Remote supervision is the safety net behind driverless vehicles, and Switzerland’s new Automated Driving Ordinance now sets clear rules for it. FEDRO has just published their report “Minimum requirements for an authorisation to remotely drive automated vehicles in Switzerland”, showing how much network delay a remote driver can manage, how to filter false alarms and what skills operators need. To unpack the key findings of the report, SAAM sat down with project leads Roland Scherwey and Raphael Murri. The questions below dig into the study’s practical lessons (from latency limits to operator training) and explain how they slot into Articles 34–43 of the ordinance so cities, fleet owners and regulators can move from pilot projects to everyday service.
1. Could you briefly summarise your project for our readers?
As of 1 March 2025, the Swiss Ordinance on Automated Driving (VAF/OCA) entered into force. This research report focuses on the ordinance’s third use case, which for the first time defines the minimum requirements for remote-operation systems to guarantee the safe and reliable running of automated vehicles. This use case concerns driverless vehicles that move without a driver on board but, under the ordinance, must be monitored in real time by an operator located in Switzerland.
These minimum requirements cover the automated vehicle, communication, IT security, the operator’s workstation and the qualifications of remote operators – see Figure 1.
The main findings of this research:
- Latency tolerance: Delays of up to 850 ms do not affect manoeuvrability at low speeds (max. 6 km/h).
- Relevance for the scenario: The issue of false-positive obstacle detection was confirmed as a critical practical problem, but can be resolved with the current LOXO and BFH system.
- Challenges for operators: Targeted training is essential to cope with high latency and complex manoeuvres.
- Taxonomy: A classification system for five Remote Operation Levels (ROL) was developed (see Figure 2), distinguishing monitoring, remote assistance, remote operation, remote control and driver-remote control. This differentiation is crucial to address the high complexity and variability of real-world driving situations and to intervene appropriately.
2. How does your project specifically contribute to the implementation of the Ordinance on Automated Driving (VAF/OCA)?
The new ordinance (VAF/OCA) requires operators to ensure that driverless vehicles are monitored during operation and can be given targeted support when needed or temporarily driven manually in exceptional situations. Our project provides the technical and methodological foundation for this and contributes to the implementation of the ordinance by:
- introducing a structured taxonomy (ROL1–ROL5) that corresponds to the operator roles set out in the ordinance,
- defining practical requirements for technical reliability, communication, cybersecurity and human intervention,
- providing validation criteria to verify that such systems comply with the VAF/OCA requirements.
The research report particularly supports Articles 34 to 43 of the ordinance, which define responsibilities, qualifications and technical conditions for the operation of driverless vehicles. In doing so, the project provides a comprehensive scientific basis for the official approval of remote-operation systems and offers practical guidelines for all industry stakeholders.
3. Which technical challenges occupied you most during the on-site tests?
A key challenge was dealing with network latency, especially in the teleoperation scenario (ROL2) where the operator directly controls the vehicle.
Our tests showed that delays of up to 850 ms at low speeds (≤ 6 km/h) do not cause significant loss of steering precision. Nevertheless, targeted operator training is essential to handle such latency safely.
Another technical focus was detecting and managing false alarms from obstacles. These false positives, for example triggered by leaves or plastic bags, could lead to unnecessary emergency braking. By tuning system responses and enabling operator intervention with the LOXO and BFH test vehicles, we were able to resolve this reliably.
Abbildung 5: Setup für Szenariotest „falsch positive“ – Hindernis installiert
In addition, typical issues—such as sensor errors in adverse weather, temporary connection losses or GNSS inaccuracies—proved relevant in practice and were specifically addressed across our eight test scenarios.
4. Have you proposed any technical or regulatory adjustments based on your project’s results?
The report sets out seven key recommendations that address both technical and regulatory aspects and can directly contribute to the further development of the VAF ordinance:
- Safety as a core principle:
Ensuring stable communication links, real-time responses and redundancy mechanisms should be at the heart of every update. The safety of all road users remains the top priority.
- Evolve scenarios:
Existing test and validation scenarios should be continuously expanded— for example, to cover higher speeds, complex urban environments or more extreme weather conditions.
- Integrate technological advancements:
New technologies such as 5G, adaptive streaming or edge computing must be included in the requirements to enable scalability and system stability, even in large fleets.
- Optimise ROL2 operation:
Teleoperation at low speeds requires improvements to operator interfaces, response logic and ergonomics. The psychological and cognitive demands on operators should also be given greater consideration.
- Regular review and update:
Requirements and standards must be continually revised—both with regard to technological innovations and regulatory developments at home and abroad.
- Operator training and certification:
Structured training programmes with practical exercises, simulation scenarios and final certification are needed. These should ensure that operators can act confidently even in complex situations.
- Harmonise with international standards:
National regulations should be aligned with international standards (e.g. UN R155, R156, ISO 16505) to ensure interoperability and scalability across borders.
These recommendations support the consistent implementation of Articles 34 to 43 of the VAF ordinance—covering operator requirements, system reliability, cybersecurity and operating licences—and at the same time lay the foundation for the ongoing evolution of the regulatory framework.
5. How do you see the evolution of the required skills for operators of automated vehicles?
With the introduction of driverless vehicles under the VAF, a new profession emerges—the Remote Operator. Their tasks and requirements vary greatly depending on the Remote Operation Level (ROL), ranging from monitoring to targeted intervention.
- At ROL5 (Monitoring), an operator can oversee multiple vehicles simultaneously—with no active intervention.
- At ROL3–ROL4 (Tele-assistance), the operator provides targeted support for specific decision processes, for example in unclear traffic situations or by suggesting alternative routes.
- At ROL2 (Tele-Driving/Teleoperation), the operator temporarily takes manual control of a single vehicle—e.g., to free it from a situation it cannot resolve on its own. According to Art. 35, the operator is considered the vehicle driver during this period.
The skills of an operator must therefore be developed in a targeted, level-appropriate manner. The research report therefore recommends the introduction of comprehensive training and certification programmes for Remote Operators, which should include, among other things:
- practical simulations of emergency scenarios,
- systematic training in human–machine interaction,
- and mandatory certification to ensure a high level of safety and competence.
6. What new challenges will arise for operators when transitioning from pilot operations to scaling driverless vehicles in real-world service?
With the growing proliferation of driverless vehicles—the third use case of the ordinance (VAF/OCA)—the challenges for operators shift when it comes to scaling these systems. The following aspects are in focus:
Limited intervention capacity per operator:
While an operator at ROL5 can monitor multiple vehicles, levels ROL2 to ROL4—involving active assistance or teleoperation by a remote operator—allow for only one vehicle to be managed at a time. As fleets grow, organizational bottlenecks arise that must be addressed through control-room concepts, prioritization and technical aids.
Increase in the total number of complex or unexpected situations:
As more driverless vehicles enter the field, the demand for remote operator support will rise when issues such as communication failures, difficult traffic scenarios, adverse weather or false alarms occur.
Scalable control-room infrastructure:
The Remote Operation Centre must be designed to operate stably and reliably even with large fleets and concurrent support requests. This requires a clear division of roles between remote operators and dispatchers, structured escalation processes, and technical and personnel redundancies to ensure flexible response during disruptions or peak loads.
Ergonomics and staff workload:
With higher vehicle volumes comes increased cognitive strain on staff—whether making decisions under time pressure, managing information overload, or coping with stress. This calls for both technical support (e.g. more automated suggestions or assistance systems) and targeted training plus organizational relief.
7. What kind of training and certification will be required for operators to effectively handle critical situations?
In Switzerland, there is so far no unified, legally mandated standard training. However, important initial foundations have been laid in projects that have now concluded:
- In the TaaS (Transport as a Service) project, a training concept for Remote Operators was developed.
- Within the Migronomous and Dynamic Micro-hub pilot projects, initial training sessions and operational deployments have already taken place.
For the safe and legally compliant operation of driverless vehicles under the VAF/OCA ordinance, comprehensive, structured training and certification programmes for Remote Operators are required. In our research report, we recommend that such a programme include the following elements, adapted to each Remote Operation Level (ROL):
- Practical simulations of real emergency scenarios, such as communication failures, false obstacle detections or ODD violations
- In-depth knowledge of vehicle systems, interfaces, sensors and teleoperation technologies
- Understanding of the applicable legal framework, especially the VAF (e.g. Articles 34–38 on operator duties and driving competence)
- Training in human factors, including response confidence, stress management and situational decision-making under uncertainty
Regular refresher courses and recertification are recommended to keep pace with evolving technical and regulatory requirements.
8. What are the main technological limitations still hindering the deployment of automated vehicles in Switzerland?
The deployment of driverless vehicles in Switzerland under the VAF/OCA ordinance is currently slowed less by fundamental technological gaps and more by real-world operational requirements and scaling challenges. Three points are especially central:
- Connectivity:
A stable, robust data link between an automated vehicle and the Remote Operation Centre is a prerequisite for remote monitoring and teleoperation. In practice, however, network-coverage gaps, lack of QoS guarantees and mobile data-volume limits hinder consistently reliable operation. Our research therefore recommends using adaptive streaming strategies and dynamic resource allocation to make efficient use of uplink capacity.
- System performance under real-world conditions:
Although the vehicles tested showed robust performance in scenarios like false-positive obstacle detections or GNSS disturbances, system reliability under variable environmental and traffic influences remains a core challenge. Regular updates to scenarios and testing strategies are required to address emerging operational demands early.
- Limitations in ROL2 operation (Tele-Drive/Teleoperation):
Tests demonstrated good latency tolerance, making teleoperation at low speeds essentially feasible. Yet beyond connectivity, the design of remote-operator interfaces and the cognitive and ergonomic demands placed on operators present further hurdles. Targeted development at both system and human-machine levels, as well as specialized training, will be needed.
9. What further technological developments are necessary to ensure the reliability of remote control for automated vehicles?
This question goes somewhat beyond the original scope of this research work. Based on our project and current trends, several key areas can be identified where technological and regulatory advances will be crucial to improving the reliability of automated-vehicle remote control systems:
Advances in sensing and environmental perception:
Further development of sensors (range, resolution, sensitivity) and intelligent fusion of data from LiDAR, radar, cameras and GNSS will enable more robust situational awareness even under adverse conditions.
Research and development for Remote Operation Centres:
As bridge technologies between manual and fully automated driving, teleoperation and tele-assistance within Remote Operation Centres are essential to cover operational gaps—for example in unclear situations. In future, AI could support remote operators’ decision-making by providing automatic risk assessments, action recommendations or vehicle-prioritisation in multi-vehicle setups.
Evolution of communication networks:
Networks offering guaranteed Quality of Service, redundant transmission channels and intelligent traffic monitoring will be vital to detect and compensate for latent interruptions, latency spikes and coverage fluctuations at an early stage. Adaptive streaming and dynamic resource allocation will also be key for operating large fleets.
The project also recommends establishing a national governance framework to continuously evaluate and integrate standards and developments. The aim is to align the VAF/OCA requirements with technical practice through clear processes for approval, evolution and evaluation.
10. Tests have shown that latencies up to 850 ms at speeds of up to 6 km/h pose no problem. What about higher speeds?
Tests with the LOXO and BFH vehicles showed that latencies up to 850 ms at speeds up to 6 km/h in ROL2 operation (tele-drive/teleoperation) were judged practicable by remote operators. Even at 1000–1250 ms, manoeuvrability was essentially maintained, though it required greater concentration.
In general, higher speeds make latency more critical, since the safety-critical window for intervention narrows. However, it is important to distinguish:
- ROL3–ROL5 (tele-assistance to monitoring): Operators do not directly control the vehicle, so latency has virtually no safety impact.
- ROL2 (tele-drive/teleoperation): This mode is not intended for continuous operation but to free a vehicle from situations it cannot resolve itself (e.g. a bottleneck or ambiguous obstacle). In such cases, speed can be limited to 6 km/h, which tests show allows safe remote control at up to 850 ms latency.
Currently, there are no tested concepts for higher speeds in ROL2, as they would carry increased risk.
11. The detection of false-positive obstacles was identified as a critical issue. Does your project offer a solution?
Yes—within the research project, false-positive obstacle detection was specifically addressed in scenario 8 (“False Positive”).
In this scenario, we tested the system’s response when harmless objects like branches or plastic bags are misclassified as obstacles, triggering an automatic emergency braking (AEB). These tests showed that:
- Both test vehicles reliably detected such objects and performed the emergency stop as designed.
- Operators then used ROL2 teleoperation—either by steering around the object or slowly driving over it (< 1 km/h)—to resolve the situation safely.
Both approaches were judged practical and safe, and could become standard in the future. The project’s results clearly demonstrate that combining automated detection with targeted remote intervention is a promising way to handle uncertainties in urban environments safely and efficiently—thereby enhancing operational stability and public acceptance of automated vehicles.
12. What role does the taxonomy of Remote Operation Levels (ROL) developed in the project play in implementing the VAF ordinance?
The five-level taxonomy of Remote Operation Levels (ROL1–ROL5) developed in the project is a central result and forms a methodological bridge between technology, operational practice and regulation.
It clearly differentiates between monitoring (ROL5), assistance (ROL3–4), direct remote driving (ROL2) and local remote control (ROL1). In doing so, it supports:
- the precise definition of the remote operator’s intervention authority,
- the assignment of responsibilities as required by the VAF ordinance (Articles 34–36),
- and the design of training, certification and operational concepts.
The taxonomy was developed and validated on a scenario basis, can be adapted flexibly to different vehicle types and use cases, and provides a practical tool for meeting the VAF requirements.
13. What potential do Remote Operation Systems have—for example, in logistics, industry or agriculture?
Remote Operation Systems offer great potential in, for example:
Last-mile logistics: Supporting automated delivery vehicles in complex manoeuvres within urban areas.
Industry and campus logistics: Operating automated vehicles on private sites or in secure environments—ideal for ROL3–4.
Agriculture: Monitoring and occasional remote control of automated machinery, e.g. when encountering obstacles, faults or challenging terrain.
The combination of high automation with targeted human intervention (tele-operation or assistance) enables economically viable deployment even where full vehicle automation is not yet possible.
14. What importance do national platforms like SAAM or SwissMoves have for the further development of automated mobility?
Associations such as SAAM and SwissMoves make a vital contribution to advancing automated mobility in Switzerland by:
- coordinating collaboration between authorities, research and industry,
- enabling and networking pilot projects (e.g. with LOXO, TaaS or the Swiss Transit Lab),
- and providing a forum for knowledge exchange.
In this project, they not only supported the content work but also ensured that the results were practical, interoperable and aligned with implementing the VAF/OCA ordinance.
15. What potential do you see for a national training and certification system for Remote Operators?
The potential is very high—and from the project’s perspective, urgently needed. With the introduction of driverless vehicles under the VAF/OCA ordinance, a new professional profile of Remote Operator emerges, requiring clearly defined skills and solid training. The research report therefore recommends:
- developing standardized training programmes, including practical simulations and legal foundations,
- as well as mandatory certifications tailored to each ROL level.
Initial steps have already been taken in the TaaS project (training concept) and through pilot projects with LOXO (Migronomous and Dynamic Micro-hub). A national certification system would help establish uniform quality standards and strengthen confidence in remote operations.
16. To what extent can tele-operation serve as a bridge technology to cover existing operational or infrastructural gaps?
Tele-operation (ROL2) can fill operational gaps where automation is not yet sufficiently reliable. In combination with monitoring (ROL5) and tele-assistance (ROL3–4), tele-operation (ROL2) is a key technology in the transition from manual to fully automated driving and can therefore bridge gaps in systems that are not yet fully reliable.
17. How was cybersecurity integrated into the project, and what methods were used to evaluate and test the requirements?
Cybersecurity was integrated by adhering to standards such as UN Regulation No. 155, ISO/IEC 27001:2024, ISA/IEC 62443, the Swiss ICT Minimum Standard and the Federal Data Protection Act. These standards help meet cybersecurity requirements and can be validated in different ways:
- Process requirements through organizational audits against specific norms or regulations.
- Technical requirements at multiple levels, from unit tests to integration tests, assessing component interaction.
- Advanced testing such as fuzzing or penetration tests (pen-tests).
In the research project, 80 of the 193 defined cybersecurity requirements were validated via pen-tests, conducted according to frameworks like the Cyber Kill Chain, MITRE ATT&CK, OWASP and ISA. Tests focused on the LOXO Alpha 1 vehicle and its LOXO TCC remote-control station to address the specific cyber-risks of remote-operation systems.
The methodology simulated an attacker targeting the external, internal and physical perimeters of the LOXO Alpha 1 and the control station. The goals were to:
- Compromise the remote-operator stations,
- Take control of one of the automated vehicles,
- Identify vulnerabilities in the security design of the vehicles.
These tests helped validate the robustness and completeness of the cybersecurity requirements and ensure compliance with international standards.
18. Who can competently support authorities or operators in implementing the new ordinance (VAF/OCA Section 3: Driverless Vehicles)?
The research consortium (HEIA-FR/ROSAS, BFH, DTC Dynamic Test Centre AG, CertX SA, Eraneos Switzerland AG, LOXO AG) and the associations SAAM (Swiss Association for Autonomous Mobility) and SwissMoves have the expertise and practical experience to actively guide implementation of Section 3 (driverless vehicles) of the ordinance. They can assist with:
- managing approval procedures under VAF/OCA,
- technical evaluation of Remote Operation Systems,
- strategic advice on integrating these systems into regular operations,
- training and certification concepts for operators.
Which specific insights did you gain by testing different types of automated vehicles (e.g. for goods and passenger transport)?
In our trials we used various automated vehicle types to cover realistic and diverse application scenarios. The LOXO automated vehicle allowed us to study challenges in autonomous goods transport in depth, while the BFH autonomous shuttle bus let us address requirements for automated passenger transport.
Key differences emerged in safety requirements:
- Goods transport prioritizes cargo securing, reliable obstacle detection and system availability.
- Passenger transport adds demands for exceptionally reliable and early hazard detection, smooth driving manoeuvres and comprehensive occupant-safety measures. Tele-operation in passenger transport also requires greater attention to psychological factors, as passengers must build trust in driving style and system safety.
Our tests also highlighted the crucial role of operator experience and skill. Both LOXO’s Claudio Panizza and BFH’s Ahmed Hanachi—experienced tele-operators and project leads—were instrumental in safely managing critical situations and generating meaningful insights for advancing remote-operation systems.
Follow us on LinkedIn to stay
up-to-date with SAAM news!
