Remote Assistance or Remote Driving: The Impact of Operational Design Domains on ADS-Supporting Systems Selection

2025-07-18

AI Summary

For SAE Level 4 autonomous driving systems operating without an on-site driver, remote intervention—via Remote Driving Services (RDS) or Remote Assistance Systems (RAS)—is critical. However, current RDS/RAS selection practices lack a systematic, ODD-aligned evaluation framework, leading to suboptimal architectural fit. Method: This paper introduces the first decision-making methodology that tightly couples Operational Design Domain (ODD) modeling with use-case analysis. Grounded in the PEGASUS framework, it establishes a structured assessment system quantifying RDS/RAS suitability across three dimensions: response capability, interaction complexity, and scenario adaptability. Contribution/Results: The method enables scientifically grounded, operationally diverse remote support system selection, significantly enhancing ADS safety and deployment efficiency. It delivers a reusable, standardized, and industry-applicable tool for architecture-level remote intervention system evaluation.

Technology Category

• Planning, Routing, and Scheduling: Model-Based Reasoning
• Multiagent Systems: Distributed Problem Solving
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

High level Automated Driving Systems (ADS) can handle many situations, but they still encounter situations where human intervention is required. In systems where a physical driver is present in the vehicle, typically SAE Level 3 systems, this intervention is relatively straightforward and is handled by the in-vehicle driver. However, the complexity increases for Level 4 systems, where, in most cases, no physical driver remains in the vehicle. The two common industry solutions for this challenge are the integration of a remote support system, such as a Remote Driving System (RDS) or Remote Assistance System (RAS). While it is clear that ADS will require one of these systems, it is less clear how the suitability of either system for a particular ADS application should be evaluated. Currently, the selection process often focuses on system architecture as well as its design and integration challenges. Furthermore, since many ADS developers choose to develop remote system solutions in-house, it is advantageous to select the simpler approach to streamline development and integration efforts. While these decision points are certainly relevant, this approach overlooks the most critical factors: the use cases and the complementarity of the ADS and the remote support system within the context of the Operational Design Design Domain (ODD). This paper proposes a structured approach for selecting between RDS and RAS as an ADS support system, based on the defined ODD and use case analysis. To achieve this, the paper applies the PEGASUS framework to systematically describe and analyze the ODD. A structured framework is introduced to evaluate and select the most suitable remote support system for an ADS based on clearly defined criteria.

Problem

• Evaluating suitability of Remote Driving vs Remote Assistance for Level 4 ADS
• Selecting ADS support systems based on ODD and use case analysis
• Addressing lack of structured framework for RDS/RAS selection criteria

Innovation

• Uses PEGASUS framework for ODD analysis
• Selects RDS or RAS based on ODD
• Evaluates remote support via use cases