Easymile software suite: Thinking about autonomous driving across platforms

Easymile builds software for autonomous driving that is not tied to a single vehicle. It is based on a suite of clearly separated, interconnected modules. Perception, localization, route finding, decision logic, and actuator control are interlinked but remain addressable as independent components. This architectural principle makes it possible to serve different vehicle platforms without having to rebuild the entire system. When a new platform is added, the interfaces are adapted and the operating domain is configured instead of replacing the core.

In practice, this means that the software family drives various use cases – from people mover shuttles and factory logistics to special applications in enclosed areas. The modular approach is not an end in itself, but a prerequisite for the targeted modernization of individual elements. If an improved AI module is added to the perception area, it can be integrated as a block without compromising the functional safety and approvals of the rest of the chain. Equally important: deployment, maintenance, and diagnostic tools are part of the suite. Autonomous driving does not end with the first trip, but begins there – with updates, monitoring, tracking, and the disciplined handling of software versions over long life cycles.

Easymile and safety: Keeping critical functions in your own hands

The architecture deliberately separates safety-critical and non-safety-critical functions. Easymile retains control in the domains of braking, steering, stability, emergency strategies, and safe states. Functional safety standards, redundant paths, and formal verification apply here. The core processes are designed in such a way that decisions remain explainable and fail-operational or fail-safe behaviors are reliably triggered. This care distinguishes long-lasting systems from showcases that only work under perfect conditions. At the same time, the software remains open where it makes sense: fleet management, user interfaces, customer-oriented integrations, data pipelines, or dashboards can be tailored to the respective IT landscape. This separation protects the level of security while still allowing for the necessary speed of implementation. After all, fleet operations at an airport are different from those at an industrial park or a municipal operator. Those who take these differences seriously integrate modules, connect to existing control centers, and respect operational routines—without compromising the security-relevant chain. Another aspect: the suite is not just a driving stack, but a supply chain of tools that can be used to validate and track release statuses. Without this discipline, operators, insurers, and authorities will not have confidence in the system. The real innovation therefore lies as much in processes and evidence as in algorithms.

Precisely defining and mastering use cases in fleet management

Autonomous driving becomes economical when the operating domain is clearly defined. Easymile focuses on highly standardized environments: airports, factory premises, ports, campuses, and local transport on defined corridors. Here, driving profiles, traffic rules, interactions, and interfaces to control centers can be clearly defined. Fleet management not only handles dispatching and load planning, but also prioritization, incident handling, remote assistance, and reporting. Because each environment uses different control centers and backends, this level must be modular. Adapters connect the vehicle suite and customer systems without touching safety-critical paths. Partnerships play a central role in this picture. When vehicle manufacturers bring robust platforms that have been proven over many years, the software closes the circle: ODD definition, sensor layout, redundancies, validation, commissioning, operation. The example of ground handling shows how much the benefits can increase: standardized routes, clear safety zones, defined handover points, and real added value in 24/7 operation. The same idea applies to buses in separate corridors or for transport tasks on industrial sites. The decisive factor is not the show effect, but the resilient integration into routine processes – including training, documentation, and service. The suite also takes into account that operators do not overhaul their processes with every technological leap. Good fleet management abstracts differences in sensors, computing platforms, or communication channels so that teams can continue working while the software iterates.

Easymile and depreciation resistance: durability as a design principle

A key concept from an operator's perspective is “depreciation resistance” – the ability to function reliably over long periods of time, to be maintained in a predictable manner, and not to disrupt financial models with frequent complete replacements. Airports, municipal transport companies, and industrial enterprises procure vehicles with long cycles. Software linked to such assets must have the same horizon: stable, updatable, legally compliant, and documented. Easymile addresses this with an architecture that allows individual blocks to be replaced while maintaining overall approval and safety arguments. This allows perception or planning modules to be renewed as sensor technology and computing hardware evolve, without having to rework the entire chain. This “targeted renewal” bridges the gap between technological progress and the financial reality of operators. Only then does a real business case emerge: less downtime, predictable opex, clear responsibilities. At the same time, the focus is on markets that are viable without subsidies. Projects that come to a standstill after funding ends do not create demand. Depreciable software, flanked by services and reliable update paths, on the other hand, builds trust. This is precisely where modularity becomes an economic tool: components can be modernized, contracts remain valid, and systems retain their value. This attracts private capital and reduces risk because value development becomes transparent.

The European approach: many use cases, partnership-based roles, sustainable financing

Europe has a dense, heterogeneous mobility landscape. Autonomous driving here is not characterized by a single dominant use case, but by many specialized use cases. Easymile therefore advocates a path that begins where the framework conditions are already “serial”: standardized areas, mature vehicle platforms, clear ODDs. From these cores, it is possible to grow step by step into more complex environments – based on evidence, not headlines. This perspective also changes the structure of the industry. The early phase was characterized by turnkey thinking: one provider shoulders everything, including liability. This made sense for the start. However, the market will only become sustainable when roles are sharpened: vehicle manufacturers, software stack providers, integrators, operators, insurers, infrastructure providers. Segmentation allows for specialization, facilitates audits, and makes responsibilities transparent. Partnerships such as the one with TLD show how robust hardware and depreciable software complement each other in a meaningful way. Financing requires a sober view: politics can set and support frameworks, but it cannot replace venture capital or the economic viability of an operator. What is needed are viable business cases. Each use case is an economic unit with measurable benefits: time, safety, availability, energy costs, personnel deployment. When these variables are translated into contracts and service levels, capital follows – especially when the life cycle, upgrade logic, and liability are clearly regulated. This creates a market that lasts longer than a funding decision.

Conclusion: Easymile combines a core safety system, modularity, and real benefits for operators.

Easymile's approach is based on a robust separation: critical functions are kept in-house, while customer-specific flexibility is provided at the edges. The result is a software suite that serves different platforms, takes ODDs seriously, and respects real operating procedures. Fleet management, deployment tools, and maintenance processes are integral components, not downstream additions. The result is a system that enables technical progress without undermining the depreciation logic of operators. Europe benefits from this approach: many use cases, clear roles, reliable integration. Partnerships with mature vehicle platforms accelerate implementation, segmentation creates transparency, and private financing becomes more likely when longevity, upgradeability, and liability are clearly regulated. In this way, autonomous driving is not marketed as a vision, but is built up as a resilient service—step by step, domain-specific, with a focus on safety, cost-effectiveness, and operation. Ultimately, what counts is whether autonomous systems drive reliably, can be maintained cleanly, and still deliver exactly what is contractually promised after years of use. Easymile designs its software with this in mind: modular in its further development, strict in its safety, open to integration—and thus suitable for translating real demand into viable, long-term applications.