Remote Assistance for Autonomous Vehicle Decision-Making

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous vehicles may encounter situations where additional intelligence or human expertise is needed to enhance decision-making, particularly in scenarios involving impaired sensing, complex navigation, or safety concerns, but existing solutions require a human driver to be present or lack protocols for violating traffic laws.

Innovation Solution

A system and method for remotely assisting autonomous vehicle operation, which includes an onboard computer, a human expert interface, and an AI expert, allowing the vehicle to request assistance from a human expert or another autonomous vehicle, and enabling remote control or route modification to navigate challenging scenarios safely without violating traffic laws.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a human driver is present in the autonomous vehicle to provide expert opinion, then decision-making capability is improved, but device complexity and operational requirements worsen

Engineering Contradiction:
Improvedecision-making capabilityVSAvoidoperational requirements
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces a remote human expert system as an intermediary between the autonomous vehicle's onboard computer and the decision-making process. When the onboard computer encounters a scenario it cannot confidently resolve, it automatically connects to a remote human expert who provides guidance through a communication interface. This mediator approach improves decision-making reliability without requiring a human driver physically present in the vehicle, thus avoiding the complexity of coordinating human driver responsibilities and liabilities.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If the autonomous vehicle system encounters scenarios requiring expert opinion, then safety is improved, but response time worsens due to lack of established protocols

Engineering Contradiction:
ImprovesafetyVSAvoidresponse time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent establishes predetermined protocols and procedures for when and how the autonomous vehicle requests remote human expert assistance. The system includes pre-defined criteria for triggering remote connections, pre-established communication channels, and pre-organized expert availability frameworks. This preliminary preparation ensures that when safety-critical scenarios arise, the vehicle can quickly initiate remote assistance without ad-hoc decision-making delays, thus improving response time while maintaining safety.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If the autonomous vehicle requests remote assistance frequently, then decision-making quality is improved, but communication bandwidth and system complexity worsen

Engineering Contradiction:
Improvedecision-making qualityVSAvoidcommunication bandwidth
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The patent implements a hierarchical assistance system where the onboard computer first attempts to resolve scenarios using its own algorithms and local data. Remote human expert assistance is requested only for specific scenarios that exceed the onboard computer's confidence threshold or capability. This localized application of remote assistance—applying it only where needed rather than universally—improves decision-making quality for critical scenarios while minimizing unnecessary communication bandwidth consumption and system complexity.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS10459440B2System and method for remotely assisting autonomous vehicle operation
Publication Date: 2019.10.29 GM GLOBAL TECHNOLOGY OPERATIONS LLC
  • US10459440B2 patent drawing
  • US10459440B2 patent drawing
  • US10459440B2 patent drawing

AI summary

Systems and methods are provided for remotely assisting an autonomous vehicle. The method includes: aggregating sensor data from the autonomous vehicle; identifying an assistance-desired scenario; generating an assistance request based on the sensor data; transmitting the assistance request to a remote assistance interface; and receiving and processing a response to the assistance request. The remote assistance interface includes a remote assistance interface that is used in generating the response to the assistance request.