Technical Summary: Towards a Joint Understanding of Remote Operation for Vehicles in Public Road Traffic

Problem Statement

Current Driving Automation Systems (DAS) are unable to handle all traffic situations that human drivers can manage, particularly when operating within their defined Operational Design Domains (ODD). To enable driverless mobility services without an in-vehicle fallback driver, remote operation (or teleoperation) is increasingly discussed as a bridging technology. However, the field suffers from inconsistent terminology and fragmented knowledge. Different research groups, standardization bodies (e.g., SAE J3016, 5GAA, BSI), and legal frameworks use varying definitions for concepts like "remote assistance," "remote driving," and "remote monitoring." These inconsistencies create communication barriers between interdisciplinary stakeholders (engineers, psychologists, legal experts, regulators) and impede the development of standardized safety concepts, training, and policy. Specifically, there is a lack of a unified framework that distinguishes between different forms of remote operation based on how information is processed between the human operator and the vehicle.

Methodology

The authors propose a joint understanding of remote operation by developing a conceptual framework grounded in information processing theory. The methodology involves three primary steps:

1. Literature and Standard Review: A comprehensive analysis of existing terminology in research, standardization (SAE J3016, 5GAA, BSI Flex 1887), and legal contexts (e.g., German StVG, AFGBV) was conducted to identify conflicting definitions and recurring dimensions of remote operation.
2. Framework Development: The authors constructed a model (Figure 1) that maps the interaction between two distinct information processing domains:
   • Human Information Processing: Modeled after Parasuraman et al. (2000), consisting of "Sense" (perception), "Plan" (decision-making), and "Act" (response execution).
   • Technical Information Processing: Modeled after the standard DAS architecture (Sense, Plan, Act), encompassing sensor data, perception, planning (global/local), and control/actuation.
3. Derivation of Definitions: The distinction between remote operation types is derived not from the physical location of the operator or the type of input device used, but from where the information transmitted from the remote control center is processed within the vehicle's DAS architecture.

Key Contributions

1. A Unified Information-Processing Framework

The paper introduces a framework that visualizes remote operation as a closed loop between a Remote Operation Control Center (containing the operator, workplace, and safety concepts) and the Vehicle (containing the DAS, remote operation modules, and vehicle-side safety concepts). The framework emphasizes that the critical differentiator is the integration point of human input into the vehicle's technical processing chain.

2. Redefinition of Remote Assistance vs. Remote Driving

The authors propose that the distinction between remote assistance and remote driving is determined by the processing stage of the vehicle's DAS:

• Remote Assistance: The remote operator provides information that is processed at the Sense (e.g., object classification, environmental context) or Plan (e.g., maneuver selection, path validation) stages of the DAS. The vehicle's DAS retains execution of the Dynamic Driving Task (DDT) and performs the "Act" stage (control/actuation). The operator acts as a remote assistant, not a driver.
• Remote Driving: The remote operator provides information that is processed at the Act stage of the DAS (i.e., directly influencing controllers or actuators). In this scenario, the remote operator executes the DDT, effectively replacing the vehicle's automated control for the duration of the intervention.

3. Clarification of Common Misconceptions

The paper explicitly states that:

• The design of the remote operator workplace (e.g., presence of a steering wheel or pedals) does not determine whether an activity is remote driving or assistance. A steering wheel input can constitute remote assistance if the vehicle interprets it as a contextual cue (e.g., "is it safe to cross?") rather than a direct control command.
• Remote monitoring is not a separate category of remote operation but a prerequisite activity inherent to both assistance and driving.

4. Characterization Across Dimensions

The authors map their new definitions against frequently discussed dimensions (Table 1), such as triggers, frequency, collision avoidance responsibility, and ODD boundaries. This demonstrates how the proposed distinction clarifies safety responsibilities (e.g., who is responsible for collision avoidance when the vehicle is outside its ODD).

Results and Application

The framework successfully categorizes existing concepts and legal definitions:

• SAE J3016 (2021): The paper aligns its definitions with SAE J3016 but clarifies that the "Act" stage processing is the definitive boundary.
• German Legal Context: The authors apply their framework to the German legal system. They identify the "Technical Supervisor" (Technische Aufsicht) under the German Road Traffic Act (StVG) as a form of Remote Assistance, as the vehicle must still verify maneuvers and execute the DDT. Conversely, the newly enacted Road Traffic Remote Driving Ordinance (StVFernLV) corresponds to Remote Driving, where the remote operator takes over the DDT execution.
• Interdisciplinary Utility: The framework provides a common language for engineers, psychologists, and regulators, allowing them to specify exactly which information processing stages are involved in a given use case without being tied to specific hardware implementations.

Significance and Claims

The paper claims that its primary significance lies in establishing a design-neutral, process-oriented foundation for the discourse on remote operation. By shifting the focus from "what the operator does" (e.g., steering) to "how the vehicle processes the input," the authors provide a robust basis for:

• Standardization: Creating consistent definitions for future technical standards.
• Safety Concepts: Enabling precise safety requirements based on the specific interaction mode (e.g., different safety protocols for when the vehicle is in "Act" control vs. "Plan" support).
• Regulatory Clarity: Assisting policymakers in distinguishing between legal requirements for remote assistance versus remote driving.

The authors maintain a modest stance, noting that the framework does not address non-driving tasks (e.g., fleet management), passenger communication, or operator training requirements, which remain areas for future research. The work serves as a "common ground" to facilitate further interdisciplinary discourse and the safe implementation of remote operation in public road traffic.