehicles (14) in a stipulated first region (16) by means of a vehicle fleet coordination system (10), in which a first regional surroundings profile (22) of the first region (16) for the vehicle fleet coordination system (10) is taken as a basis for coordinating the autonomous vehicle fleet (12), wherein coordination involves a respective position (P), which is dependent on the regional surroundings profile (22), in the first region (16) being driven to autonomously by means of a respective autonomous motor vehicle (14) from the multiplicity of autonomous motor vehicles (14), wherein the first regional surroundings profile (22) is generated on the basis of a second regional surroundings profile (24), stored in an electronic computing device (18) of the vehicle fleet coordination system (10), that is generated for a second region, which is independent of the first region (16). The invention also relates to a vehicle fleet coordination system (10).